A Machine Learning Technique to Detect Behavior Based Malware

Malware has threatened the organizations for a long time and still have not made a lot of progress in detecting the malware on time. Malware can easily harm the system by executing the unnecessary services that will put the load on the system and hinder its smooth running. There are basically two methods to detect the malware, one being the old process of detecting the malware based on the signature and the other one being the Behavior based method. The behavior of the malware is defined by the task the malware performs when it gets activated in the machine, for example, running the Operating System services, downloading the infected files from the internet. The proposed algorithm detects the malware based on its behavior. In this paper, the proposed model is the combination of Support Vector Machine and Principle Component Analysis. This proposed model achieved an accuracy of 97.75% during validation with 97% precision, 99% recall and f1-score of.98 for actual Malwares.