A Review on Feature Extraction Techniques for Sentiment Classification

Abstract

Sentiment analysis is used for social media supervising. It is used to analyze comments, recommendations provided by the reviewers. Depending on this analysis, products will be purchased through either online or offline. Machine learning is an area of artificial intelligence which gives computers the capability to learn without using exceptional computation. Traditional machine learning algorithms used for feature extraction and sentiment classification leads to scalability and computational problems while dealing with a huge amount of data. Deep learning is a field of approach concerned with neural implementations. It is a nonscalable and accurate technique for sentiment classification, automatic feature extraction and dimensionality reduction. Machine learning techniques are easier to implement but give less accuracy compared to deep learning. Deep learning methods are complex to implement from a human point of view, but machines require complex structures to learn large data set easier.