FEROM: A Pragmatic Approach for Sentiment Analysis and Opinion Mining

Abstract

In recent years, online shopping has become popular as it is convenient, reliable and cost effective. The online customer finds it difficult to make purchasing decisions based on the pictures or descriptions provided. The online review often makes it easy for the customer to make decisions for purchasing products as they are a great source to compare products and features. Unfortunately, going through all customer reviews is difficult, especially for popular items as they are in a number of hundreds or thousands.Now-a-days, a large number of availability of rich opinion resources like online review sites and blogs helps customers to understand the opinions of others about the product.We have proposed a system Feature Extraction and Refinement for Opinion Mining (FEROM) which aims to mine customer reviews of a product and extract high detailed product entities on which reviewers express their opinions. The opinions expressed by the customer are reviewed and then they are divided into multiple sentences. The Parts of Speech (POS) tagging is applied on these sentences where each sentence is tagged according to its respective parts of speech. After tagging, their expressions are identified with the help of SentiwordNet dictionary. The words in each sentence are assigned a score and an objective score is calculated with the help of SentiwordNet Dictionary.The sentiment of the customer is identified and opinion orientations for each recognized product entity are classified. These words are then compared with the dictionary of positive and negative which finally segregates the reviews into positive and negative.