Rankings and Preferences

This book examines in detail the correlation, more precisely the weighted correlation and applications involving rankings. A general application is the evaluation of methods to predict rankings. Others involve rankings representing human preferences to infer user preferences; the use of weighted correlation with microarray data and those in the domain of time series. In this book we present new weighted correlation coefficients and new methods of weighted principal component analysis.

We also introduce new methods of dimension reduction and clustering for time series data and describe some theoretical results on the weighted correlation coefficients in separate sections.

Numerous applications help the reader to learn quickly Contains two special chapters on two very popular weighted correlation coefficients Describes an easy way of using weighted correlation with already existing statistical software Includes supplementary material: sn.pub/extras

Autorentext
Joaquim Pinto da Costa received his first degree in Applied Mathematics from Porto Universitiy (Portugal), his M. Sc. degree in Applied Statistics from Oxford University and his Ph.D. degree in Applied Mathematics from University of Rennes II (France). Since 199, he is Assistant Professor at the Mathematics Department of Porto University. His research interests include Statistics, Statistical Learning Theory, Pattern Recognition, Discriminant Analysis and Clustering, Data Analysis, Neural Networks, SVMs and Machine Learning.

Inhalt

Introduction.- The Weighted Rank Correlation Coefficient rW.- The Weighted Rank Correlation Coefficient rW2 .- A Weighted Principal Component Analysis, WPCA1: Application to Gene Expression Data.- A Weighted Principal Component Analysis (WPCA2) for Time Series Data.- Weighted Clustering of Time Series.- Appendix.- References.