An Exploratory Study of K-Means and Expectation Maximization Algorithms

In this paper, K-Means and Expectation-Maximization algorithms are part of the commonly employed methods in clustering of data in relational databases. Experiments conducted with both clustering algorithms revealed that both algorithms have been found to be characterized with shortcomings. The parameters considered in evaluating the results of findings are the number of iterations (no distinct convergence, 1), the computation time (not defined, 3.2s) and the memory space (not defined, 1.1MB) consumed at the point of convergence of both K-means and Expectation-Maximization algorithms respectively. The results obtained revealed that Expectation-Maximization algorithm’s quick and premature convergence cannot be said to have guaranteed optimality of results while K-means was found not to guarantee convergence. Though reasonable conclusion could be drawn from results obtained with Expectation-Maximization algorithm, its premature convergence may raise some questions of doubt with regards to reliability of results obtained.