Clustering is one of the most common unsupervised data mining classification techniques for splitting objects into a set of meaningful groups. However, the traditional k-means algorithm is not applicable to retrieve useful information / clusters, particularly when there is an overwhelming growth of multidimensional data. Therefore, it is necessary to introduce a new strategy to determine the optimal number of clusters. To improve the clustering task on high dimensional data sets, the distance based k-means algorithm is proposed. The proposed algorithm is tested using eighteen sets of normal and non-normal multivariate simulation data under various combinations. Evidence gathered from the simulation reveal that the proposed algorithm is capable of identifying the exact number of clusters.

Clustering plays an essential role for data analysis and it has been widely used in different fields like data mining, machine learning and pattern recognition. Clustering problem divides some data, which is more similar to each other in terms of parameters under consideration. One of available methods about this area is k-means algorithm. Despite dependency of this algorithm on initial condition and convergence to local optimal points, it classifies n data to k clusters with high speed. Since we encounter a huge volume of data in clustering problems, one of suitable methods for optimal clustering is to use a meta-heuristic algorithm, which improves clustering operation. In this paper, differential evolution algorithm is utilized for solving available problems in k-means algorithm. In this paper, meta-heuristic algorithm has been used for solving data clustering problems. The applied algorithm has been compared with k-means algorithm on six known dataset from UCI database. Results show that this algorithm achieves better clustering than k-means algorithm.