Machine learning approach for mitigating security threats in iot environment

Abstract

Various entities of the human community like healthcare, smart newlinehome, smart grid etc., integrated with the internet, and thus iot emerged. in newlinethis modern era, vast amount of facts and information has been transferred via newlinewireless networks. such networks give rise to innumerable cyber threats and newlineprivacy issues. the aberrant activities which are pernicious to the network are newlinetermed intrusions. the iot networks are pervasive and thus opens to many newlinesecurity threats. if the iot framework suffers from cyber threats, it leads to newlineinformation loss or data loss and sluggishness in iot devices. intrusion newlinedetection system has been utilized for the past two decagons to secure the newlineinformation and networks. as iot has distinct standards and protocol stacks, newlinethe intrusions in iot can not be recognized by classical intrusion detection newlineapproaches. as iot generates boundless data, it is strenuous to scrutinize newlinethem frequently. an intrusion detection system acts as the surveillance to newlinedefend the system or network from attacks.machine learning approaches are excellent and influential techniques to resolve these unusual threats. a potent machine learning algorithm is deployed to obtain attack-free networks. aiming to solve thesecurity issues in the network layer of iot system, we propose a novel intrusion detection scheme that consists of three phases, namely,1. optimal feature vector selection (ofvs)2. bi-layer intrusion detection system (blid) newline3. distributed training in fog node (dt-fn)various entities of the human community like healthcare, smart home, smart grid etc., integrated with the internet, and thus iot emerged. in this modern era, vast amount of facts and information has been transferred via wireless networks. such networks give rise to innumerable cyber threats and privacy issues. the aberrant activities which are pernicious to the network are termed intrusions. the iot networks are pervasive and thus opens to many security threats. if the iot framework suffers from cyber threats, it leads to newlineinformation loss or data loss