Cyber-attacks are exponentially increasing daily with the advancements of technology. Therefore, the detection and prediction of cyber-attacks are very important for every organization that is dealing with sensitive data for business purposes. In this paper, we present a framework on cyber security using a data mining technique to predict cyber-attacks that can be helpful to take proper interventions to reduce the cyber-attacks. The two main components of the framework are the detection and prediction of cyber-attacks. The framework first extracts the patterns related to cyber-attacks from historical data using a J48 decision tree algorithm and then builds a prediction model to predict the future cyber-attacks. We then apply the framework on publicly available cyber security datasets provided by the Canadian Institute of Cybersecurity. In the datasets, several kinds of cyber-attacks are presented including DDoS, Port Scan, Bot, Brute force, SQL Injection, and Heartbleed. The proposed framework correctly detects the cyber-attacks and provides the patterns related to cyber-attacks. The overall accuracy of the proposed prediction model to detect cyber-attacks is around 99%. The extracted patterns of the prediction model on historical data can be applied to predict any future cyber-attacks. The experimental results of the prediction model indicate the superiority of the model to detect any future cyber-attacks.
|Title of host publication||The 15th IEEE Conference on Industrial Electronics and Applications (ICIEA2020)|
|Place of Publication||Kristiansand, Norway|
|Publication status||Accepted/In press - 25 May 2020|