Online Multilingual Spam Review Detection using Twin Support Vector Machine and Pre-Trained Word Embedding

The main technique for binary classification in this suggested spam review detection system, which is tailored for the English language, is the Twin Support Vector Machine (TSVM). By simultaneously taking into account both positive and negative classifications, the TSVM improves its ability to discriminate between reviews that are spam and those that are real. The proposed system uses a metaheuristic optimization strategy to maximize performance, using a proprietary method for optimizing feature weighting for the TSVM and fine-tuning hyper parameters. The system is intended to be both flexible and efficient; it can effectively identify and filter spam reviews from datasets that are in the English language. Thorough experimental assessments regularly reveal that the TSVM-based framework performs better than state-of-the-art algorithms, demonstrating its resilience and efficacy in tackling the problems associated with fraudulent online reviews written in English.