Detecting Spam Bots on Social Network

Authors

  • A. Gnanasekar
  • T. Thangam
  • S. Afraah Mariam
  • K. Deepika
  • S. Dhivya Shree

DOI:

https://doi.org/10.47059/revistageintec.v11i2.1719

Abstract

Bots have made an appearance on social media in a variety of ways. Twitter, for instance, has been particularly hard hit, with bots accounting for a shockingly large number of its users. These bots are used for nefarious purposes such as disseminating false information about politicians and inflating celebrity expectations. Furthermore, these bots have the potential to skew the results of conventional social media research. With the multiple increases in the size, speed, and style of user knowledge in online social networks, new methods of grouping and evaluating such massive knowledge are being explored. Getting rid of malicious social bots from a social media site is crucial. The most widely used methods for identifying fraudulent social bots focus on the quantitative measures of their actions. Social bots simply mimic these choices, leading to a low level of study accuracy. Transformation clickstream sequences and semi-supervised clustering were used to develop a new technique for detecting malicious social bots. This method considers not only the probability of user activity clickstreams being moved, but also the behavior's time characteristic. The detection accuracy for various kinds of malware social bots by the detection technique assisted transfer probability of user activity clickstreams will increase by a mean of 12.8 percent, as per results from our research on real online social network sites, compared to the detection method funded estimate of user behaviour.

Downloads

Published

2021-06-02

Issue

Section

Articles