The Analysis of User Reviews on Digikala with the Aim of Detecting Deceptive Opinions
Subject Areas : هوش مصنوعی و رباتیک
hosein sarlak
1
*
,
Alireza Sheikh
2
1 -
2 - Department of Business Management, Faculty of Management, Science and Technology
Keywords: Detecting Deceptive Opinions, Machine Learning, Large Language Models, Sentiment Analysis, Deep Neural Networks,
Abstract :
This research investigates and analyzes user reviews on the Digikala platform with the aim of detecting deceptive opinions. Initially, user review data was collected and preprocessed, followed by the application of various machine learning models and large language models to identify deceptive reviews. The results indicated that deceptive reviews are often written by users with lower credibility, and reviews that receive more dislikes tend to be less credible. Additionally, positive reviews are the most prevalent, and users with positive reviews generally receive more likes. This study demonstrated that employing large language models and machine learning can enhance the detection of deceptive opinions and improve the accuracy of user review monitoring systems, aiding in better identification of valuable and influential users.