A Machine Learning based Malware Classification Framework

Rapid developments in the field of computer technology introduced new factors such as sophistication, effectiveness, and increased connectivity. The evolution in computer technology involuntarily leads to the evolution of malware as more devices are connected, leading to the development of malware that can spread through interconnected devices. The increase in the use of computers in almost every part of our life, people rely more on them by storing all the crucial details and information on the computers. This will put us in a vulnerable state where cybercriminals can attack us and exploit us using our data against us. These attacks can be performed by injecting various kinds of malware into our computers which puts us at risk. This creates the necessity to discover the different malicious attacks and classify them before it takes place to be safe. Representing the malware in a visual representation makes the process of classification much easy and more efficient. This research study proposes a neural network model to evaluate the performance of different combinations of various methods available for representing malware and different models of convolutional neural networks (CNN) and selects the best model that has highest accuracy.