A Deep Analysis on Aspect based Sentiment Text Classification Approaches

Abstract

Now-a-days, people often express their opinions as reviews, comments, feedback in various social networking sites, business organizations. Feedbacks that are given by the end users have a great impact for the evolution of new version of product or service. For business invested in customers, analyzing each piece of feedback by hand can be overwhelming and similarly for an organization to rate an employee regarding his/her performance based on usual quantitative feedback system is a challenging task. Sentiment analysis, developed within this context can be helpful to solve such issues at early stage and provide guidance in improving their sales and productivity. Moreover, reviews written in natural language are mostly unstructured and needs huge time for processing. As the data is available in large size, it’s impossible to process and analyze the information manually. In order to solve this issue, many machine Learning techniques and Deep Learning models are being proposed for automatic learning, extraction and analysis. As the technology advances businesses, organizations, social media and e-commerce sites can benefit from these in-depth insights and customer satisfaction can be analyzed. Sentiment analysis is an excellent source to perform fine-grained analysis like feature-based sentiment analysis and it can be used to identify different aspects expressed at either document or sentence level. This paper highlights the insights of extracting the most important aspects from the opinions expressed in the input text using various machine learning techniques.