Fake Review Prediction and Review Analysis

Abstract

Online reviews can be deceptive or manipulative\nevaluations of services and products which are often carried out\ndeliberately for manipulation strategy to mislead the readers.\nIdentifying such reviews is an important but challenging problem.\nThere are even some associations in the merchandise industry\nwho are hiring professionals to write fake reviews so that they can\npromote their products or defame rivals products. Hence we aim to\ndevelop a method which will detect fake reviews and remove them.\nThe proposed method classifies users reviews into suspicious,\nfake, positive and negative categories by phase-wise processing. In\nthis paper, we are processing hotel reviews by using different data\nmining techniques. Moreover the reviews obtained from users are\nbeing classified into positive or negative which can be used by a\nconsumer to select a product. Organizations providing services\ncan monitor customer sentiments by scrutinizing and\nunderstanding what the customers are thinking about products\nthrough reviews. This can help buyers to purchase valuable\nproducts and spend their money on quality products. Also in our\nmodel end users see star ratings based on reviews for each hotel.