Academic literature on the topic 'Binary Logistic Regression Model'

Incorporating survey weights into likelihood-based analysis is a controversial issue because the sampling weights are not simply equal to the reciprocal of selection probabilities but they are adjusted for various characteristics such as age, race, etc. Some adjustments are based on nonresponses as well. This adjustment is accomplished using a combination of probability calculations. When we build a logistic regression model to predict categorical outcomes with survey data, the sampling weights should be considered if the sampling design does not give each individual an equal chance of being selected in the sample. We rescale these weights to sum to an equivalent sample size because the variance is too small with the original weights. These new weights are called the adjusted weights. The old method is to apply quasi-likelihood maximization to make estimation with the adjusted weights. We develop a new method based on the correct likelihood for logistic regression to include the adjusted weights. In the new method, the adjusted weights are further used to adjust for both covariates and intercepts. We explore the differences and similarities between the quasi-likelihood and the correct likelihood methods. We use both binary logistic regression model and multinomial logistic regression model to estimate parameters and apply the methods to body mass index data from the Third National Health and Nutrition Examination Survey. The results show some similarities and differences between the old and new methods in parameter estimates, standard errors and statistical p-values.

This thesis is a study of the detection of separation among the sample points in binary logistic regression models. We propose a new algorithm for detecting separation and demonstrate empirically that it can be computed fast enough to be used routinely as part of the fitting process for logistic regression models. The parameter estimates of a binary logistic regression model fit using the method of maximum likelihood sometimes do not converge to finite values. This phenomenon (also known as monotone likelihood or infinite parameters) occurs because of a condition among the sample points known as separation. There are two classes of separation. When complete separation is present among the sample points, iterative procedures for maximizing the likelihood tend to break down, when it would be clear that there is a problem with the model. However, when quasicomplete separation is present among the sample points, the iterative procedures for maximizing the likelihood tend to satisfy their convergence criterion before revealing any indication of separation. The new algorithm is based on a linear program with a nonnegative objective function that has a positive optimal value when separation is present among the sample points. We compare several approaches for solving this linear program and find that a method based on determining the feasibility of the dual to this linear program provides a numerically reliable test for separation among the sample points. A simulation study shows that this test can be computed in a similar amount of time as fitting the binary logistic regression model using the method of iteratively reweighted least squares: hence the test is fast enough to be used routinely as part of the fitting procedure. An implementation of our algorithm (as well as the other methods described in this thesis) is available in the R package safeBinaryRegression.

Thesis (Ph. D.) -- University of Maryland, College Park, 2008.<br>Thesis research directed by: Dept. of Measurement, Statistics and Evaluation. Title from t.p. of PDF. Includes bibliographical references. Published by UMI Dissertation Services, Ann Arbor, Mich. Also available in paper.

The dissertation examines advancements in the methods and techniques used in the field of econometrics. These advancements include: (i) a re-examination of the underlying statistical foundations of statistical models with binary dependent variables. (ii) using feed-forward backpropagation artificial neural networks for modeling dichotomous choice processes, and (iii) the estimation of unconditional demand elasticities using the flexible multistage demand system with asymmetric partitions and fixed effects across time. The first paper re-examines the underlying statistical foundations of statistical models with binary dependent variables using the probabilistic reduction approach. This re-examination leads to the development of the Bernoulli Regression Model, a family of statistical models arising from conditional Bernoulli distributions. The paper provides guidelines for specifying and estimating a Bernoulli Regression Model, as well as, methods for generating and simulating conditional binary choice processes. Finally, the Multinomial Regression Model is presented as a direct extension. The second paper empirically compares the out-of-sample predictive capabilities of artificial neural networks to binary logit and probit models. To facilitate this comparison, the statistical foundations of dichotomous choice models and feed-forward backpropagation artificial neural networks (FFBANNs) are re-evaluated. Using contingent valuation survey data, the paper shows that FFBANNs provide an alternative to the binary logit and probit models with linear index functions. Direct comparisons between the models showed that the FFBANNs performed marginally better than the logit and probit models for a number of within-sample and out-of-sample performance measures, but in the majority of cases these differences were not statistically significant. In addition, guidelines for modeling contingent valuation survey data and techniques for estimating median WTP measures using FFBANNs are examined. The third paper estimates a set of unconditional price and expenditure elasticities for 49 different processed food categories using scanner data and the flexible and symmetric translog (FAST) multistage demand system. Due to the use of panel data and the presence of heterogeneity across time, temporal fixed effects were incorporated into the model. Overall, estimated price elasticities are larger, in absolute terms, than previous estimates. The use of disaggregated product groupings, scanner data, and the estimation of unconditional elasticities likely accounts for these differences.<br>Ph. D.

