dharani211/Machine-learning-algorithms-for-detecting-network-attacks-with-UNSW-NB15-data-set

Due to the increasingly development of network technology recently, there are various cyber-attacks posed the huge threats to different fields around the world. Many studies and researches about cyber-security are carried out by experts in order to construct a safe network environment for people. The aim of the work is to build the detection models for classifying the attack data. Hence, we applied the UNSW-NB15 network data set which combines both normal and modern low-level attacks because we would like to create the experimental scenario close to the real world. Two classifiers are logistic regression and decision tree model for binary classification in the work. The deployed technique for decision tree achieved the highest result with 99.99% of testing accuracy compare to the 78.15% of logistic regression classifier. On the other hand, the KNN model is used for categorizing the multi-class in the project, and the averaged accuracy for testing is around 23% for ten categories classification.

Machine-learning-algorithms-for-detecting-network-attacks-with-UNSW-NB15-data-set

Due to the increasingly development of network technology recently, there are various cyber-attacks posed the huge threats to different fields around the world. Many studies and researches about cyber-security are carried out by experts in order to construct a safe network environment for people. The aim of the work is to build the detection models for classifying the attack data. Hence, we applied the UNSW-NB15 network data set which combines both normal and modern low-level attacks because we would like to create the experimental scenario close to the real world. Two classifiers are logistic regression and decision tree model for binary classification in the work. The deployed technique for decision tree achieved the highest result with 99.99% of testing accuracy compare to the 78.15% of logistic regression classifier. On the other hand, the KNN model is used for categorizing the multi-class in the project, and the averaged accuracy for testing is around 23% for ten categories classification.

The details of the UNSW-NB15 data set are published in following the papers:

Moustafa, Nour, and Jill Slay. "UNSW-NB15: a comprehensive data set for network intrusion detection systems (UNSW-NB15 network data set)."Military Communications and Information Systems Conference (MilCIS), 2015. IEEE, 2015. Moustafa, Nour, and Jill Slay. "The evaluation of Network Anomaly Detection Systems: Statistical analysis of the UNSW-NB15 data set and the comparison with the KDD99 data set." Information Security Journal: A Global Perspective (2016): 1-14.

Find the dataset here: https://www.unsw.adfa.edu.au/australian-centre-for-cyber-security/cybersecurity/ADFA-NB15-Datasets/