A Study on Detection of Malware Using Machine Learning Algorithms

Abstract

One of the most significant issues facing internet druggies currently is malware. Polymorphic malware is a new type of vicious software that's further adaptable than former generations of contagions. Polymorphic malware constantly modifies its hand traits to avoid being linked by traditional hand- grounded malware discovery models. Counter-attacking measures have been more effective, with antivirus companies expanding their signature database, which is routinely updated, although they are inefficient and ineffective in the case of polymorphic malware. To identify vicious pitfalls or malware, we used a number of machine literacy ways. A high discovery rate indicated that the algorithm with the stylish delicacy was named for operation in the system. As an advantage, the confusion matrix measured the number of false cons and false negatives, which handed fresh information regarding how well the system worked. In particular, it was demonstrated that detecting dangerous business on computer systems, and thereby perfecting the security of computer networks, was possible using the findings of malware analysis and discovery with machine literacy algorithms (Naive Byes, SVM,RF, and with the proposed approach) integrals. The results showed that when compared with other classifiers, DT( 99), CNN(97.76), and SVM(94.41) performed well in terms of discovery delicacy. These results are significant, as vicious software is getting decreasingly common and complex. Keywords: CNN, SVM, DT, cybersecurity, cyberattack, cyber warfare, cyber threats, suspicious activity.