Reliable Machine Learning Model for IIoT Botnet Detection

Due to the growing number of Internet Things (IoT) devices, network attacks like denial service (DoS) and floods are rising for security reliability issues. As a result these attacks, IoT devices suffer from disruption. Researchers have implemented different techniques identify aimed at vulnerable devices. In this study, we propose novel features selection algorithm FGOA-kNN based on hybrid filter wrapper approaches select most relevant features. The approach integrated with clustering rank then applies Grasshopper (GOA) minimize top-ranked Moreover, proposed algorithm, IHHO, selects adapts neural network’s hyper parameters detect botnets efficiently. Harris Hawks is enhanced three improvements improve global search process optimal solutions. To tackle problem population diversity, chaotic map function utilized initialization. escape energy hawks updated new nonlinear formula avoid local minima better balance between exploration exploitation. Furthermore, exploitation phase HHO using elite operator ROBL. model combines unsupervised, clustering, supervised intrusion behaviors. N-BaIoT dataset validate model. Many recent were used assess compare model’s performance. demonstrates that than other variations detecting multiclass botnet attacks.