A Multivariate Dynamic Regression Model For Categorical Data

This book studies and applies modern flexible regression models for survival data with a special focus on extensions of the Cox model and alternative models with the aim of describing time-varying effects of explanatory variables. Use of the suggested models and methods is illustrated on real data examples, using the R-package timereg developed by the authors, which is applied throughout the book with worked examples for the data sets.

State space models have gained tremendous popularity in recent years in as disparate fields as engineering, economics, genetics and ecology. After a detailed introduction to general state space models, this book focuses on dynamic linear models, emphasizing their Bayesian analysis. Whenever possible it is shown how to compute estimates and forecasts in closed form; for more complex models, simulation techniques are used. A final chapter covers modern sequential Monte Carlo algorithms. The book illustrates all the fundamental steps needed to use dynamic linear models in practice, using R. Many detailed examples based on real data sets are provided to show how to set up a specific model, estimate its parameters, and use it for forecasting. All the code used in the book is available online. No prior knowledge of Bayesian statistics or time series analysis is required, although familiarity with basic statistics and R is assumed.

One of the most widely used tools in statistical forecasting, single equation regression models is examined here. A companion to the author's earlier work, Forecasting with Univariate Box-Jenkins Models: Concepts and Cases, the present text pulls together recent time series ideas and gives special attention to possible intertemporal patterns, distributed lag responses of output to input series and the auto correlation patterns of regression disturbance. It also includes six case studies.

The use of Bayesian methods for the analysis of data has grown substantially in areas as diverse as applied statistics, psychology, economics and medical science. Bayesian Methods for Categorical Data sets out to demystify modern Bayesian methods, making them accessible to students and researchers alike. Emphasizing the use of statistical computing and applied data analysis, this book provides a comprehensive introduction to Bayesian methods of categorical outcomes. * Reviews recent Bayesian methodology for categorical outcomes (binary, count and multinomial data). * Considers missing data models techniques and non-standard models (ZIP and negative binomial). * Evaluates time series and spatio-temporal models for discrete data. * Features discussion of univariate and multivariate techniques. * Provides a set of downloadable worked examples with documented WinBUGS code, available from an ftp site. The author's previous 2 bestselling titles provided a comprehensive introduction to the theory and application of Bayesian models. Bayesian Models for Categorical Data continues to build upon this foundation by developing their application to categorical, or discrete data - one of the most common types of data available. The author's clear and logical approach makes the book accessible to a wide range of students and practitioners, including those dealing with categorical data in medicine, sociology, psychology and epidemiology.

Central to Dynamic Programming (DP) is the 'cost-to-go' or 'future value' function, which is obtained via solving the Bellman's equation, and central to many Approximate Dynamic Programming (ADP) methods is the approximation of the future value function. The exact DP algorithm seeks to compute and store a table consisting of one cost-to-go value for each point in the state space, and its usefulness is limited by the curse of dimensionality, which renders the methodology computationally intractable in the face of real life problems with high-dimensional state spaces and in the face of continuous state variables. ADP methodology seeks to address and redress the issue of the curse of dimensionality by not seeking to compute the future value function exactly and at each point of the state space; rather opting for an approximation of the future value function in the domain of the state space. Existing ADP methodologies have successfully handled 'continuous' state variables through discretization of the state space and estimation of the cost-to-go or future value function and have been challenged in scenarios involving a mix of 'continuous' and 'categorical' or qualitative state variables. The first part of this dissertation research seeks to develop a flexible, nonparametric statistical modeling method which can capture complex nonlinearity in data comprised of a mix of continuous and categorical variables and can be used to approximate future value functions in stochastic dynamic programming (SDP) problems with a mix of continuous and categorical state variables. This dissertation proposes a statistical modeling method, called 'TreeMARS' which combines the versatility of tree-models with the flexibility of multivariate adaptive regression splines (MARS). An extension of the proposed model, called 'Boosted TreeMARS', is also presented. Comparisons are made to the tree-regression model that uses a similar concept, but only permits the use of linear regression at the terminal nodes. Comparisons are presented on a 10-dimensional simulated data set. The second part of the dissertation research, seeking statistical parsimony, proposes a sequential statistical modeling methodology utilizing the 'sequential' concept from Design and Analysis of Computer Experiments (DACE) to make the grid 'only fine enough' for the 'efficient' discretization and then use MARS methods to approximate future value functions. This methodology can be extended in the future to use tree-based MARS models to approximate future value functions involving a mix of continuous and categorical state variables. This sequential grid discretization is nothing but sequential exploration of the state space and this concept of sequential exploration of the state space provides a statistically parsimonious ADP methodology which 'adaptively' captures the important variables from the state space and builds approximations around these quantities while seeking approximation of the future value functions, by using adaptive and flexible modeling.

