Dissertations / Theses: 'Customer churn prediction'

1

TANNEEDI, NAREN NAGA PAVAN PRITHVI. "Customer Churn Prediction Using Big Data Analytics." Thesis, Blekinge Tekniska Högskola, Institutionen för kommunikationssystem, 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-13518.

Full text

Abstract:

Customer churn is always a grievous issue for the Telecom industry as customers do not hesitate to leave if they don’t find what they are looking for. They certainly want competitive pricing, value for money and above all, high quality service. Customer churning is directly related to customer satisfaction. It’s a known fact that the cost of customer acquisition is far greater than cost of customer retention, that makes retention a crucial business prototype. There is no standard model which addresses the churning issues of global telecom service providers accurately. BigData analytics with Machine Learning were found to be an efficient way for identifying churn. This thesis aims to predict customer churn using Big Data analytics, namely a J48 decision tree on a Java based benchmark tool, WEKA. Three different datasets from various sources were considered; first includes Telecom operator’s six month aggregate active and churned users’ data usage volumes, second includes globally surveyed data and third dataset comprises of individual weekly data usage analysis of 22 android customers along with their average quality, annoyance and churn scores by accompanying theses. Statistical analyses and J48 Decision trees were drawn for three different datasets. From the statistics of normalized volumes, autocorrelations were small owing to reliable confidence intervals, but confidence intervals were overlapping and close by, therefore no much significance could be noticed, henceforth no strong trends could be observed. From decision tree analytics, decision trees with 52%, 70% and 95% accuracies were achieved for three different data sources respectively. Data preprocessing, data normalization and feature selection have shown to be prominently influential. Monthly data volumes have not shown much decision power. Average Quality, Churn Risk and to some extent, Annoyance scores may point out a probable churner. Weekly data volumes with customer’s recent history and necessary attributes like age, gender, tenure, bill, contract, data plan, etc., are pivotal for churn prediction.

APA, Harvard, Vancouver, ISO, and other styles

2

Hadden, John. "A customer profiling methodology for churn prediction." Thesis, Cranfield University, 2008. http://dspace.lib.cranfield.ac.uk/handle/1826/3508.

Full text

Abstract:

As markets have become increasingly saturated, companies have acknowledged that their business strategies need to focus on identifying those customers who are most likely to churn. To address this, a method is required that can identify these customers, so that proactive retention campaigns can be deployed in a bid to retain them. To further complicate this, retention campaigns can be costly. To reduce cost and maximise effectiveness, churn prediction has to be as accurate as possible to ensure that only the customers who are planning to switch their service providers are being targeted for retention. Current techniques and research as identified by literature focus primarily on the instantaneous prediction of customer churn. Much work has been invested in this method of churn prediction and significant advancement has been made. However one of the major drawbacks of current research is that the methods available do not provide adequate time for companies to identify and retain the predicted churners. There is a lack of time element in churn prediction. Current research also fails to acknowledge the expensive problem of misclassifying non-churners as churners. In addition, most research efforts base their analysis on customer demographic and usage data that can breach governing regulations. It is proposed in this research that customer complaints and repairs data could prove a suitable alternative. The doctoral research presented in this thesis aims to develop a customer profiling methodology for predicting churn in advance, while keeping the misclassification levels to a minimum. The proposed methodology incorporates time element in the prediction of customer churn for maximising future churn capture by identifying a potential loss of customer at the earliest possible point. Three case studies are identified and carried out for validating the proposed methodology using repairs and complaints data. Finally, the results from the proposed methodology are compared against popular churn prediction techniques reported in literature. The research demonstrates that customers can be placed into one of several profiles clusters according to their interactions with the service provider. Based on this, an estimate is possible regarding when the customer can be expected to terminate his/her service with the company. The proposed methodology produces better results compared to the current state-of-the-art techniques.

APA, Harvard, Vancouver, ISO, and other styles

3

Ljungehed, Jesper. "Predicting Customer Churn Using Recurrent Neural Networks." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-210670.

Full text

Abstract:

Churn prediction is used to identify customers that are becoming less loyal and is an important tool for companies that want to stay competitive in a rapidly growing market. In retail, a dynamic definition of churn is needed to identify churners correctly. Customer Lifetime Value (CLV) is the monetary value of a customer relationship. No change in CLV for a given customer indicates a decrease in loyalty. This thesis proposes a novel approach to churn prediction. The proposed model uses a Recurrent Neural Network to identify churners based on Customer Lifetime Value time series regression. The results show that the model performs better than random. This thesis also investigated the use of the K-means algorithm as a replacement to a rule-extraction algorithm. The K-means algorithm contributed to a more comprehensive analytical context regarding the churn prediction of the proposed model.
Illojalitet prediktering används för att identifiera kunder som är påväg att bli mindre lojala och är ett hjälpsamt verktyg för att ett företag ska kunna driva en konkurrenskraftig verksamhet. I detaljhandel behöves en dynamisk definition av illojalitet för att korrekt kunna identifera illojala kunder. Kundens livstidsvärde är ett mått på monetärt värde av en kundrelation. En avstannad förändring av detta värde indikerar en minskning av kundens lojalitet. Denna rapport föreslår en ny metod för att utföra illojalitet prediktering. Den föreslagna metoden består av ett återkommande neuralt nätverk som används för att identifiera illojalitet hos kunder genom att prediktera kunders livstidsvärde. Resultaten visar att den föreslagna modellen presterar bättre jämfört med slumpmässig metod. Rapporten undersöker också användningen av en k-medelvärdesalgoritm som ett substitut för en regelextraktionsalgoritm. K-medelsalgoritm bidrog till en mer omfattande analys av illojalitet predikteringen.

APA, Harvard, Vancouver, ISO, and other styles

4

Axén, Maja, and Jennifer Karlberg. "Binary Classification for Predicting Customer Churn." Thesis, Umeå universitet, Institutionen för matematik och matematisk statistik, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-171892.

Full text

Abstract:

Predicting when a customer is about to turn to a competitor can be difficult, yet extremely valuable from a business perspective. The moment a customer stops being considered a customer is known as churn, a widely researched topic in several industries when dealing with subscription-services. However, in industries with non-subscription services and products, defining churn can be a daunting task and the existing literature does not fully cover this field. Therefore, this thesis can be seen as a contribution to current research, specially when not having a set definition for churn. A definition for churn, adjusted to DIAKRIT’s business, is created. DIAKRIT is a company working in the real estate industry, which faces many challenges, such as a huge seasonality. The prediction was approached as a supervised problem, where three different Machine Learning methods were used: Logistic Regression, Random Forest and Support Vector Machine. The variables used in the predictions are predominantly activity data. With a relatively high accuracy and AUC-score, Random Forest was concluded to be the most reliable model. It is however clear that the model cannot separate between the classes perfectly. It was also visible that the Random Forest model produces a relatively high precision. Thereby, it can be settled that even though the model is not flawless the customers predicted to churn are very likely to churn.
Att prediktera när en kund är påväg att vända sig till en konkurrent kan vara svårt, dock kan det visa sig extremt värdefullt ur ett affärsperspektiv. När en kund slutar vara kund benäms det ofta som kundbortfall eller ”churn”. Detta är ett ämne som är brett forskat på i flertalet olika industrier, men då ofta i situationer med prenumenationstjänster. När man inte har en prenumerationstjänst försvåras uppgiften att definera churn och existerande studier brister i att analysera detta. Denna uppsats kan därför ses som ett bidrag till nuvarande litteratur, i synnerhet i fall där ingen tydlig definition för churn existerar. En definition för churn, anpassad efter DIAKRIT och deras affärsstruktur har skapats i det här projektet. DIAKRIT är verksamma i fastighetsbranschen, en industri som har flera utmaningar, bland annat en extrem säsongsvariaton. För att genomföra prediktionerna användes tre olika maskininlärningamodeller: Logistisk Regression, Random Forest och Support Vector Machine. De variabler som användes är mestadels aktivitetsdata. Med relativt hög noggranhet och AUC-värde anses Random Forest vara mest pålitlig. Modellen kan dock inte separera mellan de två klasserna perfekt. Random Forest modellen visade sig också genera en hög precision. Därför kan slutsatsen dras att även om modellen inte är felfri verkar det som att kunderna predikterade som churn mest sannolikt kommer churna.

APA, Harvard, Vancouver, ISO, and other styles

5

MOUNIKA, REDDY CHANDIRI. "Customer Churn Predictive Heuristics from Operator and Users' Perspective." Thesis, Blekinge Tekniska Högskola, Institutionen för kommunikationssystem, 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-13452.

Full text

Abstract:

Telecommunication organizations are confronting in expanding client administration weight as they launch various user-desired services. Conveying poor client encounters puts client connections and incomes at danger. One of the metrics used by telecommunications companies to determine their relationship with customers is “Churn”. After substantial research in the field of churn prediction over many years, Big Data analytics with Data Mining techniques was found to be an efficient way for identifying churn. These techniques are usually applied to predict customer churn by building models, pattern classification and learning from historical data. Although some work has already been undertaken with regards to users’ perspective, it appears to be in its infancy. The aim of this thesis is to validate churn predictive heuristics from the operator perspective and close to user end. Conducting experiments with different sections of people regarding their data usage, designing a model, which is close to the user end and fitting with the data obtained through the survey done. Correlating the examined churn indicators and their validation, validation with the traffic volume variation with the users’ feedback collected by accompanying theses. A Literature review is done to analyze previous works and find out the difficulties faced in analyzing the users’ feeling, also to understand methodologies to get around problems in handling the churn prediction algorithms accuracy. Experiments are conducted with different sections of people across the globe. Their experiences with quality of calls, data and if they are looking to change in future, what would be their reasons of churn be, are analyzed. Their feedback will be validated using existing heuristics. The collected data set is analyzed by statistical analysis and validated for different datasets obtained by operators’ data. Also statistical and Big Data analysis has been done with data provided by an operator’s active and churned customers monthly data volume usage. A possible correlation of the user churn with users’ feedback will be studied by calculating the percentages and further correlate the results with that of the operators’ data and the data produced by the mobile app. The results show that the monthly volumes have not shown much decision power and the need for additional attributes such as higher time resolution, age, gender and others are needed. Whereas the survey done globally has shown similarities with the operator’s customers’ feedback and issues “around the globe” such a data plan issues, pricing, issues with connectivity and speed. Nevertheless, data preprocessing and feature selection has shown to be the key factors. Churn predictive models have given a better classification of 69.7 % when more attributes were provided. Telecom Operators’ data classification have given an accuracy of 51.7 % after preprocessing and for the variables we choose. Finally, a close observation of the end user revealed the possibility to yield a much higher classification precision of 95.2 %.

