Integration of fuzzy logic in data mining to handle vagueness and uncertainty

Abstract

Recent developments in the fields of business investment, scientific research and information technology have resulted in the collection of massive data which becomes highly useful in finding certain patterns governing the data source. Clustering algorithms are popular in finding hidden patterns and information from such repository of data. The conventional clustering algorithms have difficulties in handling the challenges posed by the collection of natural data which is often vague and uncertain. This paper presents the concept of fuzzy clustering (fuzzy c-means clustering) and shows how it can handle vagueness and uncertainty in comparison with the conventional k-means clustering algorithm. © 2008 Springer-Verlag Berlin Heidelberg.