Soft computing: A new perspective for some data mining problems

Abstract

Databases in the real world are often dynamic, incomplete, noisy and very large and most of the problems which appear in the Data Mining field are caused by these characteristics. However, it is often unnecessary to obtain results (rules, clusters, etc.) with a high degree of precision since the input data is itself imprecise. For these reasons, we propose a new approach to cope with these situations. We refer to the “Soft Computing” methodology which combines the ability of Fuzzy Logic to represent and manage imprecise data and knowledge together with the accepted capacities for learning and heuristic computation of Neural Networks and Genetic Algorithms. The aim of the paper is to present some of the possiblities offered by the use of Soft Computing.