Dynamic Combination of Genetic Clustering Using Efficient Optimization Algorithm
Abstract
Data clustering is one of most commonly used techniques in the literature of Machine Learning and Pattern recognition. K-means algorithm is a well-known clustering algorithm in data mining because of its high operational speed and ease of implementation. However, problems such as sensitivity to initial solutions and the risk in getting trapped at local optima hinder its performance. Inspired by a dynamic population assignment method which is based on the rate of improvement and mean of performance, a new approach named DCGCA is proposed by employing the Dynamic Combination of Genetic Clustering Algorithm. It is shown that the performance of DCGCA is improved by employing the genetic clustering through the weighted particle swarm optimization algorithm in solving the problem getting trapped in local optima. The simulation results of the proposed method and comparison analysis with alternative methods confirms its higher efficiency and better performance.