A hybrid-multi filter-wrapper framework to identify run-time behaviour for fast malware detection

摘要: Abstract Malicious software (malware) constitute one of the most pressing cyber threats intended to cripple critical infrastructure, render infected systems unusable, permanently erase data from storage systems. The number malware has skyrocketed through use enormous development toolkit. Run-time analysis recently been used overcome limitations current detection engines due code obfuscation techniques such as polymorphism and metamorphism. However run-time approaches face a challenge processing large features which may fail provide real time protection. In this paper, we propose hybrid framework by using more than complementary filters wrapper feature selection approach identify significant behavioural characteristics malware. novelty proposed is that it exploits within-filters between wrapper-filters hybridizing discriminant, minimum redundant, maximum relevant with integrate knowledge intrinsic behaviour obtained into process. We have verified performance extensive experiments datasets. results show finds When these are in engine, computational performances accuracies also improved up 99 . 499 % compared any existing techniques.