CLASSIFICATION & CLUSTER MINING WITH PARALLEL FUZZY ACO

Abstract

The main purpose of data mining is to extract the required knowledge from data. Data mining is considered as an inter-disciplinary field. Some of the data mining tasks include classification, regression, clustering, and dependence modeling and so on. The task discussed above is solved by using many of the data mining algorithms. The first and foremost step is to design the data mining algorithm for which task the algorithm is to be solved. 

The parallel Fuzzy ACO technique is applied to the data mining tasks of both classification and clustering. The optimization algorithm to partition or creating clusters of data is more efficient. Depending on the nature of data when the level of uncertainty composition is higher in a given data set, parallel fuzzy ACO for classification and clustering gets complicated. Research works have been extensively conducted on MAX-MIN ANT for clustering and classification. The following chapter focuses on MAX-MIN ANT fuzzy for clustering and classification.The primary purpose of clustering algorithm is to segregate the data into self-similar clusters in such a way that the inter-cluster distance is maximized and the intra-cluster distance is minimized. Clustering algorithms can be divided into many types ranging from Hard, Fuzzy, Possibilistic to Probabilistic. Our work concentrates on clustering with parallel fuzzy ACO due to the reduced performance achieved using the sequential fuzzy ACO technique.