| Summary: | Reviews pose a significant challenge for e-commerce platforms that rely on online reviews to attract customers. The reviews may contain spam that resembles actual genuine reviews while current spam detection methods lack integration of linguistic and behavioural features. In this study, we propose a machine learning-based approach for detecting spam reviews using classification techniques. We explore various feature extraction methods, including n-grams and sentiment analysis, to represent the textual content of reviews. We evaluate the performance of three classifiers, including Support Vector Machines (SVM), Random Forest, and Naive Bayes, on a dataset of Shopee reviews. Our results show that SVM performs the best in terms of accuracy, precision, and recall, achieving an F1-score of 0.96. We also investigate the impact of different feature extraction methods on classification performance and find that sentiment analysis provides valuable information for detecting spam reviews. Our approach offers a promising solution for automatically identifying spam reviews, which can help e-commerce to maintain the integrity of their online reputation and enhance customer trust. © 2024 IEEE.
|
|---|