Decision Tree Based Data Pruning with the Estimation of Oversampling Attributes for the Secure Communication in IOT

Abstract

Internet of Things (IoT) exhibits a significant role to evaluate the error or supply shortage. The IoT demand for the security and authentication of the devices is considered as the most priority for software developments. As the IoT communication comprises of an interconnected environment for the both digital and physical scenarios. The IoT environment exhibits anything and anywhere services to the communication medium. In those scenarios, security is considered as the major concern to protect the data resources from unauthorized resources for the appropriate security and privacy. This paper proposed a decision tree-based pruning scheme for the IoT attributes. The proposed decision tree based pruning for the security attributes are defined as the decision tree pruning (DTP). The proposed DTP model comprises of the minority oversampling model for the estimation of the attack features. With the developed DTP model, the attack datasets were pre-processed and evaluated for the different attack environments in to consideration. The DTP processed data were applied over the conventional machine learning-based model for the computation attacks in the network. The simulation results expressed that proposed DTP model achieves the accuracy value of 98% which is ~3% higher than the conventional classifier techniques.