Modified Teacher Learning Based Optimization Method for Data Clustering
Clustering is an important task in engineering domain which can be applied in many applications. Clustering is a process to group the data items in the form of clusters such that data items in one cluster have more similarity to other clusters. On the other side, Teaching–Learning Based Optimization (TLBO) algorithm is a latest population based optimization technique that has been effectively applied to solve mechanical design problems and also utilized to solve clustering problem. This algorithm is based on unique ability of teacher i.e. how the teacher influence the learners through its teaching skills. This algorithm has shown good potential to solve clustering problems but it is still suffering with some problems. In this paper, two modifications have been proposed for TLBO method to enhance its performance in clustering domain instead of random initialization a predefined method previously used to exploit initial cluster centers as well as to deal the data vectors that cross the boundary condition. The performance of proposed modified TLBO algorithm is evaluated with six dataset using quantization error, intra cluster distance and inters cluster distance parameters and compared with K-Means, Particle Swarm Optimization (PSO) and TLBO. From the experimental results, it is clearly obvious that the proposed modifications have shown better results as compared to previously ones.
KeywordsClustering K-Means PSO TLBO
Unable to display preview. Download preview PDF.
- 2.Karaboga, D.: An idea based on honey bee swarm for numerical optimization, technical report-TR06, Erciyes University, Engineering Faculty, Computer Eng. Dep. (2005)Google Scholar
- 4.Holland, J.H.: Adaption in natural and artificial systems (1975)Google Scholar
- 5.Kennedy, J., Eberhart, R.: Particle Swarm Optimization. In: Proc. of IEEE international Conference on Neural Networks (ICW), Perth, Australia, vol. IV, pp. 1942–1948 (1995); Santosa, B., Ningrum, M.K.: Cat swarm optimization for clustering. In: International Conference on Soft Computing and Pattern Recognition (SOCPAR 2009), pp. 54–59 (2009)Google Scholar
- 10.MacQueen, J.: Some methods for classification and analysis of multivariate observations. In: Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, vol. 1(281-297), p. 14 (1967)Google Scholar