Este projeto de mestrado apresenta uma caracterização das chuvas estimadas pelo radar meteorológico Doppler de dupla polarização banda S (SPOL) do Departamento de Águas e Energia Elétrica (DAEE) e Fundação Centro Tecnológico de Hidráulica (FCTH) durante eventos com ou sem alagamento para cada bairro da cidade de São Paulo durante o ano de 2015. A caracterização foi determinada a partir da função densidade de probabilidade (PDF) da chuva acumulada e da taxa de precipitação, duração da chuva e fração da área de cada bairro onde ocorreu a chuva. Na média, os eventos de alagamento estavam associados com um volume de chuva maior que 30mm e taxa precipitação máxima maior que 30mm/h. Com relação à duração não foi possível encontrar um padrão médio, pois a chuva teve duração mínima de 20 minutos e máxima de 23 horas. Por outro lado, eventos de alagamento tinham alcançado mais de 27% da área do bairro com taxa de precipitação maior que 30 mm/h e 50 mm/h. Destaca-se ao longo desta análise que os bairros localizados próximos aos rios Tietê e Pinheiros e a região central da cidade de São Paulo apresentaram maior probabilidade de ocorrência de alagamento com volumes de chuva mais baixos do que a média de 30 mm por dia e também registraram maior recorrência de pontos alagados. Por último foi desenvolvido um método de regressão logística binária para calcular a probabilidade de ocorrência de alagamentos nos diversos bairros da cidade São Paulo. Este modelo utiliza como parâmetros de entrada a duração da chuva, a taxa de precipitação máxima e a chuva acumulada nas últimas 24 horas. O modelo apresentou uma probabilidade de detecção (POD) média de 1% e uma taxa de falso alarme média (FAR) de 0,6 para os eventos de alagamento, já para eventos sem alagamento o POD médio foi de 96% e a FAR foi de 2,5%. Portanto o modelo consegue prever os casos sem alagamento.<br>This dissertation project presents a characterization of the rainfall estimated from a dual-polarization S-band Doppler meteorological radar (SPOL) of the Department of Water and Electric Energy (DAEE) and Foundation Technological Center of Hydraulics (FCTH) during with or without flooding events for each neighborhood of the city of São Paulo over the year 2015. The characterization was determined by the probability density function (PDF) of the accumulated rainfall and the precipitation rate, rainfall duration and rainfall-area fraction in the neighborhoods. In average, flood events were associated with a rainfall volume greater than 30mm and a maximum rainfall rate greater than 30mm/h. Regarding the duration, it was not possible to find an average pattern, because the rain had a minimum duration of 20 minutes and a maximum of 23 hours. On the other hand, flood events had reached more than 27% of the neighborhood\'s area with a precipitation rate greater than 30 mm/h and 50 mm/h. It is highlighted throughout this analysis that the neighborhoods located near the Tietê and Pinheiros rivers and central region of the city of São Paulo presented a higher probability of flood occurrence with rainfall volumes lower than the average of 30 mm per day and also recorded higher recurrence of flooded spots. Finally, a binary logistic regression method was developed to estimate the probability of occurrence of flooding in the various neighborhoods of the city of São Paulo. This model uses as input parameters rainfall duration, maximum rainfall rate and accumulated rainfall in the last 24 hours. The model presented a mean probability of detection (POD) of 1% and a mean false alarm rate (FAR) of 0,6 for flood events. On the other hand, for events without occurrence of flood a mean POD was 96% and FAR 2,5. Therefore, the model can predict the events without flooding.

