Analysis of the Combination of Naïve Bayes and MHR (Mean of Horner’s Rule) for Classification of Keystroke Dynamic Authentication

Abstract

Keystroke Dynamics Authentication (KDA) is a technique used to recognize somebody dependent on typing pattern or typing rhythm in a system. Everyone's typing behavior is considered unique. One of the numerous approaches to secure private information is by utilizing a password. The development of technology is trailed by the human requirement for security concerning information and protection since hacker ability of information burglary has gotten further developed (hack the password). So that hackers can use this information for their benefit and can disadvantage others. Hence, for better security, for example, fingerprint, retina scan, et cetera are enthusiastically suggested. But these techniques are considered costly. The advantage of KDA is the user would not realize that the system is using KDA. Accordingly, we proposed the combination of Naïve Bayes and MHR (Mean of Horner’s Rule) to classify the individual as an attacker or a nonattacker. We use Naïve Bayes because it is better for classification and simple to implement than another. Furthermore, MHR is better for KDA if combined with the classification method which is based on previous research. This research showed that False Acceptance Rate (FAR) and Accuracy are improving than the previous research.

Keywords

• keystroke dynamic authentication
• classification
• naã¯ve bayes
• mean of horner’s rule