Intelligent Attack Detection Modelin lotusing Optimal Feature Selectionincorporated withOptimized DeepLearningArchitecture
Main Article Content
Attacks and problem recognition within the Internet of Things (IoT) foundation is an increasing worry in the environment of IoT.Within the expanded utilization of IoT framework in each space, dangers and assaults in these foundations are additionally developing comparably. Some attacks like node tampering, malicious code injection, malicious node injection, DoS Attacks, Malicious scripts such assaults and inconsistencies which can cause an IoT framework failure. In this study deep learning architecture models were compared to attacks and anamolies on the IoT frameworks that could be predicted.Themainphases ofthe proposedanomalyorattackdetection model are Feature Extractionand Detection. Infeatureextractionprocess,the attributesinthe datasetisconsideredasfeatures,which arelike SystemID,SystemAddress,SystemType,SystemLocation,DestinationService Address,DestinationServiceType,DestinationLocation,Accessed NodeAddress,Accessed Node Type, Operation, Value, Timestamp, and Normality. Since thenumbersoffeatureare high,which increasesthe lengthofthefeature vector, optimalfeatureselectionprocesswillbedevelopedtoextractthe mostsignificantfeatures withuni.queinformation.Finally,theoptimizedDBN(Deep Belief Network)categorizesthedataintonormalas wellasattacks,anddetectseachcategoryofattacks.