Metaheuristic Algorithm based Anomaly Detection Using Behavioral Analysis

Abstract

Computer Security is a vast field of computer science that employs different methods to ensure data and system safety. Intrusion Detection System (IDS) is one of the common mechanisms employed for the same. It is a software tool that can detect unauthorized access to a network or computer system. Anomaly detection is one of the analysis and detection algorithm used in IDS. Behavioral biometrics is a field of study that measures the uniquely identifiable and measurable patterns in human activities. Keystroke Dynamics (KD) is a behavioral biometric verification method. Based on the manner and rhythm of typing on a keyboard, this automated keystroke dynamics method confirms or identifies the identity of an individual. When a user is typing on a keyboard, the detailed timing information is captured. This information exactly describes when a key is pressed and when it was released by the user. Keystroke dynamics is also known as keystroke biometrics or typing dynamics. For future authentication, a user s unique biometric template is developed by measuring the speed of typing, keystroke rhythms and the typing pattern on a keyboard. It helps in discriminating users and thereby helps in identifying imposters (both insiders and external attackers). With keystroke dynamics, imposters attempt to authenticate using a compromised password could be detected and rejected because their typing rhythms differ significantly from those of genuine user. newlineThis thesis proposes a metaheuristic algorithm based anomaly detection using behavioral analysis. A methodology for authenticating the legal users based on the Keystroke Dynamics using a Multilayer Perceptron Neural Network (MLP-NN) and Most Valuable Player Algorithm (MVPA). In a password-based authentication system, when the password and username match with the system database, the secure application allows the user to access it. In addition to the password matching, the Keystroke Dynamics based Authentication (KDA) system validates the user by their typing rhythm of the password