Volume - 3 | Issue - 1 | march 2021
DOI
10.36548/jismac.2021.1.004
Published
07 April, 2021
Internet users are largely threatened by abuse and manipulation of several automated chat service programs called as chat bots. Malware and spam is distributed by the popular chat networks using chat bots. The commercial chat network is surveyed in this paper with a series of measurements. A series of 15 advanced to simple chatbots are used for this purpose. When compared to the bot behavior, the complexity of human behavior is high. A classification system is proposed for accurate distinguishing between human user and chatbots based on the measurements obtained from the study. Na誰ve Bayes Classifier and entropy classifier are used for the purpose of classification. Chat bot detection is performed with improved efficiency and accuracy using these classifiers. The speed of Na誰ve Bayes Classifier and accuracy of entropy classifier compliments each other in the process of detection of chat bots. The improved efficiency of the proposed system is proved by testing and comparison with the existing schemes.
KeywordsNaive Bayes Classifier Entropy classifier chatbots malware spam classification