EFFICIENT NETWORK THREAT DETECTION AND CLASSIFICATION METHOD USING SVM+NB ALGORITHM ON CLOUD COMPUTING

In this work, NSL-KDD dataset is considered for processing the session flow and to analyze threat which is related to files in the cloud. Also, feature extraction is performed based on two essential features like high-level features and low-level features which are considered for classification purposes. Here, three effectual classifiers like efficient Naive Bayes (NB) classifier, Improved Support Vector Machine (ISVM) are utilized over the features extracted from the dataset for intrusion detection. Empirical outcomes depict that the selected features will offer a better outcome to design the threat detection system which will be more efficient for cloudbased network security. Results acquired from these methods show better trade-off in comparison with existing systems like decision tree, random forest and so on. Performance metrics like precision with 100%, accuracy with 99.5%, specificity with 100%, Recall with 100% and F measure with 100%has been evaluated with this machine learning algorithms.