Construction and Assessment of Classification Rules is an accessible book the central issues and placing particular emphasis on comparison, perfomance assessment and how to match method to application. Some unusual allocation problems are outlined and a detailed discussion of perfomance assessment is included. The methods used for different appliction domains, such as parametric method, smoothing methods and recursive partitioning are described. The author reviews different approaches and guides researchers and users to suitable classes of techniques.
Part I Basic ideas; Part II Constructing Rules: Fisher's LDA, Nonlinear methods, Recursive partitioning methods, Nonparametric smoothing methods; Part III Evaluating Rules: Aspects of evaluations, Misclassification rate, Evaluating two class rules; Part IV Practical Issues: Some special problems, Some ilustrative examples, Links and comparisons between methods; References; Index