A Comprehensive Survey on Computer Forensics: State-of-the-Art, Tools, Techniques, Challenges, and Future Directions

Abstract

People produce a deluge of product evaluations and comments due to the proliferation of Internet-based applications like social networks and e-commerce websites. Therefore, processing them automatically becomes very important. There have been a lot of proposals for systems that can produce and display reputation by mining numerical and textual evaluations in the last decade. But they have overlooked the possibility that bad actors may write evaluations online with the express purpose of damaging the target product's reputation. Beyond that, these systems only care about the entity's reputation value and don't bother to generate reputation ratings for the product's individual features. In order to provide trustworthy reputation values, we built a system that uses spam filtering, review popularity, review posting time, and aspect-based sentiment analysis. With the use of user reviews gathered from several sources, the suggested model assigns numerical reputation ratings to entities and their attributes. Additionally, our suggested system provides a high-tech visualization tool that shows comprehensive data on its output. Experimental findings comparing the proposed approach to state-of-the-art reputation generating methods demonstrate its efficacy on several datasets obtained from diverse platforms (e.g., Twitter, Facebook, Amazon, etc.).