Improvement of Expectation Maximization Clustering using Select Attribute
Rupali Bhondave, Madhura Kalbhor, Supriya Shinde, K. Rajeswari?
Journal Title:International Journal of Computer Science and Mobile Computing - IJCSMC
Data mining is the process of extracting valuable information from various sources of data and produces knowledge. For mining data, WEKA tool is used. In WEKA there are various processes to produce knowledge, like Preprocess, Classification, Clustering, Select Attribute, and Association etc. This paper focuses clustering Technique. Clustering is a technique by which we can categorize similar objects or dissimilar object. There are various algorithms in clustering. A method attribute selection for experimentation on Expectation Maximization (EM) clustering is used. In attribute selection we used Best First Search (BFS), Random Search for EM clustering, which gives better results than the result obtain without using attribute selection method.