Improving Accuracy of Authentication Process via Short Free Text using Bayesian Network
The internet security problems are a crucial threat to all users in the cyber world. One of the important problems about internet security concerned with user classification and authentication. However, there are multiple components to classify and authenticate users. The first one is using username/password and the second method is OTP or Token. This paper presents a novel method which cans Classify User via Short-text and IP Model (CUSIM) to grant or reject a user in authentication. CUSIM is a Bayesian network model which utilizes the Bayesian Inference to authenticate the user. The objective of this paper is to use the model based on conditional independent with the prior knowledge, i.e. Keystroke dynamics, location used to connect to the internet, and IP address. Finally, a numerical example is provided to illustrate the probability of incorrect authentication and use an algorithm of machine learning to test the efficiency and find out the accuracy, FAR, and FRR. The model results gave better value of accuracy, FAR, and FRR.
Keywords: Authentication, User Classification, Short Free Text, Keystroke Dynamics, IP Address, Bayesian network.
Download Full-Text
ABOUT THE AUTHORS
Charoon Chantan
Ph. D Candidate.
Sukree Sinthupinyo
Advisor
Tippakorn Rungkasiri
Co-Advisor
Charoon Chantan
Ph. D Candidate.
Sukree Sinthupinyo
Advisor
Tippakorn Rungkasiri
Co-Advisor