International Conference on Computer and Knowledge Engineering

Home / 13th International Conference on Computer and Knowledge Engineering

A Chaotic Crow Search Algorithm for Overlapping Clustering

Authors :

Mostafa Sabzekar¹ Seyed Vahid Mousavainejad²

1- Department of Computer Engineering, Birjand University of Technology, Iran 2- Department of Computer Engineering and Information Technology Islamic Azad University Birjand Branch Birjand, Iran

Keywords :

clustering،chaotic crow search algorithm،overlapped data،FBCubed metric

Abstract :

Clustering is a major task of data mining that can be used to separate a set of data points into separate groups (clusters). Each clustering algorithm tries to minimize the inter-cluster similarity and, at the same time, maximize the intra-cluster similarity. However, many datasets in the real world are inherently overlapping. In such datasets, a sample can be assigned to more than one cluster. Overlapping K-Means (OKM) method is one of the most effective overlapping clustering methods. Like K-means method, it is sensitive to the selection of the initial cluster centers. In this paper, in order to improve the performance of the OKM method, the best initial centers were found by the chaotic crow search algorithm. We used the FBCubed metric, which evaluates the effectiveness of overlapping clustering algorithms. According to the obtained results of eight publicly available datasets, the proposed method reports better results than other comparing methods and, thud, it can be used as an effective method for clustering datasets.