The future of many types of businesses depends on their rating on various platforms. However, many times people try to better their reputation by posting fake reviews. Also, posting fake reviews to malign the business’s reputation is common these days. The task of detecting such fake reviews is quite complex. It involves making use of multiple techniques. After the advent of generative AI tools such as ChatGPT, Gemini, Deepseek, etc., the task has become even more complex. People have started creating very realistic looking reviews that are quite hard to distinguish from the reviews posted by real users.

This project tackles this problem. I have used the combination of sentiment analysis and graph-based analysis to detect fake reviews. I have used multiple patterns that make use of the sentiment analysis and user’s history on the platform to detect the fake reviews.