Fake Review Detection using Naive Bayesian Classifier

Abstract

The issue of fake reviews is becoming an increasingly prevalent one on the internet. The purpose of these reviews is to deliberately deceive potential customers and influence their purchasing decisions. Businesses and customers alike are looking for ways to spot and filter out these fake reviews as a result. The Naive Bayes algorithm is one effective method for identifying fake reviews. A well-known machine learning algorithm for classification tasks is Naive Bayes. It is based on the probability theorem of Bayes, which enables us to determine the probability of an event with some evidence. The Naive Bayes algorithm can be trained on a dataset of reviews that are known to be real or fake in the context of fake reviews. The characteristics of genuine and fake reviews are then learned with the algorithm by utilizing this training data. After the algorithm has been trained, it can use the characteristics of new reviews to determine whether they are genuine or fake. The fact that Naive Bayes is a relatively straightforward algorithm that can be trained quickly and easily is one advantage of using it to detect fake reviews. Additionally, it works well with text data, which is the format used by the majority of reviews. Having said that, it’s critical to keep in mind that Naive Bayes isn’t perfect and may not be able to spot all fake reviews. Cleaning and normalizing data, dealing with missing data, and dealing with outliers are all potential obstacles in data pre-processing.