A hybridise approach for data clustering based on cat swarm optimisation

摘要

Clustering is a technique that is used to find subsets of similar objects from a given set of objects such that the objects in the same subset are more similar than others. Over the time, a number of algorithms have been developed to solve the clustering problems. K-harmonic means (KHM) is one of the popular techniques that have been applied in clustering as a substitute for K-means algorithm because it is insensitive to initialisation issues due to in boosting function. But, this technique also suffered with same shortcoming of getting trapped in local optima. On the other hand, cat swarm optimisation (CSO) is the recent population-based optimisation method came in picturesque for global optimisation. So keeping this in mind, an attempt is made to integrate the CSO and KHM techniques for efficient data clustering and the proposed method is named CSOKHM. The proposed method attains the advantages of both the algorithms. In order to investigate the efficacy of the proposed method, it is applied on the seven datasets; out of seven datasets, five datasets are real which are downloaded from the UCI repository while rest are artificial ones. A comparison is also performed with other clustering methods and results favour the proposed method.