Chaos A Statistical Perspective

Chaos: A Statistical Perspective

It was none other than Henri Poincare who at the turn of the last century, recognised that initial-value sensitivity is a fundamental source of random ness. For statisticians working within the traditional statistical framework, the task of critically assimilating randomness generated by a purely de terministic system, often known as chaos, is an intellectual challenge. Like some other statisticians, we have taken up this challenge and our curiosity as reporters and participants has led us to investigate beyond the earlier discoveries in the field. Earlier statistical work in the area was mostly con cerned with the estimation of what is sometimes imprecisely called the fractal dimension. During the different stages of our writing, substantial portions of the book were used in lectures and seminars. These include the DMV (German Mathematical Society) Seminar Program, the inaugural session of lectures to the Crisis Points Project at the Peter Wall Institute of Advanced Stud ies, University of British Columbia and the graduate courses on Time Series Analysis at the University of Iowa, the University of Hong Kong, the Lon don School of Economics and Political Science, and the Chinese University of Hong Kong. We have therefore benefitted greatly from the comments and suggestions of these audiences as well as from colleagues and friends. We are grateful to them for their contributions. Our special thanks go to Colleen Cutler, Cees Diks, Barbel FinkensHidt, Cindy Greenwood, Masakazu Shi mada, Floris Takens and Qiwei Yao.

Chaos

Statistical Learning from a Regression Perspective

Statistical Learning from a Regression Perspective considers statistical learning applications when interest centers on the conditional distribution of the response variable, given a set of predictors, and when it is important to characterize how the predictors are related to the response. As a first approximation, this is can be seen as an extension of nonparametric regression. Among the statistical learning procedures examined are bagging, random forests, boosting, and support vector machines. Response variables may be quantitative or categorical. Real applications are emphasized, especially those with practical implications. One important theme is the need to explicitly take into account asymmetric costs in the fitting process. For example, in some situations false positives may be far less costly than false negatives. Another important theme is to not automatically cede modeling decisions to a fitting algorithm. In many settings, subject-matter knowledge should trump formal fitting criteria. Yet another important theme is to appreciate the limitation of one’s data and not apply statistical learning procedures that require more than the data can provide. The material is written for graduate students in the social and life sciences and for researchers who want to apply statistical learning procedures to scientific and policy problems. Intuitive explanations and visual representations are prominent. All of the analyses included are done in R.