Real Time Detection of Drifted Twitter Spam Based on Statistical Features
Sayali Kamble, Sunil Mahadev Sangve · 2018
With an increasing popularity of Twitter, huge number of tweeters tweeting across the world, real time searching techniques are emerging to permit individuals following the repercussion of news and events on Twitter. These events allow users to spread the news and enable users to talk about these actions and then post their status, these services open prologues for novel types of spam. The most happening things on Twitter at a given point in time have been viewed as a chance to create traffic and revenue. Spammers post tweets containing unrelated tweets, harmful links, repeatedly posts treading topics to try to grab attention of tweeters. The more chance it can be exposure to suspects when for a longer time a spam tweet exists. Therefore, it is very essential to discover spam tweets as soon as possible. Real time detection is in demand to decrease the loss caused by spam. For spam detection there are various machine learning techniques which consider the statistical features of tweets. In proposed system, URLs are scanned and analyse using various APIs to detect whether these URLs are malicious or not which can helps to improve the detection of spamming activity in a timely manner.