Uncertainty handling and quality assessment in data mining (Q1408062)

In the introduction the authors give an overview about data mining concepts, methods and tasks. As first step in the quality assessment data pre-processing is considered. In the evaluation of classification methods certain accuracy measures, tests, rules and metrics are presented. Cluster validity with external and internal validity indices including fuzzy clustering is discussed. Basic concepts of uncertainty handling in data mining are the probabilistic and fuzzy approach. Different algorithms, means and rules are introduced. As an application a data mining system handling uncertainty and quality called UMINER is described. Finally an example of medical data analysis serves as case studies for the mining process and the cluster analysis of epidemiological data. Each chapter ends with a summary and literature references. The last two chapters with the example of a data mining system and the case studies in medical data analysis are very useful and present results of applications and real-world cases. This book is not only of great value in advanced undergraduate or postgraduate courses, but gives also good advice to data mining applications in the industry.