Probability estimation via smoothing in sparse contingency tables with ordered categories》.pdf

Statistics Probability Letters 5 (1987) 55-63 January 1987 North-Holland PROBABILITY E S T I M A T I O N VIA S M O O T H I N G IN SPARSE CONTINGENCY TABLES W I T H O R D E R E D CATEGORIES Jeffrey S. SIMONOFF New York University, Graduate School of Business Administration, New York, N Y 10006, USA Received February 1986 Revised May 1986 Abstract: Probability estimation in sparse two-dimensional contingency tables with ordered categories is examined. Several smoothing procedures are compared to analysis of the unsmoothed table. It is shown that probability estimates obtained via maximum penalized likelihood smoothing are consistent under a sparse asymptotic framework if the underlying probability matrix is smooth, and are more accurate than kernel-based and other smoothing techniques. In fact, computer simulations indicate that smoothing based on a product kernel is less effective than no smoothing at all. An example is given to illustrate the smoothing technique. Possible extensions to model building and higher dimensional tables are discussed. Keywords: ordered categorical data, maximum penalized likelihood, kernel-based smoothing, Bayes methods, computer simulations, analysis of association. 1. Introduction categories, since it would be reasonable to suppose that the probabilities were generated from a Although the analysis of contingency tables has smooth bivariate density function. been extensively studied, and is fairly we