L-nearest neighbors ant colony optimization for data clustering

Abstract

It is an important trend to apply the metaheuristics, such as ant colony optimization (ACO ), to data clustering. In general, the ACO for data clustering can accomplish better quality of cluster ing. In this paper, we proposed an improved ACO , to enhance the efficiency of ACO for data clustering. It is based on the assump tion that there are at least one or more neighbors belong to the same cluster in the L nearest neighbors of each instance. It mod ifies the operation of constructing solution to reduce the compu tation time of Euclidean distance. The experimental results show that the L-NNACO is faster than ACO about 38% to 54% • In addition, the L-NNACO is with greater or equal accuracy to the ACO for the various datasets of real world.