Improving security in malware detection using aggregate algorithms

In the age of communication and the digital world, the security of computer systems is considered one of the most controversial security issues. The upcoming research is trying to extract useful data from the Microsoft malware dataset named BIG 2015, a classifier that is very simple and has little computational complexity both in the field of feature extraction and in the field of the classifier mechanism. It provides security and malware detection. Of the 1804 extracted features, some of which have played a more important and colorful role in the classification, the section_name_headre feature has been calculated with a weight of 0.2160. The accuracy of the classifier is 99.81 and the predictor error is 0.00774. In this regard, in order to achieve better predictions and higher accuracy than the aggregate algorithm and methods of selecting suitable features from the data sets used, the techniques of Feature selection, Feature Importance Xgboost & Lgb, and Permutation Importance have been used. Therefore, by using the findings of this research in IDS and IPS systems, it is possible to increase the accuracy of malware detection and reduce the detection error rate.