Deriving the Partial Order of Documents to Extend Clustering Applications

Abstract

The exponential growth of text documents over the internet has paved the way for systematic document organization. It is widely accepted that the document clustering has augmented the information retrieval process to a greater extend. Basically all the text clustering algorithms tend to establish more appropriate clusters of text documents, and the accuracy of text clustering algorithms are measured based on cluster cohesion and separation. Keeping to the basic principle of clustering to minimize cohesion and maximize separation, all the algorithms deploy different strategies to generate better quality clusters. It is observed from the detailed literature survey that Classification, Categorization, Plagiarism Detection and Clustering are correlated. All these text mining tasks are performed based on indexing, searching or relating the key terms present in the documents. Moreover, all the text mining methods focuses on establishing the similarity or difference among the text documents, by which they perform their intended tasks. Hence, they tend to limit the application of clustering only to complement information retrieval task. This paper tries to present an algorithm to establish the partial order among the text documents and thus to extend the applications of clustering.