Fake Review Detection Using Extravagant Words: A Comparative Study of K-means, K-mode, and Hierarchical Classification

Abstract

In the digital age, online reviews playa critical role in influencing consumer decisions and business reputations. How-ever, the prevalence of fake reviews undermines the credibility of online platforms. In this study, we propose a new method based on the use of extravagant words as indicators of fake reviews, leveraging their tendency to exaggerate and manipulate readers. We aim to evaluate the effectiveness of this approach by comparing three distinct methods: K-means, K-mode, and Hierarchical clustering. Utilizing two comprehensive datasets, we preprocess the data, extract features, and implement the tree classification methods. Additionally, we incorporate BERT for improved feature extraction. The performance of each method is assessed using metrics such as accuracy, precision, recall, and Fl-score. Our findings reveal significant insights into the utility of extravagant words and BERT in distinguishing fake reviews and highlight the comparative strengths and weaknesses of K-means, K-modes and Hierarchical clustering. This study provides a novel approach to enhance review authenticity verification, with implications for improving consumer trust and platform integrity.