Detection of Fake Online Reviews using Semi-Supervised and Supervised: Machine Learning Algorithms

Fake online reviews have become a significant challenge in e-commerce and online platforms, affecting consumer trust and business reputations. This study presents an approach to detecting fake reviews using both semi-supervised and supervised learning techniques. The proposed methodology leverages machine learning models trained on labeled and unlabeled datasets to improve classification accuracy. Experimental results demonstrate the effectiveness of various algorithms in distinguishing between genuine and fake reviews. The findings contribute to the development of robust fraud detection systems for online platforms. The study employs a combination of support vector machines (SVM), Navie's Bayes, and Expectation Maximization (EM) algorithms to detect false reviews. Performance is evaluated using metrics such as accuracy, precision, recall, and F1 score, showing that semi-supervised learning improves classification effectiveness by leveraging unlabeled data.