Secure Trust-Based Attribute Access Control Mechanism Using FK-MFCMC and MOEHO-XGBOOST Techniques

The Internet of things (IoT) possesses various devices, which are switching data continuously and are acquired throughout via lossy networks. This promotes the necessity for a lightweight, flexible along with adaptive access control (AC) method to handle the insidious behavior of such global ecological units and also guarantees the reliability betwixt the trusted devices. To provide users with adaptive authentication centered on dynamic trust estimation, a secure trust along with attribute-based AC (ABAC) approach is invented in this research. First, Trust Evaluation (TE) is executed utilizing Fisher score Kernel-based Minkowski fuzzy C-Means Clustering (FK-MFCMC) centered on trusted data balance by means of instructive features being extricated. A successful association is kept up by TE betwixt the nodes installed in the network. And, it guarantees that the entire nodes work in a reliable form. Subsequently, to ingress the request, the TE makes a decision utilizing the multi-objective elephant herding optimization based on an extreme gradient boosting (MOEHO-XGBoost) methodology. In this methodology, the ABAC request is converted into a permission decision vector. Then, it is converted into a binary classification problem that checks whether to permit or reject access. On the whole, the proposed methodology attains higher trust accuracy and TE for several nodes in IoT together with obtains a higher security level in correlation with the prevailing novel methodologies.