Determination of Potential Criminals in Social Network

Abstract

This paper proposes a new approach to determine potential criminals by performing content analysis on Tweets. The analysis is done with the help of machine learning technologies and big data analysis. In this study, we have utilized from the MLP algorithm. A dataset consisting of 384 words are used to make the classification process. Dataset consist of two classes that are organized crime and cyber-crimes. In the analysis process, date, time, location values of sent twitter sharings are also used. Criminals can be detected with a success rate of 71,61%. Also, the developed system identifies potential criminals who commit an offence of organized crime and cybercrime with the world's most widely used social media. User can use the system to get accurate results for scanning potential criminals with analyzing their sharings or scanning keywords to reach potential criminals. In addition to this property, user can get results that are more accurate with narrowing the content screening with location and date information. It is thought that the proposed system might help to find criminals and the security forces can easily detect them by the developed software.