APA, Harvard, Vancouver, ISO, and other styles

6

Tryggvadottir, Valgerdur. "Customer Churn Prediction for PC Games : Probability of churn predicted for big-spenders usingsupervised machine learning." Thesis, KTH, Optimeringslära och systemteori, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-254198.

Full text

Abstract:

Paradox Interactive is a Swedish video game developer and publisher which has players all around the world. Paradox’s largest platform in terms of amount of players and revenue is the PC. The goal of this thesis was to make a churn predic-tion model to predict the probability of players churning in order to know which players to focus on in retention campaigns. Since the purpose of churn prediction is to minimize loss due to customers churning the focus was on big-spenders (whales) in Paradox PC games. In order to define which players are big-spenders the spending for players over a 12 month rolling period (from 2016-01-01 until 2018-12-31) was investigated. The players spending more than the 95th-percentile of the total spending for each pe-riod were defined as whales. Defining when a whale has churned, i.e. stopped being a big-spender in Paradox PC games, was done by looking at how many days had passed since the players bought something. A whale has churned if he has not bought anything for the past 28 days. When data had been collected about the whales the data set was prepared for a number of di˙erent supervised machine learning methods. Logistic Regression, L1 Regularized Logistic Regression, Decision Tree and Random Forest were the meth-ods tested. Random Forest performed best in terms of AUC, with AUC = 0.7162. The conclusion is that it seems to be possible to predict the probability of churning for Paradox whales. It might be possible to improve the model further by investi-gating more data and fine tuning the definition of churn.
Paradox Interactive är en svensk videospelutvecklare och utgivare som har spelare över hela världen. Paradox största plattform när det gäller antal spelare och intäk-ter är PC:n. Målet med detta exjobb var att göra en churn-predikterings modell för att förutsäga sannolikheten för att spelare har "churnat" för att veta vilka spelare fokusen ska vara på i retentionskampanjer. Eftersom syftet med churn-prediktering är att minimera förlust på grund av kunderna som "churnar", var fokusen på spelare som spenderar mest pengar (valar) i Paradox PC-spel.För att definiera vilka spelare som är valar undersöktes hur mycket spelarna spenderar under en 12 månaders rullande period (från 2016-01-01 till 2018-12-31). Spelarna som spenderade mer än 95:e percentilen av den totala spenderingen för varje period definierades som valar. För att definiera när en val har "churnat", det vill säga slutat vara en kund som spenderar mycket pengar i Paradox PC-spel, tittade man på hur många dagar som gått sedan spelarna köpte någonting. En val har "churnat" om han inte har köpt något under de senaste 28 dagarna.När data hade varit samlad om valarna var datan förberedd för ett antal olika maskininlärningsmetoder. Logistic Regression, L1 Regularized Logistic Regression, Decision Tree och Random Forest var de metoder som testades. Random Forest var den metoden som gav bäst resultat med avseende på AUC, med AUC = 0, 7162. Slutsatsen är att det verkar vara möjligt att förutsäga sannolikheten att Paradox valar "churnar". Det kan vara möjligt att förbättra modellen ytterligare genom att undersöka mer data och finjustera definitionen av churn.

APA, Harvard, Vancouver, ISO, and other styles

7

Tandan, Isabelle, and Erika Goteman. "Bank Customer Churn Prediction : A comparison between classification and evaluation methods." Thesis, Uppsala universitet, Statistiska institutionen, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-411918.

Full text

Abstract:

This study aims to assess which supervised statistical learning method; random forest, logistic regression or K-nearest neighbor, that is the best at predicting banks customer churn. Additionally, the study evaluates which cross-validation set approach; k-Fold cross-validation or leave-one-out cross-validation that yields the most reliable results. Predicting customer churn has increased in popularity since new technology, regulation and changed demand has led to an increase in competition for banks. Thus, with greater reason, banks acknowledge the importance of maintaining their customer base. The findings of this study are that unrestricted random forest model estimated using k-Fold is to prefer out of performance measurements, computational efficiency and a theoretical point of view. Albeit, k-Fold cross-validation and leave-one-out cross-validation yield similar results, k-Fold cross-validation is to prefer due to computational advantages. For future research, methods that generate models with both good interpretability and high predictability would be beneficial. In order to combine the knowledge of which customers end their engagement as well as understanding why. Moreover, interesting future research would be to analyze at which dataset size leave-one-out cross-validation and k-Fold cross-validation yield the same results.

APA, Harvard, Vancouver, ISO, and other styles

8

Sergue, Marie. "Customer Churn Analysis and Prediction using Machine Learning for a B2B SaaS company." Thesis, KTH, Matematisk statistik, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-269540.

Full text

Abstract:

This past decade, the majority of services have been digitalized and data more and more available, easy to store and to process in order to understand customers behaviors. In order to be leaders in their proper industries, subscription-based businesses must focus on their Customer Relationship Management and in particular churn management, that is understanding customers cancelling their subscription. In this thesis, churn analysis is performed on real life data from a Software as a Service (SaaS) company selling an advanced cloud-based business phone system, Aircall. This use case has the particularity that the available dataset gathers customers data on a monthly basis and has a very imbalanced distribution of the target: a large majority of customers do not churn. Therefore, several methods are tried in order to diminish the impact of the imbalance while remaining as close as possible to the real world and the temporal framework. These methods include oversampling and undersampling (SMOTE and Tomek's link) and time series cross-validation. Then logistic regression and random forest models are used with an aim to both predict and explain churn.The non-linear method performed better than logistic regression, suggesting the limitation of linear models for our use case. Moreover, mixing oversampling with undersampling gives better performances in terms of precision/recall trade-off. Time series cross-validation also happens to be an efficient method to improve performance of the model. Overall, the resulting model is more useful to explain churn than to predict it. It highlighted some features majorly influencing churn, mostly related to product usage.
Under det senaste decenniet har många tjänster digitaliserats och data blivit mer och mer tillgängliga, enkla att lagra och bearbeta med syftet att förstå kundbeteende. För att kunna vara ledande inom sina branscher måste prenumerationsbaserade företag fokusera på kundrelationshantering och i synnerhet churn management, det vill säga förståelse för hur kunder avbryter sin prenumeration. I denna uppsats utförs kärnanalys på verkliga data från ett SaaS-företag (software as a service) som säljer ett avancerat molnbaserat företagstelefonsystem, Aircall. Denna fallstudie är speciell på så sätt att den tillgängliga datamängden består av månatlig kunddata med en mycket ojämn fördelning: en stor majoritet av kunderna avbryter inte sina prenumerationer. Därför undersöks flera metoder för att minska effekten av denna obalans, samtidigt som de förblir så nära den verkliga världen och den tidsmässiga ramen. Dessa metoder inkluderar översampling och undersampling (SMOTE och Tomeks länk) och korsvalidering av tidsserier. Sedan används logistisk regression och random forests i syfte att både förutsäga och förklara prenumerationsbortfall. Den icke-linjära metoden presterade bättre än logistisk regression, vilket tyder på en begränsning hos linjära modeller i vårt användningsfall. Dessutom ger blandning av översampling med undersampling bättre prestanda när det gäller precision och återkoppling. Korsvalidering av tidsserier är också en effektiv metod för att förbättra modellens prestanda. Sammantaget är den resulterande modellen mer användbar för att förklara bortfall än att förutsäga dessa. Med hjälp av modellen kunde vissa faktorer, främst relaterade till produktanvändning, som påverkar bortfallet identifieras.

APA, Harvard, Vancouver, ISO, and other styles

9

Shan, Min. "Building Customer Churn Prediction Models in Fitness Industry with Machine Learning Methods." Thesis, Umeå universitet, Institutionen för datavetenskap, 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-142515.

Full text

Abstract:

With the rapid growth of digital systems, churn management has become a major focus within customer relationship management in many industries. Ample research has been conducted for churn prediction in different industries with various machine learning methods. This thesis aims to combine feature selection and supervised machine learning methods for defining models of churn prediction and apply them on fitness industry. Forward selection is chosen as feature selection methods. Support Vector Machine, Boosted Decision Tree and Artificial Neural Network are used and compared as learning algorithms. The experiment shows the model trained by Boosted Decision Tree delivers the best result in this project. Moreover, the discussion about the findings in the project are presented.

APA, Harvard, Vancouver, ISO, and other styles

10

Osman, Yasin, and Benjamin Ghaffari. "Customer churn prediction using machine learning : A study in the B2B subscription based service context." Thesis, Blekinge Tekniska Högskola, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-21872.

Full text

Abstract:

The rapid growth of technological infrastructure has changed the way companies do business. Subscription based services are one of the outcomes of the ongoing digitalization, and with more and more products and services to choose from, customer churning has become a major problem and a threat to all firms. We propose a machine learning based churn prediction model for a subscription based service provider, within the domain of financial administration in the business-to-business (B2B) context. The aim of our study is to contribute knowledge within the field of churn prediction. For the proposed model, we compare two ensemble learners, XGBoost and Random Forest, with a single base learner, Naïve Bayes. The study follows the guidelines of the design science methodology, where we used the machine learning process to iteratively build and evaluate the generated model, using the metrics, accuracy, precision, recall, and F1- score. The data has been collected from a subscription-based service provider, within the financial administration sector. Since the used dataset is imbalanced with a majority of non- churners, we evaluated three different sampling methods, that is, SMOTE, SMOTEENN and RandomUnderSampler, in order to balance the dataset. From the results of our study, we conclude that machine learning is a useful approach for prediction of customer churning. In addition, our results show that ensemble learners perform better than single base learners and that a balanced training dataset is expected to improve the performance of the classifiers.

