Network Traffic Analysis using Feature-Based Trojan Detection Method

Malware assault cases are currently steadily rising in both the private and public sectors. To analyzed the behavior features of trojan based on the characteristics of host and semantics of code. These techniques are having several drawbacks examining numerous common Trojans’ network behavior, characteristics, and network traffic. In this study, to train the Trojan Detection Algorithm, Random Forest Algorithm, Naïve Bayes Algorithm, and Decision Tree Algorithms are used. The classification and evaluation of performance process are carried out using WEKA. The features of Trojan behavior and communication are extracted using a model. Then locate and catch the traffic of Trojans. The accuracy is up to 98%. The proposed Trojans’ detection model when compared with different machine learning algorithms, experiment shows that proposed algorithm is beneficial and effective for detecting Trojans.