User authentication based on mouse movement data using normalized features

This paper presents a user authentication system based on mouse movement data. An available logging tool named Recording User Input (RUI) is used to collect three types of mouse actions — Mouse Move, Point-and-Click on Left or Right mouse button and Drag-and-Drop. Collected data are divided into N-number of blocks consisting of specific number of actions. From each block seventy four features are extracted to form feature vectors where number of new features is forty eight. Two types of classifiers are used to identify the user: Support Vector Machine (SVM) and Artificial Neural Network (ANN) are used to identify the user. A benchmark data is used to train and test the system. Experimental result shows that for both classifiers system with proposed features perform better. The results also show that Support Vector Machine outperforms the Artificial Neural Network Classifier.