APA, Harvard, Vancouver, ISO, and other styles

11

Granov, Anida. "Customer loyalty, return and churn prediction through machine learning methods : for a Swedish fashion and e-commerce company." Thesis, Umeå universitet, Institutionen för matematik och matematisk statistik, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-184709.

Full text

Abstract:

The analysis of gaining, retaining and maintaining customer trust is a highly topical issue in the e-commerce industry to mitigate the challenges of increased competition and volatile customer relationships as an effect of the increasing use of the internet to purchase goods. This study is conducted at the Swedish online fashion retailer NA-KD with the aim of gaining better insight into customer behavior that determines purchases, returns and churn. Therefore, the objectives for this study are to identify the group of loyal customers as well as construct models to predict customer loyalty, frequent returns and customer churn. Two separate approaches are used for solving the problem where a clustering model is constructed to divide the data into different customer segments that can explain customer behaviour. Then a classification model is constructed to classify the customers into the classes of churners, returners and loyal customers based on the exploratory data analysis and previous insights and knowledge from the company. By using the unsupervised machine learning method K-prototypes clustering for mixed data, six clusters are identified and defined as churned, potential, loyal customers and Brand Champions, indecisive shoppers, and high-risky churners. The supervised classification method of bias reduced binary Logistic Regression is used to classify customers into the classes of loyal customers, customers of frequent returns and churners. The final models had an accuracy of 0.68, 0.75 and 0.98 for the three separate binary classification models classifying Churners, Returners and Loyalists respectively.

APA, Harvard, Vancouver, ISO, and other styles

12

Magnani, Martina. "Progettazione e implementazione di un sistema di predizione dell'abbandono degli utenti in una catena di palestre." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2020. http://amslaurea.unibo.it/22144/.

Full text

Abstract:

L'abbandono del cliente, ossia il customer churn, si riferisce a quando un cliente cessa il suo rapporto con l'azienda. In genere, le aziende considerano un cliente come perso quando un determinato periodo di tempo è trascorso dall'ultima interazione del cliente con i servizi dell'azienda. La riduzione del tasso di abbandono è quindi un obiettivo di business chiave per ogni attività. Per riuscire a trattenere i clienti che stanno per abbandonare l'azienda, è necessario: prevedere in anticipo quali clienti abbandoneranno; sapere quali azioni di marketing avranno maggiore impatto sulla fidelizzazione di ogni particolare cliente. L'obiettivo della tesi è lo studio e l'implementazione di un sistema di previsione dell'abbandono dei clienti in una catena di palestre: il sistema è realizzato per conto di Technogym, azienda leader nel mercato del fitness. Technogym offre già un servizio di previsione del rischio di abbandono basato su regole statiche. Tale servizio offre risultati accettabili ma è un sistema che non si adatta automaticamente al variare delle caratteristiche dei clienti nel tempo. Con questa tesi si sono sfruttate le potenzialità offerte dalle tecnologie di apprendimento automatico, per cercare di far fronte ai limiti del sistema storicamente utilizzato dall'azienda. Il lavoro di tesi ha previsto tre macro-fasi: la prima fase è la comprensione e l'analisi del sistema storico, con lo scopo di capire la struttura dei dati, di migliorarne la qualità e di approfondirne tramite analisi statistiche il contenuto informativo in relazione alle features definite dagli algoritmi di apprendimento automatico. La seconda fase ha previsto lo studio, la definizione e la realizzazione di due modelli di ML basati sulle stesse features ma utilizzando due tecnologie differenti: Random Forest Classifier e il servizio AutoML Tables di Google. La terza fase si è concentrata su una valutazione comparativa delle performance dei modelli di ML rispetto al sistema storico.

APA, Harvard, Vancouver, ISO, and other styles

13

Buö, David, and Magnus Kjellander. "Predicting Customer Churn at a Swedish CRM-system Company." Thesis, Linköpings universitet, Databas och informationsteknik, 2014. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-107904.

Full text

Abstract:

This master thesis investigates if customer churn can be predicted at the Swedish CRM-system provider Lundalogik. Churn occurs when a customer leaves a company and is a relevant issue since it is cheaper to keep an existing customer than finding a new one. If churn can be predicted, the company can target their resources to those customers and hopefully keep them. Finding the customers likely to churn is done through mining Lundalogik's customer database to find patterns that results in churn. Customer attributes considered relevant for the analysis are collected and prepared for mining. In addition, new attributes are created from information in the database and added to the analysis. The data mining was performed with Microsoft SQL Server Data Tools in iterations, where the data was prepared differently in each iteration. The major conclusion from this thesis is that churn can be predicted at Lundalogik. The mining resulted in new insights regarding churn but also confirmed some of Lundalogik's existing theories regarding churn. There are many factors that needs to be taken into consideration when evaluating the results and which preparation gives the best results. To further improve the prediction there are some final recommendations, i.e. including invoice data, to Lundalogik of what can be done.

APA, Harvard, Vancouver, ISO, and other styles

14

Grönros, Lovisa, and Ida Janér. "Predicting Customer Churn Rate in the iGaming Industry using Supervised Machine Learning." Thesis, KTH, Matematisk statistik, 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-228609.

Full text

Abstract:

Mr Green is one of the leading online game providers in the European market. Their mission is to o˙er entertainment and a superior user experience to their customers. To be able to better understand each individual customer and the entire customer life cycle the concept of churn rate is essential, which is also an important input value when calculating the return on marketing e˙orts. This thesis analyzes the feasibility to use 24 hours of initial data on player characteristics and behaviour to predict the probability of each customer churning or not. This is done by examining various supervised machine learning models to determine which model best captures the customer behaviour. The evaluated models are logistic regression, random forest and linear discriminant analysis, as well as two ensemble methods using stacking and voting classifiers. The main finding is that the best accuracy is obtained using a voting ensemble method with the three base models logistic regression, random forest and linear discriminant analysis weighted as w = (0.005, 0.80, 0.015). With this model the attained accuracy is 75.94 %.
Mr Green är en av de ledande onlinespelsleverantörerna på den europeiska mark-naden. Deras mission är att erbjuda underhållning och en överlägsen användarup-plevelse till sina kunder. För att bättre kunna förstå sina kunder och deras livs-cykel är kundbortfall ett ytterst viktigt koncept. Det är också ett viktigt mått för att kunna utvärdera resultaten av marknadsföring. Denna rapport analyserar möjligheten att, med 24 timmars data över kundbeteende, kunna avgöra vilka kun-der som kommer att lämna siten. Detta görs genom att undersöka olika modeller inom övervakad maskininlärning för att avgöra vilken som bäst fångar kundernas be-teende. Modellerna som undersöks är logistisk regression, random forest och en linjär diskriminantanalys, samt två olika sammansättningsmodeller som använder sig av stacking och voting. Resultatet av denna studie är att en sammansättningsmodell som väger modellerna logistisk regression, random forest och en linjär diskriminan-tanalys ger den högsta förklaringsgraden på 75.94 %.

APA, Harvard, Vancouver, ISO, and other styles

15

Barr, Kajsa, and Hampus Pettersson. "Predicting and Explaining Customer Churn for an Audio/e-book Subscription Service using Statistical Analysis and Machine Learning." Thesis, KTH, Matematisk statistik, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-252723.

Full text

Abstract:

The current technology shift has contributed to increased consumption of media and entertainment through various mobile devices, and especially through subscription based services. Storytel is a company offering a subscription based streaming service for audio and e-books, and has grown rapidly in the last couple of years. However, when operating in a competitive market, it is of great importance to understand the behavior and demands of the customer base. It has been shown that it is more profitable to retain existing customers than to acquire new ones, which is why a large focus should be directed towards preventing customers from leaving the service, that is preventing customer churn. One way to cope with this problem is by applying statistical analysis and machine learning in order to identify patterns and customer behavior in data. In this thesis, the models logistic regression and random forest are used with an aim to both predict and explain churn in early stages of a customer's subscription. The models are tested together with the feature selection methods Elastic Net, RFE and PCA, as well as with the oversampling method SMOTE. One main finding is that the best predictive model is obtained by using random forest together with RFE, producing a prediction score of 0.2427 and a recall score of 0.7699. The other main finding is that the explanatory model is given by logistic regression together with Elastic Net, where significant regression coefficient estimates can be used to explain patterns associated with churn and give useful findings from a business perspective.
Det pågående teknologiskiftet har bidragit till en ökad konsumtion av digital media och underhållning via olika typer av mobila enheter, t.ex. smarttelefoner. Storytel är ett företag som erbjuder en prenumerationstjänst för ljud- och e-böcker och har haft en kraftig tillväxt de senaste åren. När företag befinner sig i en konkurrensutsatt marknad är det av stor vikt att förstå sig på kunders beteende samt vilka krav och önskemål kunder har på tjänsten. Det har nämligen visat sig vara mer lönsamt att behålla existerande kunder i tjänsten än hela tiden värva nya, och det är därför viktigt att se till att en befintlig kund inte avslutar sin prenumeration. Ett sätt att hantera detta är genom att använda statistisk analys och maskininlärningsmetoder för att identifiera mönster och beteenden i data. I denna uppsats används både logistisk regression och random forest med syfte att både prediktera och förklara uppsägning av tjänsten i ett tidigt stadie av en kunds prenumeration. Modellerna testas tillsammans med variabelselektionsmetoderna Elastic Net, RFE och PCA, samt tillsammans med översamplingsmetoden SMOTE. Resultatet blev att random forest tillsammans med RFE bäst predikterade uppsägning av tjänsten med 0.2427 i måttet precision och 0.7699 i måttet recall. Ett annat viktigt resultat är att den förklarande modellen ges av logistisk regression tillsammans med Elastic Net, där signifikanta estimat av regressionskoefficienterna ökar förklaringsgraden för beteenden och mönster relaterade till kunders uppsägning av tjänsten. Därmed ges användbara insikter ur ett företagsperspektiv.

