An Efficient Worm Detection System Using Multi Feature Analysis and Classification Techniques

Signature based pattern identification is used to detect the worm hole attack in networks. This identification of signature pattern stores the input data and reduce the overhead problem. The malicious node identification based on the packets dropped during the communication between the users. Here applying machine learning approaches to solve the malicious in networks. In this paper, distinguished behaviors of Transferrable Executable headers, API function calls and DLL files are used. The machine learning algorithm, Adaboost ensemble classifier, Naive Bayesian classifier and Decision tree are applied for improving to detection rate. The performance of each algorithm firstly evaluated to find the most outperformed algorithm each for classifying benign executable and malicious worm executable. And then, these features were combined to detect worm more precisely.