An Ant Colony Clustering Algorithm Using Fuzzy Logic

The performance of Data partitioning using machine learning techniques is calculated only with distance measures i.e similarity between the transactions is carried out with the help of distance measurement algorithms such as Euclidian distance measure and cosine distance measure. The distance with connectivity (DWC) model is used to estimate distance between transactions with local consistency and global connectivity information. The ant colony optimization (ACO) techniques are used for the data clustering process. In this paper we propose distance measure model of DWC by enhancing the model using fuzzy logic. The transaction weights are updated using fuzzification process. All the attribute weight values are updated with a fuzzy set weight value. The distance with connectivity model is tuned to estimate distance between the transactions using the fuzzy set values. The distance measure model efficiently handles the uneven transaction distributions. The ant colonyclustering algorithm is also improved with fuzzy logic. The similarity computations are carried out with fuzzy distance measurement models. Un-even data distribution handling, accurate distance measure and cluster accuracy are the features of the proposed clustering algorithm.