Academic literature on the topic 'Medical statistics : statistical software'

Medical research is often weakened by poor statistical practice, and inappropriate use of statistical computer software is part of this problem. The statistical knowledge that medical researchers require has traditionally been gained in both dedicated and ad hoc learning time, often separate from the research processes in which the statistical methods are applied. Computer software, however, can be written to flexibly support statistical practice. The work of this thesis was to explore the possibility of, and if possible, to create, a resource supporting medical researchers in statistical knowledge and calculation at the point of need. The work was carried out over eleven years, and was directed towards the medical research community in general. Statistical and Software Engineering methods were used to produce a unified statistical computational and knowledge support resource. Mathematically and computationally robust approaches to statistical methods were continually sought from current literature. The type of evaluation undertaken was formative; this included monitoring uptake of the software and feedback from its users, comparisons with other software, reviews in peer reviewed publications, and testing of results against classical and reference data. Large-scale opportunistic feedback from users of this resource was employed in its continuous improvement. The software resulting from the work of this thesis is provided herein as supportive evidence. Results of applying the software to classical reference data are shown in the written thesis. The scope and presentation of statistical methods are considered in a comparison of the software with common statistical software resources. This comparison showed that the software written for this thesis more closely matched statistical methods commonly used in medical research, and contained more statistical knowledge support materials. Up to October 31st 2000, uptake of the software was recorded for 5621 separate instances by individuals or institutions. The development has been self-sustaining. Medical researchers need to have sufficient statistical understanding, just as statistical researchers need to sufficiently understand the nature of data. Statistical software tools may damage statistical practice if they distract attention from statistical goals and tasks, onto the tools themselves. The work of this thesis provides a practical computing framework supporting statistical knowledge and calculation in medical research. This work has shown that sustainable software can be engineered to improve statistical appreciation and practice in ways that are beyond the reach of traditional medical statistical education.

Systems have been designed and synthesized using CMOS technology for many years, with improvements in the fabrication process allowing designs to be scaled onto smaller areas with relative ease. The introduction of nano scale CMOS technologies has ended this time of simple scaling, as variations within the silicon now dramatically affect circuit performance and manufacturing yield. These random physical variations cannot be removed from the manufacturing process, requiring that their affects are modelled, predicted and accommodated within the design process. This thesis presents an investigation into the challenges of including these affects within the design process, with a review of the recent research conducted in incorporating variability within timing analysis tools. The conclusion from the literature review is that an accurate, efficient and transparent method of predicting the impact of statistical process variations on the performance of a circuit has not yet been created and adopted by the IC design industry. The investigation begins with the modelling of transistor based statistical process variations at the standard cell level, where it is determined that simple statistical models do not accurately reflect the extremes in performance, and can provide overly pessimistic predictions. The techniques of Monte Carlo Cell Characterisation (MCCC) and Monte Carlo Static Timing Analysis (MCSTA) are introduced as more suitable approaches, which accurately reflect the performance of circuits as modelled by Monte Carlo SPICE simulations, with far less pessimism than the traditional method of Corner Analysis or even modern Statistical Static Timing Analysis. The final section of this thesis focuses on practical implementations of MCSTA, where the sample sizes required to accurately predict circuit behaviour (to within 1% of SPICE)can be reduced to as few as ten, using simple statistical sampling techniques.

Die vorliegende Doktorarbeit bündelt die Veröffentlichungen des Autors und seiner Koautoren zu den Themen e-Learning und statistischer Software. Die Kapitel 2 bis 5 sind Aspekten des e-Learning gewidmet, die Kapitel 6 bis 9 beschreiben die Entwicklung der statistischen Programmiersprache Yxilon. In Kapitel 2, Koautoren Wolfgang Härdle und Sigbert Klinke, wird erörtert, ob und wie computerbasierte Elemente in den Kanon der methodischen Bildung integriert werden sollen und wo die Grenzen des e-Learning in der Statistik-Ausbildung liegen. Kapitel 3, Koautoren Wolfgang Härdle und Sigbert Klinke, gibt Einschätzungen verschiedener e-Learning Plattformen und beschreibt Punkte, die bei der Entwicklung von e-Learning Plattformen berücksichtigt werden sollten. Kapitel 4, geschrieben mit Wolfgang Härdle und Sigbert Klinke, diskutiert zwei Veröffentlichungen in der "International Statistical Review", die eine technische Lösung für die Verbesserung des Verständnisses der Statistik-Lehre vorstellen. Kapitel 5, Koautoren Wolfgang Härdle und Sigbert Klinke, beschreibt die Anwendung von Web-Techniken für die Lehre in Statistik. Weiterhin stellt es die Quantnet Plattform vor, eine Plattform für die Verwaltung von Programmen und Daten. In Kapitel 6, Koautoren Wolfgang Härdle und Sigbert Klinke, diskutieren die Autoren die Anforderungen an eine Statistical Engine. Kapitel 7, geschrieben mit Yuval Guri und Sigbert Klinke, erläutert die Ideen, die zur Re-Implementierung der XploRe Sprache geführt haben und diskutiert ausgewählte technische Aspekte der Yxilon Plattform wie Objektdatenbank und die Erzeugung von kompilierbarem Code für Hochsprachen. In Kapitel 8, Koautoren Wolfgang Härdle und Sigbert Klinke, wird die implementierte Client-Server Struktur beschrieben. Server und Kommunikationsprotokoll werden zusammen mit dem entwickelten Client und der Grafik-Engine beschrieben. Das letzte Kapitel, beschreibt die Struktur der Yxilon Plattform in ihrer jetzigen Form.<br>The following doctoral thesis collects the papers the author has written with his coauthors on e-Learning and statistical software. The chapters 2 to 5 are devoted to selected aspects of e-Learning, the chapters 6 to 9 describe the development of the statistical programming environment Yxilon. In chapter 2, coauthored by Wolfgang Härdle and Sigbert Klinke, the question whether and how computational elements should be integrated into the canon of methodological education and where e-techniques have their limits in statistics education is discussed. Chapter 3, coauthored by Wolfgang Härdle and Sigbert Klinke, gives reviews of different e-learning platforms for statistics and reveals facts that may be taken into account for future e-learning platforms in statistics and related fields. Chapter 4, written with Wolfgang Härdle and Sigbert Klinke, discusses two papers published in International Statistical Review which both offer a technical solution to improve the understanding of statistics by students. Chapter 5, coauthored by Wolfgang Härdle and Sigbert Klinke, describes web-related techniques for teaching statistics. It furthermore introduces the Quantnet platform, a framework to manage scientific code and data. In chapter 6, coauthored by Wolfgang Härdle and Sigbert Klinke, the requirements for a statistical engine are discussed. Chapter 7, written jointly with Yuval Guri and Sigbert Klinke, explains ideas which led to the reimplementation of the XploRe language. In chapter 8, coauthored by Wolfgang Härdle and Sigbert Klinke, the implemented client/server structure of the Yxilon platform is laid out in terms of technical features. The server and the communication protocol are described together with the developed Java client featuring the Jasplot graphics engine. Finally chapter 9 describes the structure of the Yxilon environment in its present form.