Enforcing security and resilience in a cloud platform is an essential but challenging problem due to the presence of large number heterogeneous applications running on shared resources. A analysis system that can detect threats or malware must exist inside infrastructure. Much research has been done machine learning-driven analysis, it limited computational complexity detection accuracy. To overcome these drawbacks, we proposed new based concept clustering trend micro locality sensitive hashing (TLSH). We used Cuckoo sandbox, which provides dynamic reports files by executing them isolated environment. novel feature extraction algorithm extract features from obtained sandbox. Further, most important are selected using principal component (PCA), random forest, Chi-square selection methods. Subsequently, experimental results for non-clustering approaches three classifiers, including Decision Tree, Random Forest, Logistic Regression. The model performance shows better classification accuracy false positive rate (FPR) as compared state-of-the-art works approach at significantly lesser computation cost.
- Rahul Kumar
Indian Institute of Technology Bhubaneswar
- Kamalakanta SethiCorresponding
Indian Institute of Technology Bhubaneswar
- Nishant Prajapati
National Institute of Technology Rourkela
- Rashmi Ranjan Rout
National Institute of Technology Warangal
- Padmalochan Bera
Indian Institute of Technology Bhubaneswar