Supervised Classifier Approach for Intrusion Detection on KDD with Optimal MapReduce Framework Model in Cloud Computing

Abstract

Background: Cloud computing is characterized as the arrangement of assets or accessible administrations by the cloud service providers through web to their clients. It communicates everything as administrations over the web as per the client request, for example, the operating system, organization of equipment, storage, assets, and software. Nowadays, Intrusion Detection Systems (IDS) play a powerful role while working under the influence of the experts who act when a system is hacked or under some intrusions. Most intrusion detection frameworks are created based on the machine learning strategies. Since the datasets play a major role, this is utilized as part of the intrusion detection i.e., Knowledge Discovery in Database (KDD). Methods: In this paper, the intruded data was detected and classified utilizing Machine Learning (ML) with MapReduce model. The primary objective of the Hadoop MapReduce model is to reduce the extent of database ideal weight that was decided for reducer model and second stage by utilizing Decision Tree (DT) classifier in data detection. This DT classifier utilizes an appropriate classifier to decide the class labels for non-homogeneous leaf nodes. The decision tree fragment provided a coarse section profile while the leaf level classifier yielded the data about the qualities that influence the label inside a portion. Results: From the proposed results, the accuracy for detection was 96.21% in comparison with the existing classifiers, for example, Neural Network (NN), Naive Bayes (NB) and K Nearest Neighbor (KNN). Conclusion: This study introduced a Hadoop Map-reduce model to create diverse mappers and diminish the data utilizing OBL-GWO strategy.