Adrian B.W., Azzalini A. Applied Smoothing Techniques For Data Analysis The Kernel Approach

Oxford University Press, 1997. — 208 p. — (Oxford Statistical Science Series, Book 18). — ISBN13: 978-0198523963; ISBN10: 0198523963.The book is about the use of nonparametric smoothing tools in practical data analysis problems. There already exist a number of books in this general area, and these give excellent descriptions of the theoretical ideas, and include good illustrations of applications. The aim of this text is to complement the existing books in two ways. Firstly, this book focuses particularly on intuitive presentation of the underlying ideas across a wide variety of different contexts and data structures. The level of mathematical treatment is correspondingly rather light and where details are appropriate these have been separated from the main text. The emphasis is on introducing the principles of smoothing in an accessible manner to a wide audience. Secondly, the book emphasizes practical issues of inference, where well defined procedures for this exist, rather than of estimation. These aims give particular prominence to the role of graphics, firstly to communicate basic ideas in an informal, intuitive way, and secondly to provide tools for exploring and analyzing a variety of different types of data.
This monograph aims to introduce nonparametric smoothing to statisticians and researchers in other scientific areas who seek a practical introduction to the topic. In view of its style of presentation, the text could also form the basis of a course in nonparametric smoothing suitable for students of statistics. Both density estimation and nonparametric regression are covered although there is considerably more emphasis on the latter in view of its more numerous applications. For each of these two areas a general introduction is provided by one chapter focusing on basic ideas and graphical explorations of data and another which prepares the ground for inferential techniques. The remainder of the text extends these ideas to a variety of different data structures. The final chapter provides a brief introduction to the more general tools provided by generalized additive models. Throughout the book, the authors have endeavored to provide a wide variety of illustrations and exercises.
At the end of each section of this book, examples are given of how graphics and analysis can be carried out in the S-Plus computing environment. Software specifically designed to support the illustrations discussed in the book is freely available. It should be emphasized that the book can easily be used without reference to this. However, for those who have access to S-Plus, the software will provide a means of exploring the concepts discussed in the text, and to provide simple tools for analysis which can be applied in a straightforward and accessible manner. Novice users will be able to explore data without the need to learn the detailed syntax of S-Plus, while experienced users will be able to extend these tools and analyses. Details of how to obtain the S-Plus material, which includes functions, data and scripts, are given in an Appendix.