Electronic Recommendation Based on Customer Review Using Text Mining

Abstract

In this paper, buyers are looking for reviews of the products before purchasing them. In view of this, online shopping platforms are encouraging their customers to provide reviews on products that would help future customers and the service provider to enhance their services. These reviews are normally in natural language mostly in the English language. These are used to analyze and provide data that is used for repairing and building new products because most services are unable to review consumers' reviews at the same time regularly, so they need mining tools to learn about those reviewers, which is what consumers need for their goods. Users review assessments for upgrading their products. Reviews are analyzed by customers to decide whether to purchase or not to purchase. The main objectives of this study are to develop a recommendation system based on customer reviews, develop a dataset of customer reviews from Amazon and eBay, analyze the reviews to create a database of products, develop an algorithm for generating positive/negative scores for a product, develop a method for gathering user requirements in natural language and to identify the main product, and also develop an algorithm to match the user request with the product and generate recommendations. The study concluded that the selected program was suitable for analyzing and managing the review data.