Analysis of Network log data using Machine Learning

The proliferation of web base usage has also resulted in an escalation in unauthorized network access. In this scenario, it is imperative to periodically analyze log records of the network so that malicious users can be identified. This process can be automated using machine learning techniques. In this paper, analysis of log records of a network is carried out using supervised machine learning techniques. K-means and Self organizing feature map (SOFM) algorithms have been used with the data set obtained from the UCI machine learning repository. An accuracy of 97.2% has been archived.