Laying chicken algorithm (LCA) based for clustering

Abstract

Many researches and applications related to fuzzy clustering are still very important and interesting. One of them is focusing on sensitive fuzzy based for clustering regarding to the initial centroid. If the initial centroid is bad, it will not converge to a good clustering results. This is due to iteration process easily stuck in local optimum. A stochastic global optimization is used to handle this optimization problem. Laying Chicken Algorithm (LCA) is one current of the stochastic global optimization as a multi swarm optimization. It adapts individual laying hens in the process of incubating their eggs to improve the Chicken Swarm Optimization (CSO). In this article, Fuzzy C-Means (FCM) and LCA were modified to repair local optimum of Fuzzy Clustering proposed. The LCA is used to find the global optimum of FCM. Data were redefined to be a matrix of identity as initial population chicken of LCA. Then, the improvement of population and solution were updated to get the optimal solution. The experiment was run by using UCI dataset. The comparison was conducted by evaluating the term of Davies Boulding index, rank index and accuracy of overall average. The results indicate that FCMLCA method has better performance than the compared methods.