APA, Harvard, Vancouver, ISO, and other styles

16

Chen, Po-Yu, and 陳柏宇. "Customer Churn Prediction in Virtual Worlds." Thesis, 2013. http://ndltd.ncl.edu.tw/handle/27497138210438784042.

Full text

Abstract:

碩士
國立交通大學
資訊管理研究所
102
With the rapid development of internet websites, more and more online games are produced. Virtual Worlds (VWs) are getting more attention because of the booming trend of on-line games. The highly growth market of VWs attract many companies to join the contest. But the fierce competitions result in a high customer turnover and shortage of profit. Moreover, the unsatisfied customer may spread negative word-of-mouth effect to the company. Therefore, how to predict the churner in the virtual worlds and satisfy them has becoming an important issue. Even though customers churn prediction has been studying in the telecom, financial and retail industry to reduce customer turnover rate, but has not been applied in the virtual worlds to solve the customers’ turnover problem. The objective of this research is to develop a novel virtual world customer churn prediction method. This study analyzes the relationship between customer churn and three kinds of user behaviors in virtual world. The behaviors include virtual life behaviors, social contact behavior, and social influences of social circle neighbors. Our proposed model use random forest and neural network to classify the customer churn in virtual worlds by the three user behaviors mentioned above. The results shows our propose model considering both user’s activity energy and social circle neighbors’ social influence will have better performance. Also, the result shows the performance of decision tree is better than neural network for customer churn prediction in virtual worlds.

APA, Harvard, Vancouver, ISO, and other styles

17

Weng, Pu-Dong, and 翁樸棟. "Customer Churn Prediction of VoIP Service Industry." Thesis, 2007. http://ndltd.ncl.edu.tw/handle/00133400982319283213.

Full text

Abstract:

碩士
國立臺北大學
企業管理學系碩士在職專班
95
Ever since the deregulation of Taiwan telecommunication voice service in July, 2001, Voice over IP (VoIP) has become popular and competition among telephone carriers has been fierce. Retaining value customers, i.e. to lower customer churn rate, became more important than acquiring new customers in maintaining profitability. Employing the decision tree based technique and a private Taiwanese telephone carrier’s VoIP corporate account data, this study proposes a model of churn prediction and investigates the underlying factors of churners. The main results of this study are as follows. First, the churn model generates a satisfactory rate of predictive correctness of 81.44%. Factors, according to their significance, that contribute to customer churn are length of usage days, cellular phone rate, long distance phone rate, service location, and average successful rate. Second, the length of usage days is negatively correlated with the churn rate and customers become stable once usage surpasses 70.6 days. Third, this study also shows that approximately 31% of churn was caused by the cellular phone rate. This outcome is consistent with the general view that a lower cellular phone rate is one of the key motivations for using VoIP. However, it should be noted that the 69% of churn rate for the less than 70.6 days target class contrasts sharply to the 20% average churn rate for the remaining classes. This implies that the cellular phone rate is not the primary factor that affects the churn rate. Finally, it is also generally considered that quality is important and poor quality might be the main reason for churning. This study, however, shows that the negative effects of short usage days on customer churn outweigh the positive effects of other factors such as quality, etc. For maintaining profitability, the VoIP service provider should improve competitiveness of price and quality of service, enthusiasm of problem solving, and capability of customer-end system integration. One implication of this study is that up-front professional suggestions based on customer needs, rather than over-promising, are the keys to eliminate customers’ dissatisfaction caused by unrealistic expectations on VoIP service. This study has shown that customers with the longest customer usage days have the highest customer royalty (zero churn). Keywords: VoIP, Data Mining, Decision Tree, Churn Prediction, Customer Retention

APA, Harvard, Vancouver, ISO, and other styles

18

Hsieh, Yi-Fan, and 謝逸凡. "The Research of ISP Customer Churn Prediction Model." Thesis, 2004. http://ndltd.ncl.edu.tw/handle/r5w28g.

Full text

Abstract:

碩士
銘傳大學
資訊管理學系碩士在職專班
92
This research uses four constructs with 14 variables to build a customer churn prediction model. The constructs are customer classification, customer satisfaction, customer loyalty, and service quality. At first, samples are collected through eliminating incomplete data by using the data mining technique. Then, the data is sequenced and statistically analyzed for its distribution. Furthermore, a prediction is given based on the resulting determining factor. Finally, the customer churn prediction model is built based on the above determining factor associated with data mining techniques. The prediction provides a goal, which helps the company to adjust their customer keeping policy, which will decrease customer churn rate therefore benefit the company. This research is based on the customer data of an internet service provider (ISP) of Taiwan. The total effective data are 12177 records. A total of 10655 records, or 87.5 % effective data, are assigned as training data. Finally, the decision tree classification technique was applied onto the customer churn prediction model. The result shows that customer loyalty is the determining construct of prediction customer churn rate. In addition, there are 9 major factors, which yield good prediction. The major factors are total used time, customer service record, discounts, application time, purchase amount, disconnection rate, date of most recent purchase, charge rate, and customer classification. Of the nine major factors, two of them are Categorical variables, and the remaining seven are numerical variables. In general, continuous numerical variables yields better prediction, because that ISP, like water supply company and electric company, provides services continuously. Customers behave differently as time changes. Therefore in order to predict a better customer churn rate, continuous numerical data should weigh more in the prediction model once a potential customer turned customer. This research model is based on the analysis of traits of corporate ADSL subscribers. When the model applies on different products or industries, several types of data need to be fine tuned before they are collected, based on characteristics of the product or industry. After the data is collected, it will be filtered based on its predictability. Some variables does not yield precise prediction by themselves. In order to decide whether these variables are deciding factors, they have to be considered collaterately with their distributions. The data, grouped by prediction that it is going to yield, is divided evenly among each group. The more diverse the data lies, the better prediction it yields.

APA, Harvard, Vancouver, ISO, and other styles

19

Chiu, Wan-yu, and 邱婉玉. "Data and dimensionality reduction in customer churn prediction." Thesis, 2009. http://ndltd.ncl.edu.tw/handle/79309103036696425973.

Full text

Abstract:

碩士
國立中正大學
會計與資訊科技研究所
97
Nowadays, the competition faced by enterprises is a global competition, which is not limited to regional or national competition. The ‘product-oriented’ model has been transferred to the ‘customer-oriented’ one. This results in the importance of Customer Relationship Management (CRM). Customer retention is one major problem of CRM. Predicting customer behavior helps companies build customer loyalty and maximize profitability. In CRM it is vital forces and essential to use advanced technology to predict the behavior of customers such as churn, especially in the highly competitive telecommunications industry, many companies have used machine learning techniques to predict customer churn and thus to effectively decrease the churn rate. In literature, churn analysis is usually based on either classification or cluster data mining methods. In this thesis, we propose to use the dimensionality and data reduction methods to find representative attributes and useful data from a given dataset, then train the prediction model by the processed dataset. The purpose is to improve the prediction performance of the customer churn prediction model and testify which one of the data pre-processing method is the best. The prediction model is based on Neural Networks, and the training dataset is different from various data pre-processing methods. The experimental result shows that the best data pre-processing method is to consider both the dimensionality and data reduction processes, but the priority of data reduction must higher than dimensionality reduction in terms of higher prediction accuracy. In addition, this combination provides the lowest Type I and II error rates.

APA, Harvard, Vancouver, ISO, and other styles

20

廖秀玉. "Friend Recommendation and Customer Churn Prediction in Virtual Worlds." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/zq3hep.

Full text

Abstract:

博士
國立交通大學
資訊管理研究所
103
Virtual worlds (VWs) are becoming effective interactive platforms in the fields of education, social sciences and humanities. User communities in virtual worlds tend to have fewer real world linkages and more entertainment-related goals than those in social networks. The above characteristics result in an ineffective modality with respect to applying existing friend recommendation and customer churn prediction methods in virtual worlds. Firstly, this study develops a virtual friend recommendation approach based on user similarity and contact strengths in virtual worlds. Then, it proposes a customer churn prediction method taking users’ monetary cost, activity energy and social neighbor features into considerations. In the proposed friend recommendation approach, users’ contact activities in virtual worlds are characterized into dynamic features and contact types to derive their contact strengths in communication-based, social-based, transaction-based, quest-based and relationship-based contact types. Classification approaches were developed to predict friend relationships based on user similarity and contact strengths among users. A novel friend recommendation approach is further developed herein to recommend friends as regards certain virtual worlds based on friend-classifiers. In the customer churn prediction approach, users are segmented into stable and unstable groups. Users’ consumption behaviors, virtual life and social life activity energy and social neighbors influence are analyzed by user segments. Different classification methods are applied to predict customer churn. The evaluation uses mass data collected from an online virtual world in Taiwan, and validates the effectiveness of the proposed methodology. The experiment results show that the friend classifier and customer churn prediction that take into account contact strengths can elicit stronger prediction performance than the friend-classifier and churn prediction that considers only user similarity or monetary methods in the existing research.

APA, Harvard, Vancouver, ISO, and other styles

21

Chen, Tai-Ling, and 陳岱伶. "Managing Customer Churn using a Churn Prediction Model and a Profit Maximization Retention Strategy." Thesis, 2019. http://ndltd.ncl.edu.tw/cgi-bin/gs32/gsweb.cgi/login?o=dnclcdr&s=id=%22107NCHU5402019%22.&searchmode=basic.

Full text

Abstract:

碩士
國立中興大學
行銷學系所
107
Retaining an existing customer costs much less than acquiring a new customer, and increasing customer retention by 5% can increase profits from 25-85% (Reichheld & Sasser, 1990). Customer churn management is very important for companies, but many studies focus on improving the accuracy of predictive models. A churn prediction models lacks a matching strategy, making it difficult to select target customers according to corporate goals. Therefore, the study adopts the profit maximization retention strategy for customer churn management, taking the telecommunications company as an example. CART, logistic regression and neural network are compared according to their ability to predict churners, and the outperformer is applied to adopt a profit maximization strategy in order to accurately target profitable customers and determine the optimal target customer size. The important influence variables of the model are discussed and verified. The results show that the neural network has the best prediction performance, and the profit maximization strategy has higher expected profit in various situations than the churn probability strategy, which confirms the advantage of the profit maximization retention strategy. This study provides companies with the ability to accurately target profitable customers in retention decisions and anticipate the expected profit from a retention activity.

