Effective Pattern Discovery for Text Mining Using Pattern Taxonomy Model

We describe an effective and innovative pattern discovery technique. In order to overcome the problem of misinterpretation and low frequency pattern taxonomy model is used. It makes use of closed sequential patterns and pruning nonclosed patterns to obtain the d-patterns. And reshuffle the terms support by using normal forms to get relevant terms from negative documents. This includes pattern deploying and pattern evolving for improving the effectiveness of using and updating discovered patterns for finding related and interested information. For deploying patterns the D-pattern mining algorithm and for evolution of patterns IPEvolving and shuffling algorithms are used. Deployment based on positive documents while evolution is based on negative documents. It requires less number of patterns for training phase. This model is effective in time complexity and coding also.