A Practitioner's Guide to Resampling for Data Analysis, Data Mining, and Modeling

摘要

Researchers in a wide variety of disciplines, ranging from engineering, through biology, to medicine. This is an elementary introduction to the use of resampling methods, such as permutation tests and bootstrap methods, applied in a wide variety of statistical problems. It assumes a basic knowledge of statistics, but avoids deep technicalities, with the perhaps inevitable consequence that sometimes some of the explanations or justifications seem a little superficial, even ambiguous. The choice of material struck me as rather ad hoc. For example, there is no discussion of generalised linear models or time series (the two paragraphs on block bootstrap are too brief to be helpful), and although there are two chapters on multivariate problems, one covering testing and the other classification trees, there is no discussion of principal components analysis, factor analysis, or a host of other multivariate methods (correspondence analysis is covered, in a half page in the categorical data chapter).