Feature selection has acquired importance in view of its commitments in saving classification costs as to time/computational loads. Searching for significant attributes, a feature search method is through decision trees. Features selection falls into 2 gatherings: filter as well as wrapper based techniques. Filters orders attributes through assessments models holding exclusively those features with values more noteworthy than a threshold. Wrappers search feature set for optimum sub-sets in a specific classifier. Execution measurements are connected to sub sets based on their presentation using a specific learning data set classifier. In this paper we proposed parametric feature weight equivalence based feature selection (PFWEFS) method for select a feature of polarity multi – view textual data. The proposed PFWEFS method provides the great result in experiment part.

Feature selection, filter, wrapper, feature set, parametric feature weight equivalence.