<p> The first year of college is a critical period of transition for incoming college students. Learning communities have been identified as an approach to link students together in courses that are intentionally integrated and designed with first-year students' needs in mind. Yet, learning community teaching teams are often not provided with data prior to the start of the semester about their students in order to target interventions. Also, it remains unclear as to which students are most benefitted by participating in learning communities. One question then becomes, what variables known on or before the first day of classes are predictive of first-year student success, in terms of retention and probation status, for first-year college students in learning communities? </p><p> The correlational study employed univariate and multivariate analyses on pre-college data about three consecutive cohorts of first-year students in learning communities at a regional public university in South Texas. Logistic regression models were developed to predict retention and probation status without respect to learning community membership, as well as for each learning community category. </p><p> Results indicated that group differences were not statistically significant based on either first-generation status or age for retention, while group differences were statistically significant for probation status on the basis of all of the pre-college variables except age. Although statistically significant differences were found among the learning community categories for each of the pre-college variables, there were no statistically significant group differences in their retention or probation rates. </p><p> The model to predict retention regardless of learning community membership included five variables, while the model to predict probation status included eight variables. The models for each learning community contained different sets of predictor variables; the most common predictors of retention or probation status were high school percentile and orientation date. </p><p> The study has practical implications for admissions officers, orientation planners, student support services, and learning community practitioners. It is recommended to replicate the study with more recent learning community cohorts and additional pre-college variables, as well as in programs across the nation, to contribute to the literature about the potential for learning communities to enhance first-year student success.</p>

Panel data, also known as longitudinal data, are composed of repeated measurements taken from the same subject over different time points. Although it is generally used in time series applications, forecasting can also be used in panel data due to its time dimension. However, there is limited number of studies in this area in the literature. In this thesis, forecasting is studied for panel data with binary response because of its increasing importance and increasing fundamental roles. A simulation study is held to compare the efficiency of different methods and to find the one that gives the optimal forecast values. In this simulation, 21 different methods, including na&iuml<br>ve and complex ones, are used by the help of R software. It is concluded that transition models and random effects models with no lag of response can be chosen for getting the most accurate forecasts, especially for the first two years of forecasting.

<p>The main objective of this study was to identify ways in which transaction costs can be lowered to improve small-scale farmers&rsquo<br>participation in and returns from agricultural output markets, with specific reference to small-scale mango farmers in Limpopo province. This study hypothesizes that transaction costs are lower in informal spot markets and increase when small-scale farmers sell in more structured markets (formal markets). This study builds on transaction cost economics (TCE) to demonstrate how to overcome transaction cost barriers that small-scale mango farmers face in the agro-food markets. The approach to collect primary information was sequenced in two steps: first, key informant and focus group interviews were conducted and, secondly, a structured survey instrument was administered in two districts of Limpopo. A total of 235 smallscale mango farmers were interviewed. A binary logistic regression model was used to estimate the impact of transaction costs on the likelihood of households&rsquo<br>participation in formal (=1) and informal (=0) agro-food markets. STATA Version 10 was used to analyse the data. This study found that a larger proportion of male than female farming households reported participation in the formal markets, suggesting deep-seated gender differentiation in market participation. The average age of small farmers participating in formal markets is 52, compared to 44 for those in informal markets, implying that older farmers might have established stronger networks and acquired experience over a longer period. Farmers staying very far from the densely populated towns (more than 50 km) participate less in the formal markets than those staying closer (0 &ndash<br>25 km and 26 &ndash<br>49 km), which implies that the further they are from the towns, the less the likelihood of farmers selling in the formal markets. Farmers who own storage facilities and a bakkie (transportation means) participate more in formal markets compared to those who do not own these assets, which suggests that these farmers are able to store mangoes, retaining their freshness and subsequently delivering them to various agro-food markets on time. Households that participate in formal markets have high mean values of income and social grants. However, this study found that the likelihood of a household&rsquo<br>s participation in the markets is less as income and social grants increase. This suggests that households do not invest their financial assets in order to overcome market access barriers. A large proportion of households that own larger pieces of arable land participate in the formal markets, which implies that they are able to produce marketable surplus. Households that have a high mean value (in Rand) of cattle participate more in formal markets than in informal markets. However, this study found that the likelihood of a household&rsquo<br>s participation in the markets does not change with an increase in the value of its livestock. These findings suggest that households do not sell their cattle in order to overcome market access barriers. Reduced transaction costs for small-scale mango farmers in Limpopo should improve their participation in and returns from the agro-food markets. Policy interventions to support this need to focus on: access to storage and transportation facilities, enforcement of gender equity requirements in existing policies, and better access to information about markets.</p>