Performance Analysis of Standard k-Means Clustering Algorithm on Clustering TMG format Document Data

Document clustering is useful in many information retrieval operations such as document browsing, organization and viewing of retrieval results, generation of Yahoo-like hierarchies of documents, etc. The general goal of clustering is to group data elements such that the intra-group similarities are high and the inter-group similarities are low. Generative models based on the multivariate Bernoulli and multinomial distributions have been widely used for text classification. In this work, we explore the k-means clustering algorithm for document clustering problem. The proposed work implements the standard k-mean clustering algorithm and tests it with TMG format document data and L2normalized document data. The results of the k-means clustering algorithm are compared with von Mises-Fisher model-based clustering (vMF-based k-means) algorithm. Key WordsText Clustering, Classification, Document Clustering, Model Based Clustering, Term Document Matrix, Text to Matrix Generator (TMG), k-means, MisesFisher Clustering