An Efficient Stabbing Based Intrusion Detection Framework for Sensor Networks

摘要

Intelligent Intrusion Detection System (IIDS) for networks provide aresourceful solution to network security than conventional intrusion defencemechanisms like a firewall. The efficiency of IIDS highly relies on the algorithmperformance. The enhancements towards these methods are utilized to enhancethe classification accuracy and diminish the testing and training time of thesealgorithms. Here, a novel and intelligent learning approach are known as the stabbingof intrusion with learning framework (SILF), is proposed to learn the attackfeatures and reduce the dimensionality. It also reduces the testing and trainingtime effectively and enhances Linear Support Vector Machine (l-SVM). It constructsan auto-encoder method, an efficient learning approach for feature constructionunsupervised manner. Here, the inclusive certified signature (ICS) isadded to the encoder and decoder to preserve the sensitive data without beingharmed by the attackers. By training the samples in the preliminary stage, theselected features are provided into the classifier (lSVM) to enhance the predictionability for intrusion and classification accuracy. Thus, the model efficiency islearned linearly. The multi-classification is examined and compared with variousclassifier approaches like conventional SVM, Random Forest (RF), RecurrentNeural Network (RNN), STL-IDS and game theory. The outcomes show thatthe proposed l-SVM has triggered the prediction rate by effectual testing andtraining and proves that the model is more efficient than the traditional approachesin terms of performance metrics like accuracy, precision, recall, F-measure, pvalue,MCC and so on. The proposed SILF enhances network intrusion detectionand offers a novel research methodology for intrusion detection. Here, the simulationis done with a MATLAB environment where the proposed model shows abetter trade-off compared to prevailing approaches.