Time Efficiency on Computational Performance of PCA, FA and TSVD on Ransomware Detection

Benni Purnama (Faculty of Computer Science, Universitas Dinamika Bangsa)
Deris Stiawan (Department of Computer Engineering, Universitas Sriwijaya, Indonesia)
Darmawijoyo Hanapi (Mathematics Education, Sriwijaya University, Indonesia)
Mohd. Yazid Idris (School of Computing, Faculty of Engineering, Universiti Teknologi Malaysia)
Sharipuddin Sharipuddin (Faculty of Computer Science, Universitas Dinamika Bangsa)
Nurul Afifah (Department of Computer Engineering, Universitas Sriwijaya, Indonesia)
Rahmat Budiarto (College of Computer Science and Information Technology, Al Baha University)

Indonesian Journal of Electrical Engineering and Informatics (IJEEI)
Vol 10, No 1: March 2022

Ransomware is able to attack and take over access of the targeted user'scomputer. Then the hackers demand a ransom to restore the user's accessrights. Ransomware detection process especially in big data has problems interm of computational processing time or detection speed. Thus, it requires adimensionality reduction method for computational process efficiency. Thisresearch work investigates the efficiency of three dimensionality reductionmethods, i.e.: Principal Component Analysis (PCA), Factor Analysis (FA) andTruncated Singular Value Decomposition (TSVD). Experimental results onCICAndMal2017 dataset show that PCA is the fastest and most significantmethod in the computational process with average detection time of 34.33s.Furthermore, result of accuracy, precision and recall also show that the PCAis superior compared to FA and TSVD.