An Experimental Comparison of Several Clustering and Initialization Methods
This work addresses clustering method selection and initialization for high-dimensional data, but it is incremental as it compares existing methods without introducing new ones.
The paper experimentally compares three batch clustering algorithms (EM, a winner-take-all EM variant, and model-based hierarchical agglomerative clustering) on high-dimensional discrete data, finding that EM significantly outperforms the others, and investigates initialization schemes for EM, showing they yield similar model quality despite differences.
We examine methods for clustering in high dimensions. In the first part of the paper, we perform an experimental comparison between three batch clustering algorithms: the Expectation-Maximization (EM) algorithm, a winner take all version of the EM algorithm reminiscent of the K-means algorithm, and model-based hierarchical agglomerative clustering. We learn naive-Bayes models with a hidden root node, using high-dimensional discrete-variable data sets (both real and synthetic). We find that the EM algorithm significantly outperforms the other methods, and proceed to investigate the effect of various initialization schemes on the final solution produced by the EM algorithm. The initializations that we consider are (1) parameters sampled from an uninformative prior, (2) random perturbations of the marginal distribution of the data, and (3) the output of hierarchical agglomerative clustering. Although the methods are substantially different, they lead to learned models that are strikingly similar in quality.