Cyber Crime Detection and Prevention Techniques on Cyber Cased Objects Using SVM and Smote

Abstract

Conventional cybersecurity employs crime prevention mechanisms over distributed networks. This demands crime event management at the network level where Detection and Prevention of cybercrimes is a must. A new Framework IDSEM has been introduced in this paper to handle the contemporary heterogeneous objects in cloud environment. This may aid for deployment of analytical tools over the network. A supervised machine learning algorithm like SVM has been implemented to support IDSEM. A machine learning technique Like SMOTE has been implemented to handle imbalanced classification of the sample data. This approach addresses imbalanced datasets by oversampling the minority classes. This will help to solve Social Engineering Attacks (SEA) like Phishing and Vishing. Classification mechanisms like decision trees and probability functions are used in this context. The IDSEM framework could minimize traffic across the cloud network and detect cybercrimes maximally. When results were compared with existing approaches, the results were found to be good, leading to the development of a unique SMOTE algorithm.