Detection of Malicious Data in Twitter Using Machine Learning Approaches

Abstract

: Unlike traditional media social media is populated by unknown individuals who can broadcast whatever they like. This online social media culture is dynamic in its nature and transition to digital media is becoming a trend among people. In upcoming years the use of traditional media will decline, and the increasing use of Online Social Networks(OSNs) blur the actual information of the traditional media. The information generated by the authentic users gives useful information to the general users, on the other hand,Spammers spread irrelevant or misleading information that makes social media a plot for false news. So unwanted text or vulnerable links can be distributed to specific users. These false texts are anonymous and sometimes linked with potential URLs. Due to data restrictions and communication categories, the current systems do not deserve an exact statistical classification for a piece of news. We will study different research papers using various techniques for master training in the prediction and detection of malicious data on social networks online. We tried to find spam tweets from the tweets collected by using Enhanced Random forest classifications and NaiveBayes in this research. To evaluate the work, different validation metrics such as F1-scoring, accurcy and precision values are calculated.