Research on product review analysis and spam review detection

Many e-commerce web sites enable their customers to write product reviews and feedback in the form of ratings. This gives the company personnel an indication about their products' standing in the market, while also enabling fellow customers to form an opinion and help purchase a product. However, due to the reason of profit or fame, many target products are promoted or demoted in the form of spam. It may contain fake reviews or malicious opinions, which is misleading. In this paper, we make an attempt to detect spam and fake reviews, and filter out reviews with expletives, vulgar and curse words, by incorporating sentiment analysis. Other studies solve this by using just the ratings as a parameter. This paper, however, by taking upon consideration Amazon dataset, matches the posted rating with the calculated ratings of each review, by producing a sentiment score with the help of an inhouse dictionary. Finally, we graphically show and analyze the different features of the product which adds to its popularity or demotion.