Authenticity of System Users via Mouse Handling Method

Abstract

The proposed work narrate a behavioural bio-metric approach to confirm authenticated users dynamically based on their mouse motion. A self–generated mouse data [9] was used to extract features to categorize the user’s mouse handling pattern which is different from other users. The model built is trained using the Gaussian Naive Bayes Classifier for quick and accurate classification of data. The proposed model performs better than previously used models in all evaluation metrics including, accuracy, false accept rate, false reject rate.