Data Mining and Knowledge Engineering

K. Karthika ¹, G. Komarasamy ¹

Affiliations
1 Bannari Amman Institute of Technology, Sathyamangalam, Tamil Nadu, IN

Abstract

This paper presents a new data clustering algorithm called KPSO algorithm, a combination on K-means and Particle swarm Optimization algorithms. Unlike traditional K-means method, KPSO need not specify the number of clusters to be given prior the clustering process and is able to find the optimal number of clusters during the clustering process. In each and every iteration of KPSO, a discrete PSO is used to optimize the number of clusters with which the K-means is used to find the best clustering result.

Keywords

Data Clustering, K-Means, Particle Swarm Optimization, Clustering Process.

Full Text

     

G. Komarasamy ¹, Amitabh Wahi ²

Affiliations
1 Department of Computer Science and Engineering, Bannari Amman Institute of Technology, Sathyamangalam, Tamilnadu, IN
2 Department of Information Technology, Bannari Amman Institute of Technology, Sathyamangalam, Tamilnadu, IN

Abstract

Cluster analysis is one of the primary data analysis methods and K-means algorithm is well known for its efficiency in clustering large data sets. The K-means (KM) algorithm is one of the popular unsupervised learning clustering algorithms for cluster the large datasets but it is sensitive to the selection of initial cluster centroid, and selection of K value is an issue also sometimes it is hard to predict before the number of clusters that would be there in data. There are inefficient and universal methods for the selection of K value, till now we selected that as random value. In this paper, we propose a new metaheuristic method KMBA, the KM and Bat Algorithm (BA) based on the echolocation behavior of bats to identify the initial values for overcome the KM issues. The algorithm does not require the user to give in advance the number of clusters and cluster centre, it resolves the K-means (KM) cluster problem. This method finds the cluster centre which is generated by using the BA, and then it forms the cluster by using the KM. The combination of both KM and BA provides an efficient clustering and achieves higher efficiency. These clusters are formed by the minimal computational resources and time. The experimental result shows that proposed algorithm is better than the existing algorithms.

Keywords

Centroid, Clustering, Metaheuristic, BAT Algorithm.

Full Text

     

D. Sharmila Rani ¹, N. Kousika ², G. Komarasamy ³

Affiliations
1 Sri Krishna College of Engineering and Technology, Coimbatore, IN
2 Department of Computer Science, Sri Krishna College of Engineering and Technology, Coimbatore, IN
3 Computer Science and Engineering Department, Bannari Amman Institute of Technology, IN

Abstract

Clustering is an unsupervised classification that is the partitioning of a data set in a set of meaningful subsets. Each object in dataset shares some common property often proximity according to some defined distance measure. Among various types of clustering techniques, K-Means is one of the most popular algorithms. The objective of K-means algorithm is to make the distances of objects in the same cluster as small as possible. Algorithms, systems and frameworks that address clustering challenges have been more elaborated over the past years. In this review paper, we present the K-Means algorithm and its improved techniques.

Keywords

Classification, Clustering, K-Means Clustering, Partitioning Clustering.

Full Text