Bangla Fake Food Review Detection Using Machine Learning Algorithms

Abstract

In recent years, Fake Review Detection has emerged as a key and well-liked topic for both business and research due to the explosive growth of e-commerce platforms. In today's e-commerce, internet reviews are crucial for helping consumers make the best decisions. Numerous social media platforms and marketing websites host millions of reviews of various goods and services. Customers post reviews of products based on their personal experiences, and other buyers can learn more from these reviews before making a purchase choice. However, since there are no restrictions on what can be posted in a review on any internet platform, it can often be quite difficult to identify the real reviews. According to them, anyone may submit reviews, which increases the quantity of bogus reviews and might provide false information, misleading a buyer. These reviews, whether they are phony or real, have a big impact on an organization's revenue and reputation. Positive reviews typically increase sales and profit while negative reviews negatively impact a company's reputation. We are motivated by this situation to discern between bogus and real reviews. In this study, we cover several machine learning techniques (Naive Bayes, Random Forrest, K-Nearest Neighbor, Logistic Regression, Support Vector Machine (Linear), Decision Tree), and we try to determine whether the review is credible or not using those algorithms. We used a confusion matrix and examined the outcomes of the experiment. We also talk about some of the difficulties we had writing this thesis and some of our future plans for it.