|
Abstract
|
Data clustering is a well-known data analysis technique needed in many fields such as Engineering, Education, Medicine,
and Bioinformatics. The-means algorithm is one of the most common algorithms used for data clustering. However, the
results of K-means highly depend on the initial centers and may converge to local optima. In order to overcome these
problems, many studies have been done in clustering. This paper proposes hybrid approach for data clustering using
cuckoo search and differential evolution algorithms. Cuckoo Search (CS) is a novel swarm-intelligence based algorithm
that has been proposed recently. This algorithm has less control parameters and good ability in solving different kinds of
problems, however it has some problems like much number of functional evaluation and it sometimes easily get trapped
in local minima. The proposed algorithm benefits from Differential Evolution Algorithm (DE) and also benefits from
Mantegna levy distribution to promote the CS algorithm in solving the number of functional evaluation and also obtaining
better convergence speed and high precision in a short time. For testing the performance of the proposed algorithm this
paper employed six real standard benchmark datasets from UCI machine learning repository. The simulation results show
that the proposed algorithm has ability in obtaining better results in terms of Convergence Speed, Accuracy, and also
reducing number of functional evaluation.
|