APA, Harvard, Vancouver, ISO, and other styles

22

HSU, CHUAN-YAO, and 徐雋堯. "Constructing Customer Churn Prediction Model by Entity Embedded Vector Method." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/e2q84t.

Full text

Abstract:

碩士
東吳大學
巨量資料管理學院碩士學位學程
107
With the advancement of technology, the telecommunication companies face more competitive environment in the marketing. It is an important issue for the industry to maintain a good revenue and members. If the customers can effectively stay within the scope of business operations, they will not lose or transfer services to other operators. The enterprises can obtain a good competitive advantage. Therefore, building a predictive model for customer churn is a key role for marketing in the business model. In the past, the categorical variables in the data were usually converted to the numerical data with one-hot encoding method. However, this method may generate a large number of sparse matrices. Therefore, we applied entity embedding to generate vector space in response to the shortcomings of previous one-hot encoding method, and also simultaneously compared the performance between those two coding methods with various machine learning models. The results show that the performance of entity embedding is higher than those of one-hot encoding method with five models including random forest, SVM, Bayesian, KNN and neural network. The important features for customer churn are "day calls", "service calls", "eve charge" and "intl charges". We suggested that entity embedding can generate the proper vector space instead of the sparse metrics and also improve the performance of the machine learning model. Finally, the results that we found can help the telecommunication companies detect the dynamic behavior of the customers and ensure customer retention.

APA, Harvard, Vancouver, ISO, and other styles

23

Díaz, Méndez German Augusto. "Data mining guided process for churn prediction in retail: from descriptive to predictive analytics." Master's thesis, 2021. http://hdl.handle.net/10362/112895.

Full text

Abstract:

Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Information Systems and Technologies Management
In recent years, the development of new technologies has permeated all industries, and with its rapid introduction, technology has brought the need to solve uncertainty in processes. The need to understand and collect data by companies has become a central paradigm, but the journey continues in the efforts to transform it into powerful insight into new processes, goods, and services. In the grocery retail industry has been essential to understanding the need to include academic research to understand different commercial purposes (Perloff & Denbaly, 2007). It has become an essential issue to understand the data coming from all the sources in the industries, allowing to focus the efforts to reduce the gap between the vertical and horizontal relationships and from the different stakeholders in the supply chain. That is why it became relevant to understand the customer experience along the supply chain and maximized by the marketing chain. The complexity of the transactions and the crescent number of customers define challenges for the grocery retail stores to process and provide a high-quality service based on data to their customers. The key to gaining competitive advantage is to understand, classify, and prevent customer churn to maximize profit. It is used to attract and retain new customers with data-driven decisions. For this, it is necessary to understand and label the customers as churners. The organizations tend to focus more on developing plans to deal with the Customers, using CRM (Customer Relationship Management) as the core strategy to handle, maintain and build new long-lasting relationships with the customer as a critical stakeholder (Chorianopoulos, 2015). Data mining techniques help CRM to achieve their goals building tools that lead to informed decisions, creating better, stronger and long-lasting relationships thanks to the analysis of the customer-organization interaction and application of complex models.

APA, Harvard, Vancouver, ISO, and other styles

24

Chen, Mao-Yuan, and 陳茂元. "Applying Data Mining to Multimedia on Demand (MOD) Customer Churn Prediction." Thesis, 2007. http://ndltd.ncl.edu.tw/handle/72100691987278444844.

Full text

Abstract:

碩士
國立中正大學
會計所
95
Because of globalization, loosening of laws, and decreasing constraints of investment, our telecommunication market opens up worldwide. As a result, fixed network industry faces up unprecedented intense competition. And the product-oriented mode turns into customer-oriented mode. As far as telecommunications industry is concerned, customers’ demands must be considered as priorities. The latest product “Triple play”, which combines voice, data, and video, is the most-wanted service to all the competitors. Multimedia on Demand (MOD) has become available in the market for more than two years. During the product inducts the time to enter the growing stage, owing to the fact that customers are hard to accept the new product and some people may withdraw the service after installed, it is necessary to prevent customers from churning to the steady growth of market share. In literature, churn analysis is usually based on either classification or cluster data mining methods. In this thesis, we propose to use the association rule method to find the churn attributes and rules. Then, important variables are selected by the rules. Next, both decision trees and neural networks are developed for comparisons. The training dataset composed of fifteen months information, which is divided into eight different sizes of training and validation sets for model construction. The predicting dataset is based on the later four months information to find the best customer churn prediction model. The experimental results show that association rule to select useful variables and decision trees to build the model provide the accuracy rate of 90%. Meanwhile, the model can help us find the useful information and offer the decision makers some related knowledge so that they can work out marketing strategies.

APA, Harvard, Vancouver, ISO, and other styles

25

Wu, Chia-An, and 吳佳安. "The Study of Customer Churn Prediction Model in Hot Spring Hotels." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/97257365615471608388.

Full text

Abstract:

碩士
輔仁大學
科技管理學程碩士在職專班
103
The main purposes of this study were to understand the customer constitutive structure of a case company, to build the customer churn prediction model of a case company by using data mining classification technologies (discriminant analysis, logistic regression, artificial neural networks), and to find out the significant characteristics of churn customer. The customer database which contained 34,786 records was provided by a case company in order to perform the empirical research. After merging the repeated accommodation records from the same customer, the database had a total of 6,711 records. The results were as followed: 1. In this case company, the customer constitutive structure was that “Taiwanese customers” accounted the most (97.56%), the “male customers” accounted for about seventy percent (74.53%), the average “age” of customers was 45.19-year-old, the customers who were not living in Taichung City accounted for 86.77%, the average “total amount of consumption” of customers was NT$26,654.59, the average “total number of accommodations” of customers was 2.73 times, the average “total number of nights” of customers was 5.18 nights, and the “churn customers” accounted for 35.08%. 2. The main purpose of the constructional process of customer churn prediction model which this study proposed was using three classification methods to obtain the one best discriminating model. Beside, in order to verify the effectivity of the discriminating model, this study used the customer database provided by a case company to perform the empirical research. The result showed that the whole correct classification rate of artificial neural networks was the highest in the three analytical tools. Therefore, this study suggested that artificial neural networks was a worth tool to use. 3. Overall, the characteristics of churn customer through the customer churn perdiction model using artificial neural networks were the customers of the age between 51-year-old and 60-year-old, and the total amount of consumption between NT$5,000 and NT$10,000.

APA, Harvard, Vancouver, ISO, and other styles

26

Santos, Hugo Filipe Paulino dos. "Social network embeddings for churn prediction." Master's thesis, 2020. http://hdl.handle.net/10071/22051.

Full text

Abstract:

With the large adoption of Internet customers became more aware of existing services and their prices. From the perspective of companies acquiring a new customer is more expensive than maintaining existing ones. In this sense, companies began to address the challenge of leaving customers to other companies. Customer churn is even more challenging in the telecommunications sector, because customers can change operator faster due to shorter loyalty period and easy migration service to other telecommunications operators without associated costs. Anticipating churn is therefore a major concern for telecommunication companies, which leads them to carry out retention campaigns for these customers. Predictive models allows us to predict whether a customer will leave their operator using that client’s past information. The present work describes how a predictive model was build to predict the outflow of customers exploring customer relationships. Unlike other works, it uses a social network analysis that takes advantage of small customer representations (network embeddings) and allows to obtain better results than other methods.
Com a generalização da Internet os clientes tornaram-se mais informados dos serviços existentes e dos seus preços. Na perspetiva das empresas, adquirir um novo cliente é mais dispendioso que manter os existentes. Nesse sentido as empresas começaram a abordar o desafio da saída de clientes para outras companhias. A saída de clientes é ainda mais desafiante no setor das telecomunicações, porque os clientes podem mudar de operador com maior rapidez devido ao período de fidelização mais curto e à fácil migração do serviço para outros operadores de telecomunicações sem custos associados. Antecipar a saída é, portanto, uma grande preocupação para as empresas de telecomunicações, que as leva a realizar campanhas de retenção para esses clientes. Modelos preditivos permitem prever se um cliente vai abandonar a sua operadora atual usando informação passada desse cliente. O presente trabalho detalha como foi construído um modelo preditivo para prever a saída de clientes explorando relacionamentos entre clientes. Ao contrário de outros trabalhos, este utiliza uma análise de rede social que tira partido de representações de baixa dimensionalidade dos clientes (network embeddings) e permite obter melhores resultados que outros métodos.

APA, Harvard, Vancouver, ISO, and other styles

27

Hsieh, Yi-Cheng, and 謝易澄. "Integrating Classification and Clustering Techniques for Customer Churn Prediction in Logistics Industry." Thesis, 2013. http://ndltd.ncl.edu.tw/handle/hgm73r.

Full text

Abstract:

碩士
國立中正大學
資訊管理學系暨研究所
101
Customer value and churn has always been the most important issue of concern to enterprise, in the logistics industry as well. In recent years, with the rise of the Internet, the rise of consumer awareness, coupled with horizontal competition in the market, making the life cycle of the customer becomes more short-term than in the past. How to establish long-term cooperative relationship with valuable customers, is the key to stability in today's competitive market. For businesses, it is important to establish a stable partnership with customers, but the cost is expensive. When the corporate resources are limited, high-value customer retention will get higher benefits. In the first part of this study, we perform data extraction to historical transaction data. By extended customer value analysis model and customer satisfaction index for logistics, construct a new and appropriate research variables to assess the customer value and churn. In the second part, using data mining clustering techniques to separate the high and low value customers by the customer value research variables. In the third part, using data mining classification techniques to performed churn prediction analysis for high-value customers. Finally, discover the key factors of customer churn to provide business decision-makers to develop a marketing strategy.

APA, Harvard, Vancouver, ISO, and other styles

28

CHEN, E.-LING, and 陳依伶. "The Study on Customer Churn Prediction Model for a Direct Selling Company." Thesis, 2016. http://ndltd.ncl.edu.tw/handle/ys5g35.

