Ulteriori informazioni
The author has written an accessible book presenting the central issues and placing particular emphasis on comparison, performance assessment and how to match method to application. He reviews the different approaches and guides researchers (and the huge class of users and potential users) to suitable classes of techniques. He points the way to more detailed study in "further reading" sections in each chapter.
Sommario
BASIC IDEAS.
Introduction.
CONSTRUCTING RULES.
Fisher's LDA and Other Methods Based on Covariance Matrices.
Nonlinear Methods.
Recursive Partitioning Methods.
Nonparametric Smoothing Methods.
EVALUATING RULES.
Aspects of Evaluation.
Misclassification Rate.
Evaluating Two- Class Rules.
PRACTICAL ISSUES.
Some Special Problems.
Some Illustrative Applications.
Links and Comparisons Between Methods.
References.
Index.
Info autore
David J. Hand, Professor em. Für Mathematik und Senior Research Investigator am Imperial College London, ist ehemaliger Präsident der Royal Statistical Society und derzeit wissenschaftlicher Chefberater von Winston Capital Management, einem der erfolgreichsten auf automatisierten (algorithmischen) Handel spezialisierten Hedgefonds in Europa. David Hand lebt in London.
Riassunto
Classification rules can be defined as objective, formal methods used for statistical classification. This text presents the central issues and placing particular emphasis on comparison, assessment and how to match method to application.
Testo aggiuntivo
"Evaluation rules are handled and the most important methods to calculate misclassification rates are presented." (Monatshefte fur Mathematik, Vol. 127, No. 2, 1999)
"...an accessible book..." (Zentralblatt Math, Vol. 997, No.22, 2002)
Relazione
"Evaluation rules are handled and the most important methods to calculate misclassification rates are presented." (Monatshefte fur Mathematik, Vol. 127, No. 2, 1999)
"...an accessible book..." (Zentralblatt Math, Vol. 997, No.22, 2002)
