Comparative Study: Machine Learning Classification Algorithms for Backdoor Detection

Abstract

In the last ten years, malware attacks have become a common crime story online. Nowadays, well-known threats, including viruses, worms, trojans, backdoors, exploits, password stealers, and spyware, have reached millions, and among these threats, the backdoor attack has a high rate of intrusion across global networks around the world. Backdoor attack is a hidden technique is used for getting remote access to a machine or other system that without authentication. In this study ten different supervised learning techniques such as Bayes Net, Bayesian LR, Naives Bayes, Naive Bayes, Multi Layer Perceptron, Lib SVM, K-star, Stacking, Threshold Selection, Randomization filter Classifier and Zero R were employed to achieve the comparative analysis of machine classifier. The performance of the classifier algorithms was rated using Accuracy, Precision, Recall, F-Measure, False Positive Rate and True Positive Rate using WEKA data mining tool. Multi Layer Perceptron was found to be an excellent classifier that gives the best accuracy of 99.97% and a false positive rate of 0.00. The comparative analysis result indicates the achievement of low false-positive rate for backdoor classification which suggests that anti-phishing application developer can implement the machine learning classification algorithm that was discovered to be the best in this study to enhance the feature of Backdoor attack detection and classification.