Full text

Abstract:

碩士
輔仁大學
統計資訊學系應用統計碩士在職專班
104
Loss of customers is one of the most potential issues in many companies; unfortunately most of those companies are not aware of which, when or why they lose their customers, nor be aware of what kind of the impact would have to their income and profit. However, instead of making all the effort to attract new customers, in many successful companies one of the most important factors to success is to keep loyal customers. In fact, to prevent a company from losing its customers in early phase would be helpful in raising its revenue. Discovered from this study, the main causes of losing customers include customer behavior, customer gender and their group. In addition, this study separate two models by the number of times each individual customer purchase product E, as a result, helps to increase the accuracy of the non-loss customers about 6.53%, also enhance over-all accuracy by 3.63%. Therefore, the decision tree and logistic regression mixed model in this study may not only to investigate the effects of variables related to the loss of customers, but also provide a reference for companies to respond to and prevent from loss of customers, and then to create the maximum profit of company.

APA, Harvard, Vancouver, ISO, and other styles

29

Chen, Huei-Ting, and 陳蕙婷. "Applying Decision Tree for Customer Churn Prediction Model of Telecom Prepaid Products." Thesis, 2017. http://ndltd.ncl.edu.tw/handle/umb943.

Full text

Abstract:

碩士
國立臺北科技大學
工業工程與管理系EMBA班
105
Since Taiwan government privatized telecommunication service industry in 1997, the competition were severely among mobile companies in Taiwan. There are 5 mobile companies running 2G, 3G, 4G system in Taiwan now. With increasingly popular smart devices (smart phones, tablets, etc.), the mobile service needs continue growth and the market has become more competitive than that in before. There are two types of products in mobile services： “Postpaid product” and “Prepaid product”. “Postpaid product” means that customers subscribe mobile services and pay by monthly bill. “Prepaid product (prepaid card)” means that customers prepay a certain amount for mobile voice or internet services and need to recharge after the credit run out. Because companies usually ask mobile phone users sign on at least one year contract for postpaid customers, prepaid card has faced more churn risk in comparing with postpaid products. In addition to attract new customers, reducing customer churn should be a better strategy in today’s competitive markets. This study examines the key factors of customer churn and builds the churn prediction model to help mobile companies to identify the possible threat and take action to anti-churn. The results show that the change of customer answering call behavior and the change of off-net percentage and are highly related to customer churn. The obtained key factors of churn are suggested to merge in the early warning report or system to help telecom company anti-churn and customer retention.

APA, Harvard, Vancouver, ISO, and other styles

30

Tsai, Wan-Ching, and 蔡宛靜. "The Application of Data Mining in Customer Churn Prediction in Mobile Telecommunication Industry." Thesis, 2013. http://ndltd.ncl.edu.tw/handle/88497175019095702935.

Full text

Abstract:

碩士
國防大學理工學院
資訊工程碩士班
101
Telecommunications market in Taiwan has become seriously challenging since 1997 when the industry was fully liberalized. More and more telecommunication companies join the competition to grab the market share, as well as the strong demand from various kinds of target consumers for wireless communications services makes the business become flourishing. With such a fiercely saturated market and keen competition, a serious loss of customers (Churn) problem is the common difficulty in the industry. In order to maintain the market share, all the companies push themselves as far as they can go not only to attract consumers via providing numerous promotional campaigns but also to utilize new analytically predictive tools to maintain existing customers. Since the cost of maintaining old customers is much lower than that of recruiting new customers, telecommunication operators presently are eager to resolve the issues of losing customers promptly in order to consolidate their business. This study applies the datasets which are also used for Knowledge Discovery and Data mining (KDD) CUP 2009 contest and provided by a well-known French operator— the Orange Company. Using data mining method and the Weka machine learning programs, we propose two algorithms to established customer churn prediction models: 1) Outlier Discarding Algorithm(OD) filtering outlier data in order to obtain a representative training data; 2) High Missing rate Feature Discarding Algorithm(HMFD) reducing the impact of irrelevant features to enhance the accuracy of data mining classification. In accordance with the experimental results, it is proved that through applying the data mining methods and the two algorithms mentioned above on the KDD CUP 2009 datasets, we can obtain better Area under Receiver Operating Characteristic (AUC, also known as ROC Area)performance, furthermore, by establishing customer churn prediction model for telecommunication industry in advance can do great help to set up corresponding marketing measures (Marketing Strategy) in order to reduce the effect of the loss of customers.

APA, Harvard, Vancouver, ISO, and other styles

31

柳靜慧. "A Study of The Prediction Model for Customer churn-An Example of Telecommunication Industry." Thesis, 2003. http://ndltd.ncl.edu.tw/handle/90283399414789783268.

Full text

Abstract:

碩士
輔仁大學
資訊管理學系
91
To a telecommunication company, high quality service and low operating cost are the key to a successful telecommunication business. The primary issue a product or service provider in a market functioning by the principle of "customer first” (or "the customer is God") has to address is to meet the customers’ needs. Customers have various needs, namely quality, price, after sale service, etc. Service quality directly determines how customers are satisfied with the service and thus their decision to purchase the service/product. Hence, enhancement of service quality is a necessary condition for increasing customers. On the other hand, reduction in operating cost is a requisite for a telecommunication dealer to make any profits. Although the search for new customers is a top priority, the issue of customer churn is even more worthy of attention. This Study involves discovering customer churn by means of data mining, exploring and discovering data by means of clustering, K-means and Self-organizing Map (SOM) and sorting out models embodied in voluminous data but hardly detectable. In this Study, personal data about telecommunication customers and contents of the contracts they entered into with the telecommunication company are regarded as important parameters. With clustering, customers are divided into different customer blocks, minimizing the discrepancies found within the same group but maximizing the discrepancies among different groups. The researcher identifies the important factors in customer churn and puts forth feasible strategies for coping with customer churn, with a view to uncovering whatever unidentified customer churn and eventually achieving the management goals. This Study of the prediction model for customer churn is not only to discover customer churn, but it is also designed to keep the customers. In other words, the ultimate goal of this Study is, given the findings, to keep the promotion objects, create the prediction model for customer churn and identify any latent customer churn efficiently, so as to provide reference for future customer relation work.

APA, Harvard, Vancouver, ISO, and other styles

32

Wen, Huang-I., and 黃怡雯. "A research of the prediction model for Customer Churn - by using data mining technology." Thesis, 2008. http://ndltd.ncl.edu.tw/handle/axad5x.

Full text

Abstract:

碩士
銘傳大學
資訊管理學系碩士在職專班
96
From 2006 ,the credit card market in Taiwan has shown a decreasing tendency ,so does the amount of spending, cash advance and revolving . It is because of the cost of developing new customers is five times higher than keeping the old ones. Therefore, it is important to a customer churn prediction model to retent profItable customers and increase the benefits for the enterprise. The sample data of this research is the total churn customers in 2007 of a banking institute. The RFM model is applied with four main contributors, which is characteristics of customer, quality of service, satisfaction of customers , loyalty of customers and JCIC, to find out the reasons of customer churn and build a prediction model. We found that the customer churn related to sex,marriage,education,customer and bank intercourse period, yearly income, FICO,Non-profit organization porxy pay, own other bank cards, activity, spending, frequency and benefit point. The main effect variables were two categories variables and seven range variables including : own other bank cards, FICO, point, spending, frequency, activity rate, customer and bank intercourse period. The hit rate of the predication model is 83.26%. We also found that the FICO is an important variable to prediction customer churn, however, this data is difficult to collect , If the sampling data is not enough which would make fall prediction. We propose to combine the customer expense behavior variables and segment characteristic to increase the prediction accuracy.

APA, Harvard, Vancouver, ISO, and other styles

33

Kandel, Ibrahem Hamdy Abdelhamid. "A comparative study of tree-based models for churn prediction : a case study in the telecommunication sector." Master's thesis, 2019. http://hdl.handle.net/10362/60302.

Full text

Abstract:

Dissertation presented as the partial requirement for obtaining a Master's degree in Statistics and Information Management, specialization in Marketing Research e CRM
In the recent years the topic of customer churn gains an increasing importance, which is the phenomena of the customers abandoning the company to another in the future. Customer churn plays an important role especially in the more saturated industries like telecommunication industry. Since the existing customers are very valuable and the acquisition cost of new customers is very high nowadays. The companies want to know which of their customers and when are they going to churn to another provider, so that measures can be taken to retain the customers who are at risk of churning. Such measures could be in the form of incentives to the churners, but the downside is the wrong classification of a churners will cost the company a lot, especially when incentives are given to some non-churner customers. The common challenge to predict customer churn will be how to pre-process the data and which algorithm to choose, especially when the dataset is heterogeneous which is very common for telecommunication companies’ datasets. The presented thesis aims at predicting customer churn for telecommunication sector using different decision tree algorithms and its ensemble models.

APA, Harvard, Vancouver, ISO, and other styles

34

Fu, Yung-Lin, and 伏泳霖. "Prediction of Customer Churn based on Social Behavior Analysis and Topic Model with Big Data." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/z3cqwe.

Full text

Abstract:

碩士
中原大學
資訊管理研究所
106
Because the convenience of the network, so that customers in the online shopping conversion costs are lower. For enterprises, how to use limited resources to prediction of customer churn and customer retention, is a very important issue. In the past research, customer churn was mainly predicted by customer value. Today''s online shopping platform provides a platform for customers to write reviews and social. Therefore, the past customer churn prediction method, because the different data characteristics, has gradually become not applicable. Customers share and exchange information about purchase, including product reviews, ratings as a consideration for purchasing decisions and influence the possibility of continued consumption in the future. Such decisions may be influenced by the opinions of their friends who have a relationship. Purchase records and product reviews will be with the accumulation of time, so that enterprises can analyze the amount of data gradually increased. Therefore, this project proposes a prediction of customer churn model based on social behavior analysis and topic model with big data. It considers the social behavior of customers on the Internet and the information implied in the reviews written by the customers. And through the topic model of the building, customers often express the words can be topic classified, and in order to build the customer’s own preferences, and to Hadoop platform for the experimental infrastructure, saving a large amount of data required to calculated the amount of time. The experimental results show that our prediction of customer churn model has better prediction results compared with the other customer churn prediction method, and the execution efficiency on the Hadoop platform is obviously higher than that on the general computer implementation efficiency.

