Leveraging Extravagant Linguistic Patterns to Enhance Fake Review Detection: A Comparative Study on Clustering Methods

Abstract

Online reviews are crucial in shaping consumer behavior and business success, but the growing presence of fake reviews threatens the credibility of these platforms. This paper introduces a novel approach to fake review detection by focusing on the use of extravagant language words that are excessively positive or negative, often used to manipulate opinions. We explore how extravagant words can serve as strong indicators of review authenticity. By leveraging advanced clustering methods and BERT embeddings for linguistic feature extraction, we examine patterns in deceptive and genuine reviews. Two diverse datasets are utilized to validate our approach, and the model's performance is evaluated through accuracy, precision, recall, and F1-score. Our findings demonstrate the effectiveness of extravagant words in enhancing the detection of fake reviews and provide valuable insights into improving the reliability of online review systems. This approach offers practical implications for boosting consumer trust and enhancing the integrity of digital platforms.