A Framework for the Detection of Suspicious Discussion on Online Forums using Integrated approach of Support Vector Machine and Particle Swarm Optimization

Abstract

With the advancements of technology, human has become more efficient to share the information via internet. It has more often observed that the news stories are initially broken up on social media sites like Twitter, Facebook, etc. and later have taken up by the other channels like printing media or electronic media channels. Any social media based message board where information and user opinion can be discussed is considered as online forums. Most of the data of online forums are stored in text format, hence the present work make use of only text format of suspected postings as evidence for investigation. But there are many suspicious users such as spammers, fraudsters, and other types of attackers that use the latest technology for the criminal activities. So, there is the need to develop tools for the recognition of suspicious activities. There is the existence of tools and methods for the recognition of suspicious information available on internet in the form of user comments or views. In this research paper, we are presenting the existing research work for the detection of suspicious information. The key features and drawbacks of existing concepts have also presented. To improve the autonomous approach, we have also presented a framework using integrated approach of Support Vector Machine (SVM) and Particle Swarm Optimization (PSO). SVM is a statistical learning based data mining approach and PSO is swarm intelligence based concept considered to optimize the parameters of SVM.