Detecion of Opinion Spam in Online Reviews

Abstract

The rise of Internet has led to consumers constantly and increasingly review and research products and services online. Consequently, websites that garner such reviews become primary targets for opinion Spam, which essentially means to sway public opinion by posting deceptive reviews. In this work, we have worked on integrating linguistic features and N-gram modeling to develop a feature set that can be used to detect authentic sounding yet fake reviews. A data set of 1600 reviews from 20 dierent hotels is used for experimentation and results. From the findings, we also try to gure out what can possibly be the factors that help to detect the spammers, and, additionally, make suggestions that can be incorporated by websites to control Spam based on user information