APA, Harvard, Vancouver, ISO, and other styles

35

Ko, Kuang-Zn, and 柯光任. "The Research in the Prediction Model of Customer Churn - Using an example of Credit Card Revolving." Thesis, 2005. http://ndltd.ncl.edu.tw/handle/rtyzc7.

Full text

Abstract:

碩士
銘傳大學
資訊管理學系碩士在職專班
93
The transitions of the social environment changed the consumption value point of people, and promoted the developing of products and services. Service qualities and customer satisfactions affect customers’ consumption sensations. Businesses should understand what customers want and need, and enhance their purchase desire. That means the banks must sense that how the customer chooses a bank? Why the customer changes their behavior of using credit cards? This research’s purpose is to create a customer churn prediction model. The research subject is the credit card customer who uses revolving interest. There are 5 constructs that constitute our framework. We referenced the RMF model along with the data mining technology. Finally we got 15 important variables for this prediction model. The 15 customer behavioral variables were all happened in the late 6 months of their consumption. Such as, money spending growing months, or department store’s promotion during the period. Furthermore, this research used the attribute prediction evaluation model to measure all the variables, and it exactly improve the degree of accuracy. After the model training, we got four findings:1.The customers, who didn’t have many consumption times or didn’t spend lots of money, was not easy to use revolving interest. It means the customer is easier to leave. 2.The customer, who uses lots of revolving interest, is not easier to leave. 3.The customer, who had ever got into debt or past the payment time limit, means he/she has worse economic situation, so he/she is easier to leave too. 4.Long-term customers have better loyalty, so they are not easy to leave. Keyword：Customer Churn Prediction Model, Credit Card Revolving, Data Mining, Decision Tree, Neural Network

APA, Harvard, Vancouver, ISO, and other styles

36

HUANG, PENG-HSIANG, and 黃朋祥. "Building Customer Churn Prediction Model Using Data Mining Technique: A Case study of Lien-Chin Computer Training Center." Thesis, 2013. http://ndltd.ncl.edu.tw/handle/55537294833399833105.

Full text

Abstract:

碩士
致理技術學院
服務業經營管理研究所
101
The competition is originally very sharp in the market of Taiwanese Computer Cram Schools. Moreover, consumers certainly take care of their disposal incomes owing to some reasons, such as birth rate decrease, economic downturn, and the negative growth of GDP. The enterprises of Computer Cram Schools have put their focus on customer retention management, in order to maintain the survival of enterprises. Accordingly, in this thesis, we conduct a case study using the existing data of this case company and the techniques of data mining analysis. We aim to find the economic combination of factors, which can effectively forecast the leave of customers in previous. Three tools of data mining analysis were applied to this case study. They include decision trees, association rules, and artificial neural networks. We found the two factors of “age” and “fee paid” can definitely predict the churn of customers and the forecasting results of all models were up to 96%. In conclusion, the enterprises of Taiwanese Computer Cram Schools can utilize the characteristics of consumers’ age and fee paid to further develop the ex ante churn-prevention and ex post remedy strategy.

APA, Harvard, Vancouver, ISO, and other styles

37

Chen, Ying Feng, and 陳盈峰. "A Study of the Prediction Model for Customer Churn Based on Genetic Programming-An Example of Security Service Industry." Thesis, 2007. http://ndltd.ncl.edu.tw/handle/18232310291513985060.

Full text

Abstract:

碩士
輔仁大學
資訊管理學系
95
Security service industry have developed in Taiwan about 30 years, under the firece market competition day by day, how to protect customer to become an important subject. To customer’s loss, the characteristic of group that can not grasp customers to engage in commonly used statistical analysis effectively to security service industry. So, set up the prediction model that this research uses the genetic programming that can solve the complicated problem or variety, expect knowing customer's loss factor and customer outline , is regarded as and made because of the reference in conformity with the countermeasure, reduce the customer's running off rate. 　　This six characteristics attribute with the customer of security service industry of research is done for the setting-up foundation of the prediction model, the experimental result can increase the loss customer's grasping rate by 30%. Is it find to research, “serviced month” and “area” for important factor of running off, and limiting conditions on “area” can district separate appear of different nature customer group. The model quotes RFM concept, can show high-quality customer group suddenly. After the model joins the mechanism of strengthening, reduce necessary evolution generation while solving of asking, it can accord with the logic rule too to answer, the decision rule can possess the explaining too, evolving efficiency wholly can be promoted by a wide margin. 　　According to the experimental result, the research could be summarized in four aspects: 　　1.The prediction model can find out customer's loss factor. 　　2.As to the thing that the loss customer's ability of grasping, the prediction model is obviously superior to the existing statistical analysis way. 　　3.Use customer's contribution degree weight value instead to guide the direction of evolving, can excavate out the high-quality customer's characteristic outline, increase benefit that the customer keeps by a wide margin . 　　4.Can promote and evolve efficiency, a stability of model and validity by a wide margin after joining the mechanism of strengthening, verify the modeling feasibility of the prediction model.

APA, Harvard, Vancouver, ISO, and other styles

38

Ming-hui, Wu, and 吳明輝. "A Study of Gene Expression Programming on the Prediction Model for Customer Churn – An Example of A Telecommunication Company." Thesis, 2010. http://ndltd.ncl.edu.tw/handle/59612877702360189337.

Full text

Abstract:

碩士
輔仁大學
資訊管理學系
98
In the highly competitive mobile communications market, the various telecom companies to develop new customers high cost, the fight to the customers and high turnover rate. Therefore, if the telecom companies to invest their effort in maintaining and obtain the trust of existing customers, understand customer needs and provide customers with the best of products and services, in order to avoid the loss of existing customers, its cost, economy and profitability of enterprises benefits will contribute to higher than other industries. This has sparked research on the telecommunications company of the customer churn prediction model for high level of interest. This study aimed to explore the feasibility and performance of combination of GEP (Gene Expression Programming) algorithm and RFM (Recency, Frenquency, Monetary) model that applied to the prediction model for customer churn. The results of the first Model (M1) in this study show that, the three indicators of RFM model, a significant implicit in customer behavior patterns about the churn of non-contracted customer, however, the second model (M2) results show, for contracted customer, M index the least significant. The Specificity, one performance indicator of the model, indicators of M1 and M2 were both controlled in 60 percent index level, the other performance indicator, Sensitivity, of M1 model has been caused by better performance (up to the test group, the average performance over 78.06%), and although the rate of churn of M2 model is only 5.55% of the sample, obviously is not conducive to forecast, although its average performance of the highest test group also reached more than 70% performance. Comprehensive experimental results and analysis, this study found the Gene Expression Programming algorithm used in customer churn prediction model, is technically feasible, but in application, it should be noticed that the influence of the sample rate of churn on the stability of model, large distribution differences of the small sample may reduce the stability of the model.

APA, Harvard, Vancouver, ISO, and other styles

39

Wang, Chih-Chi, and 王志吉. "The Research of applying Data Mining Techniques on the Prediction Model of Customer Churn – Using an example of a certain Security Broker." Thesis, 2012. http://ndltd.ncl.edu.tw/handle/92332163542332954752.

Full text

Abstract:

碩士
銘傳大學
資訊管理學系碩士在職專班
100
The merges and acquisitions (M&A) actions in securities industry has become a tendency for years. It toward to the prospects of enterprise getting bigger gradually. Securities industry offers products and services with high homogeneity. Therefore, the innovative financial products and better quality of customer service have been required to improved sharply on the market to avoid customer churn. This study uses of 4 dimensions, the business cycle on the market, macroeconomics index, service quality, and consumption characteristics of customer, to explore the influence factors that caused customer churn. We used a certain security broker’s customer data as samples, according to the complete procedures of data mining to mine knowledge. First, we eliminating the incomplete data, doing appropriate coding, using the attribute prediction evaluation model to identify the key prediction variables, to explore the critical factors that affects customer churn and provide the results of this research as reference to security brokers for formulating their marketing strategies. The results showed that frequency and monetary; the critical factors that affected the prediction of customer churn are recency, frequency and monetary; the composite index of leading business cycle indicators is the most leading and effective factor to sense the change of business cycle, also has significant impact on the customer churn.

APA, Harvard, Vancouver, ISO, and other styles

40

Chiu, Yi-Ling, and 丘翊伶. "Churn Prediction of High-value Customers in Virtual Worlds." Thesis, 2014. http://ndltd.ncl.edu.tw/handle/73708090772758937273.

Full text

Abstract:

碩士
國立交通大學
資訊管理研究所
102
With the emerging of social network websites, more and more social network online games are produced. Also, the trend of playing online games with friends encourages the flourishing VWs market. Nowadays, users own more selections of VWs games while companies consequently suffer from the problems of high customer turnover rate and low-customer-loyalty. Moreover, according to the Pareto principle, 80% of a company's profits come from 20% of its customers (the high value segment). Losing a high-value customer will naturally be more damaging than the loss of a low-value one. Therefore, building a churn prediction model to facilitate subsequent churn management and customer retention is the best core marketing strategy. In this paper, we put emphasis on modeling a hybrid classification, which takes monetary cost, user behavior and social neighbor features into consideration. Through the perspective of RFM model, we can predict more precisely. Our research applies the dataset from Roomi and several common classification methods to conduct the prediction. The experimental results show that the proposed hybrid model is more thoughtful and well-suited for this problem compare to the traditional way in each classification method.

APA, Harvard, Vancouver, ISO, and other styles

41

Chen, Chih-yi, and 陳志毅. "Predicting customer churn through analysis of credit card data." Thesis, 2012. http://ndltd.ncl.edu.tw/handle/31270804283493109349.