Forecasting is required in many situations. Stocking an inventory may require forecasts of demand months in advance. Telecommunication routing requires traffic forecasts a few minutes ahead. Whatever the circumstances or time horizons involved, forecasting is an important aid in effective and efficient planning. This textbook provides a comprehensive introduction to forecasting methods and presents enough information about each method for readers to use them sensibly.

This book is concerned with the use of generalized linear models for univariate and multivariate regression analysis. Its emphasis is to provide a detailed introductory survey of the subject based on the analysis of real data drawn from a variety of subjects including the biological sciences, economics, and the social sciences. Where possible, technical details and proofs are deferred to an appendix in order to provide an accessible account for non-experts. Topics covered include: models for multi-categorical responses, model checking, time series and longitudinal data, random effects models, and state-space models. Throughout, the authors have taken great pains to discuss the underlying theoretical ideas in ways that relate well to the data at hand. As a result, numerous researchers whose work relies on the use of these models will find this an invaluable account to have on their desks. "The basic aim of the authors is to bring together and review a large part of recent advances in statistical modelling of multivariate and multicategorical models within the generalized linear models framework. Based on well-chosen sets of data, these new developments are introduced to a not necessarily expert audience. Completeness was not an aim. The result is a self-contained, well-written text offering the applied researcher a useful insight into the applicability of the general linear model methodology." P.A.L. Embrechts, ETH-Zentrum, Z rich, Switzerland.

This is the first book in longitudinal categorical data analysis with parametric correlation models developed based on dynamic relationships among repeated categorical responses. This book is a natural generalization of the longitudinal binary data analysis to the multinomial data setup with more than two categories. Thus, unlike the existing books on cross-sectional categorical data analysis using log linear models, this book uses multinomial probability models both in cross-sectional and longitudinal setups. A theoretical foundation is provided for the analysis of univariate multinomial responses, by developing models systematically for the cases with no covariates as well as categorical covariates, both in cross-sectional and longitudinal setups. In the longitudinal setup, both stationary and non-stationary covariates are considered. These models have also been extended to the bivariate multinomial setup along with suitable covariates. For the inferences, the book uses the generalized quasi-likelihood as well as the exact likelihood approaches. The book is technically rigorous, and, it also presents illustrations of the statistical analysis of various real life data involving univariate multinomial responses both in cross-sectional and longitudinal setups. This book is written mainly for the graduate students and researchers in statistics and social sciences, among other applied statistics research areas. However, the rest of the book, specifically the chapters from 1 to 3, may also be used for a senior undergraduate course in statistics.

Continuous Multivariate Distributions, Volume 1, Second Edition provides a remarkably comprehensive, self-contained resource for this critical statistical area. It covers all significant advances that have occurred in the field over the past quarter century in the theory, methodology, inferential procedures, computational and simulational aspects, and applications of continuous multivariate distributions. In-depth coverage includes MV systems of distributions, MV normal, MV exponential, MV extreme value, MV beta, MV gamma, MV logistic, MV Liouville, and MV Pareto distributions, as well as MV natural exponential families, which have grown immensely since the 1970s. Each distribution is presented in its own chapter along with descriptions of real-world applications gleaned from the current literature on continuous multivariate distributions and their applications.