Various Feature Extraction and Selection Techniques for Lexicon Based and Machine learning Sentiment Classification

Abstract

Introduction: Natural Language Processing (NLP) is a kind of software that gives computers the ability to comprehend human languages. Words are often used as the fundamental unit for grammatical and semantic analysis on a deeper level, and the major objective of most natural language processing (NLP) projects is word segmentation

Objectives: Proposed System uses the feature extraction method for generating the hybrid feature to get good accuracy for classification in order to solve the practical problem of huge structural differences between different data modalities in a multi-modal environment, where traditional machine learning methods cannot be directly applied to solve the problem.

Methods: In order to do so, this paper introduces the method for generating the hybrid feature. In this System, we utilise a various feature extraction and selection technique from large text. The data has collected from students’ feedback and these feature extraction techniques has applied

Results: Each technique provides different feature extraction while NLP based dependency features provides homogeneous feature set with relationship.

Conclusions: In extensive experimental analysis NLP based features obtains higher precision over the other feature extraction techniques in classification.