Effective Pattern Deploying Approach in Pattern Taxonomy Model for Text Mining

Abstract

Text mining is a technique that helps user find useful information from a large amount of digital text document. Most existing text mining methods adopted term-based approaches, but they all suffer from the problem of polysemy and synonymy. The next phrase-based approach could not perform better than term based approach. Instead of using typical term-based method many data mining techniques have been proposed for mining useful patterns, however effective usage and updation of discovered patterns is still an open research issue. Pattern deploing and pattern evolving method has also been proposed in order to refine the patterns that helps in improving the effectiveness of pattern discovery. This paper presents an innovative pattern deploying technique based on pattern support to improve effectiveness of using and updating patterns. In existing method called PDM [1] it simply consider the number of sequential patterns containing the given term to compute term weight. The proposed method suggest a probabilistic method to estimate and compute term weight, in which we consider pattern support property which is already discovered in PTM model.