Unstructured Data Learning using Relation Pattern Knowledge for Efficient Classification through Probabilistic Approach

Abstract

The web is a huge source of information that makes it complicated for humans to choose significant information without knowledge. However, traditional classification methods face significant challenges due to the complex and unregulated distribution of data from large, heterogeneous sources. In such a case, current classification methods or methods of computational complexity and high time to process such large text data. Therefore, it is necessary to develop effective classification methods that can regulate real-time information requirements for different fields. This paper proposed classification by Relation Pattern Knowledge (RPK) and building an FP-Tree for data functions, and applying it to the Probabilistic Relation Pattern Approach (PRPA) to data correlation and classification. FP-Tree uses feature reduction based on covariance deviation deviations and feature relationship trees to provide robust relationship patterns between feature patterns used to efficiently classifies data. The approach is evaluated in comparison with the existing feature reduction and classification approaches measure to evaluate the accuracy to show the improvisation of the proposed approach in various classification.

 Keywords:  Unstructured, Relation Pattern, FP-Tree, Probabilistic, Classification.