BUYER RATING AND SENTIMENT ANALYSIS USING DISCRIMINATORY SUPPORT VECTOR MACHINE (SVM)

Abstract

This substantial issue is increasingly important in business and culture. It presents many challenging research scenarios but guarantees a relevant insight for everybody interested in view evaluation and social networking analysis. This paper's key aim is to detect sentiment polarity such as positive, negative, and emoji representation with customer feedback on various products. Opinion mining from e-commerce sites has a significant part in making purchase decisions and founders to boost their product and marketing strategies. But, it becomes very difficult for the clients to understand and assess the product's actual view manually. Because of this, we need an automated way. The majority of the researchers used machine learning algorithms to do an automated representation of phrase embedding. Among the popular techniques in machine learning has been used the support vector machine (SVM). The weighted support vector machine (WSVM) is the improved version for the standard SVM to grow the outlier sensitivity issue. In this paper, the word2Vec version uses to extract the attributes from customer reviews in WSVM based on opinion analysis of product reviews in E-commerce websites. The experiment result shows that the suggested WSVM can works better on the opinion classification job doing any version applied.