Practical Analysis of Representative Models in Classifier: A Review

Abstract

Handling large amount of information with the advancement of technologies is beyond the scope of conventional database technologies. As a solution to the above task, document management systems and machine learning techniques are used. The widely used supervised machine learning technique employing representative models are reviewed in this paper. A classifier creates a model for prediction. The fundamental concern in the classifier is to create an efficient document representation for representative modeling. The model represents documents in an algebraic vector form. The classical models for document representation are numerical vector representational model which gives the frequency of occurrence of the feature as the vector element. Feature extraction and feature selection are done for dimension reduction techniques. In this paper, we have reviewed some existing methods for document representation. The practical analysis and comparison on various techniques for document representation are presented here.