An Effective Detection of Version Number Attacks in the IoT using Neural Networks

Day by Day human dependency on connected technology is rapidly growing with the deployment of IoT based real-time sensing applications that serve various business or human requirements. Therefore, there are various researches to secure connected devices like IoT from a wide range of attacks like Version Number Attack. Detection of Version Number Attack in the network infrastructure is the impart part of the cyber security to watch and block the malicious traffic flows on the IoT network. To overcome this major issue, proposing a machine learning and deep learning approach. First, implement a qualitative feature extraction approach based on filter technique that are independent of any classifiers. Therefore, the selected features are not dependent on any Machine Learning and Deep Learning algorithm. Next, implemented a SMOTE Oversampling technique to overcome the imbalanced dataset problem. The proposed approach has greater benefits in terms of Higher ROC and can be effectively implemented in a real world complex IoT network as a unified framework for detecting Version Number attacks.