Non-Class Element based Iterative Text Clustering Algorithm for Improved Clustering Accuracy using Semantic Ontology

Abstract

In this method, a non-class element based iterative clustering approach is discussed. First the method extracts the terms of the class using the preprocessing algorithm. The preprocessing algorithm extract the terms and for each term the method performs stemming and tagging. Finally the method selects a subset of terms from the entire term set. For each term identified, the method computes the semantic bound measure using the semantic ontology. Then the method computes the semantic closeness measure for the document. Based on computed semantic closeness measure, the method selects the class for the document. Then the method identifies the non class elements from the document and for each class, the method computes non class weight. Based on computed semantic closeness and non class weight, the method computes the true weight for each class. Finally a class is selected based on computed true weight.