Leveraging AI and ML for Predictive Analysis and Feature Attribution in Abnormal Network Behavior Detection
Shridhar Allagi, Toralkar Pawan, Kavita Mainalli, Nagaraj V. Dharwadkar · 2024
India has witnessed 15% increase in cyber-attacks in 2023 since 2022. Aadhar data breach is one of example among many cyber-attacks in India that cause millions of citizens to lose the personal details such as name, phone number, address, Aadhar number to hackers. Covid 19 of 500 million people were leaked by hacker is the second example faced by Indian organization in medical domain. There was a need to overcome the limitation of current security method organization have imposed Artificial Intelligence (AI) and Machine Learning (ML) to stop the cyber-attacks on Indian organization. With AI and ML model observe the patterns in the data and predict the trends in data and with continuous monitoring identifies the abnormal attacks. The aim of this paper is to find the contribution of attributes / features in dataset towards prediction of abnormal behavior in the network. Model uses SelectKBest() method to select the best feature among the other attributes. These attributes act as inputs to support vector machine that uses linear kernel to predict the attacks. The model archives an accuracy of 97% and Shapley additive explanations is used to explain feature contribution.