Full text

Abstract:

碩士
東吳大學
資訊管理學系
100
Taiwanese Banks are faced with a tough environment with fierce competition in its domestic credit card market and subdued global market. Those banks are increasingly emphasizing on the importance of retaining customers in order to sustain market shares and remain profitable. The banks utilize marketing campaigns and targeted telemarketing to retain customers and build on brand loyalty. Given the tough environment, limited resources and budget are available; it is crucial for banks to analyze consumer behavior in order to efficiently allocate resources to achieve the optimal outcome. This study has outlined a new model for detecting potential customer churn and provides an early warning indicator for local banks. The model incorporates data sources from credit card account records, customer characteristics, and business relation with built-in time factor of temporal abstraction to organize the data prior to analysis. This allows time sensitivity data to be preserved and be further applied with the association rule to analyze and detect abnormal customer behavior. The research outcome indicates that the system is relatively effective in detecting customer churn early in the life cycle and assists banks to address the issue before further escalation. The process of customer churn always leaves trail of evidence; it is unlikely that high profile customers significantly change their transaction volume and behavior overnight. This study aims to prove that the association rules obtained from the research are indeed applicable to detect customer churn, thus providing basis for further analysis. The initial assumption formulated on the association rules are test against real data for a period of six months to reconcile on the relation of the rules to abnormal customer behavior. Furthermore, the tested rules are summited to experts for further scrutinization, hence establishing the relationship between defined rules and management. The goal of this study is to provide measuring tool for banks to assess the quality of marketing campaign and reestablish the decent business-customer relationship.

APA, Harvard, Vancouver, ISO, and other styles

42

WANG, CHIH-CHIA, and 王至嘉. "The Study of Predicting Customer Churn Based on Social Network Analysis." Thesis, 2017. http://ndltd.ncl.edu.tw/handle/7ndszk.

Full text

Abstract:

碩士
銘傳大學
資訊工程學系碩士班
105
Network grows up, mobile devices spread among the people and social websites like facebook, twitter etc. rise. The information that people can receive a day is much more than before, however the information may be spread by a group of people which close to their social scope so that when people receive the information they will be effect by the group. Therefore, an individual in a social group makes may be suffered an influence from some people whatever they inside social network the individual belong to or outside, when making a decision. In the past, most people made a precision model according to individual information or attributes, but they were not to consider to their social relevant. Human is kind of group live animal, most of their behaviors often affect by their live group. For instance, when people's mobile contract is expired, they need to make a decision to decide whether keep the same mobile operator or change others. Most of kind of such decision will affect from the past experience or their friends. If people’s friend user the other mobile operator, the possibility that people change their mobile operator will rise up. For mobile operator, the cost to keep a client will be much cheaper than to develop a client. In this paper, we will use the concept of social network to construct a churn precision model based on social network analysis. Further, find the key performance indicator that affects a client churn to rise the accuracy of precision of client churn.

APA, Harvard, Vancouver, ISO, and other styles

43

FANG, LUNG-WEI, and 方龍偉. "The Model of Customer Churn PredictionOn Machine Learning–A Case of An Asset ManagementCompany." Thesis, 2017. http://ndltd.ncl.edu.tw/handle/5h8ezx.

Full text

Abstract:

碩士
輔仁大學
資訊管理學系碩士在職專班
105
In the highly competitive wealth management market, the investment company to develop new customers high cost, the fight to the customers and high turnover rate. Therefore, if the investment company to focus on maintaining and obtaining the trust of existing customers, understand customer needs and provide customers with the best products and rewards, in order to avoid the loss of existing customers, its costs, economy and profitability of enterprises benefits willcontribute to higher than other industries. This has sparked research onthe telecommunications company of the customer churn prediction modelfor high level of interest. This study is divided into two topics, one is to define the RFM model of the object in line with the indicators and explore whether the RFM model indicators help to enhance the ability of model prediction The second is comparative gene expression programming method, C4.5 Decision Tree, Random Forest, Support Vector Machine the Correct Rate of Customer Loss Prediction Model and Its Advantages and Disadvantages of Vector Machine Outputand. In experimentalconclusion,the Model 2 is based on the model 1 variables and added RFM variables,the four algorithms of the Accuracy, Precision, FPR and F-Measure are better than the modelone, showing that RFM is one of the factors affecting customer churn.In addition,this study also summarizes the characteristics of each algorithm, if the efficiency of the algorithm to evaluate, you can choose C4.5 Decision tree modeling, if the effect of the algorithm to evaluate, choose Random Forest modeling, The explanatory power of the model can be evaluated by C4.5 or Gene Expression Programming.

APA, Harvard, Vancouver, ISO, and other styles

44

MAO, HUI-WEN, and 毛慧雯. "Apply Data Mining Techniques to a Telecom for VIP and Churn Customers Prediction using Decision Tree." Thesis, 2008. http://ndltd.ncl.edu.tw/handle/55133181218047318075.

Full text

Abstract:

碩士
輔仁大學
資訊工程學系
96
In 2006 years, as the telecom industry more and fiercer. We lost very much revenue. We analyze various viewpoints and take two conclusions: One is that we fall out of more and more original customers. Another is that the customer’s contribution decreases. These two reasons result from that the customers have no truehearted attitude toward telcom industry. The quarter of customer amount run away to other telecom providers or break off contract. CRM (Customer Relationship Management) is very important because the customers understand their request and they know how to choose different product before they make decision. CRM raise the interaction between the customers and our services. More detailed we understand the customers’ requirement, more suitable is our product that we design for specific customers. Then we can promote our product to the vast market. This thesis proposed a customer prediction mechanism. Two core concepts are integrated into our research: prediction and customer relationship. There are many relations between customers and providers. For VIP customers we need to enhance the VIP customers’ interaction and for implicit-loss customers we need to struggle to increase the confidence to our product. Our research can reach the following purposes: 1. Our mechanism can predict if the customer is VIP or implicit-lost. 2. We can know if the customer is excellent or bad quality. Customer attribute can help us to analyze the customer’s behavior. Our research uses C4.5 decision-tree solution to classify the customer rank by analyzing the customers’ attributes and finding some rules then to finding our want to get customers Keywords: Decision Tree、Prediction、VIP、Churn、Data Mining

APA, Harvard, Vancouver, ISO, and other styles

45

Muwawa, Jean Nestor Dahj. "Data mining and predictive analytics application on cellular networks to monitor and optimize quality of service and customer experience." Diss., 2018. http://hdl.handle.net/10500/25875.

Full text

Abstract:

This research study focuses on the application models of Data Mining and Machine Learning covering cellular network traffic, in the objective to arm Mobile Network Operators with full view of performance branches (Services, Device, Subscribers). The purpose is to optimize and minimize the time to detect service and subscriber patterns behaviour. Different data mining techniques and predictive algorithms have been applied on real cellular network datasets to uncover different data usage patterns using specific Key Performance Indicators (KPIs) and Key Quality Indicators (KQI). The following tools will be used to develop the concept: RStudio for Machine Learning and process visualization, Apache Spark, SparkSQL for data and big data processing and clicData for service Visualization. Two use cases have been studied during this research. In the first study, the process of Data and predictive Analytics are fully applied in the field of Telecommunications to efficiently address users’ experience, in the goal of increasing customer loyalty and decreasing churn or customer attrition. Using real cellular network transactions, prediction analytics are used to predict customers who are likely to churn, which can result in revenue loss. Prediction algorithms and models including Classification Tree, Random Forest, Neural Networks and Gradient boosting have been used with an exploratory Data Analysis, determining relationship between predicting variables. The data is segmented in to two, a training set to train the model and a testing set to test the model. The evaluation of the best performing model is based on the prediction accuracy, sensitivity, specificity and the Confusion Matrix on the test set. The second use case analyses Service Quality Management using modern data mining techniques and the advantages of in-memory big data processing with Apache Spark and SparkSQL to save cost on tool investment; thus, a low-cost Service Quality Management model is proposed and analyzed. With increase in Smart phone adoption, access to mobile internet services, applications such as streaming, interactive chats require a certain service level to ensure customer satisfaction. As a result, an SQM framework is developed with Service Quality Index (SQI) and Key Performance Index (KPI). The research concludes with recommendations and future studies around modern technology applications in Telecommunications including Internet of Things (IoT), Cloud and recommender systems.
Cellular networks have evolved and are still evolving, from traditional GSM (Global System for Mobile Communication) Circuit switched which only supported voice services and extremely low data rate, to LTE all Packet networks accommodating high speed data used for various service applications such as video streaming, video conferencing, heavy torrent download; and for say in a near future the roll-out of the Fifth generation (5G) cellular networks, intended to support complex technologies such as IoT (Internet of Things), High Definition video streaming and projected to cater massive amount of data. With high demand on network services and easy access to mobile phones, billions of transactions are performed by subscribers. The transactions appear in the form of SMSs, Handovers, voice calls, web browsing activities, video and audio streaming, heavy downloads and uploads. Nevertheless, the stormy growth in data traffic and the high requirements of new services introduce bigger challenges to Mobile Network Operators (NMOs) in analysing the big data traffic flowing in the network. Therefore, Quality of Service (QoS) and Quality of Experience (QoE) turn in to a challenge. Inefficiency in mining, analysing data and applying predictive intelligence on network traffic can produce high rate of unhappy customers or subscribers, loss on revenue and negative services’ perspective. Researchers and Service Providers are investing in Data mining, Machine Learning and AI (Artificial Intelligence) methods to manage services and experience. This research study focuses on the application models of Data Mining and Machine Learning covering network traffic, in the objective to arm Mobile Network Operators with full view of performance branches (Services, Device, Subscribers). The purpose is to optimize and minimize the time to detect service and subscriber patterns behaviour. Different data mining techniques and predictive algorithms will be applied on cellular network datasets to uncover different data usage patterns using specific Key Performance Indicators (KPIs) and Key Quality Indicators (KQI). The following tools will be used to develop the concept: R-Studio for Machine Learning, Apache Spark, SparkSQL for data processing and clicData for Visualization.
Electrical and Mining Engineering
M. Tech (Electrical Engineering)

APA, Harvard, Vancouver, ISO, and other styles

Dissertations / Theses on the topic 'Customer churn prediction'

Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles