Metaheuristic Clustering Algorithm

Abstract

In this thesis we describe an essential problem in data clustering and present some solutions for it. We investigate using distance measures other than Euclidean type for improving the performance of clustering. We also develop a new point symmetry-based distance measure and prove its efficiency. We develop a novel effective k-means algorithm which improves the performance of the k-mean algorithm. We develop a dynamic linkage clustering algorithm using kd-tree and we prove its high performance. The Automatic Clustering Differential Evolution (ACDE) is specific to clustering simple data sets and finding the optimal number of clusters automatically. We improve ACDE for classifying more complex data sets using kd-tree. The proposed algorithms do not have a worst-case bound on running time that exists in many similar algorithms in the literature. Experimental results shown in this thesis demonstrate the effectiveness of the proposed algorithms. We compare the proposed algorithms with other famous similar algorithms. We present the proposed algorithms and their performance results in detail along with promising avenues of future research.