Comparison of twitter spam detection using various machine learning algorithms

Abstract

Online Social Networks(OSNs) have mutual themes such as information sharing, person-to-person interaction and creation of shared and collaborative content.  Lots of micro blogging websites available like Twitter, Instagram, Tumblr. A standout amongst the most prominent online networking stages is Twitter. It has 313 million months to month dynamic clients which post of 500 million tweets for each day. Twitter allows users to send short text based messages with up to 140-character letters called "tweets". Enlisted clients can read and post tweets however the individuals who are unregistered can just read them. Due to the reputation it attracts the consideration of spammers for their vindictive points, for example, phishing true blue clients or spreading malevolent programming and promotes through URLs shared inside tweets, forcefully take after/unfollow valid clients and commandeer drifting subjects to draw in their consideration, proliferating obscenity. Twitter Spam has become a critical problem nowadays. By looking at the execution of an extensive variety of standard machine learning calculations, fundamentally expecting to distinguish the acceptable location execution in light of a lot of information by utilizing account-based and tweet content-based highlights.