Effective IOT Attack Detection with Optimal Feature Selection and Random Forest: Performance Enhancement using Spider Monkey Optimization

IoT is a combination of software and hardware that collects and shares data, using networks, sensors, and electronics. IP addresses allow devices to communicate with one other over the internet and exchange data inside the network. The IoT idea is crucial for creating smart surroundings, such as smart cities and houses, applicable to many sectors. Smart environments address living environment, industrial demands, and energy consumption, promoting creativity and competitiveness in everyday life. The growing prominence of IoT applications and services across networks highlights the primary objective. The first paper introduces the SMO-RF technique to detect IoT device threats. Use this optimization strategy to choose attributes related to restrictions. The SMO method combines optimal characteristics for RF categorization during selection. As per the 80-20 rule, features undergo initial classification before being classified in the RF. When applying higher feature values to RF models, the best fitness values are considered. SMO minimizes overfitting by using information to categorization. AUC, recall, F-Measure, accuracy, and precision are used to assess the performance of this approach. NSL-KDD is freely accessible and used with SMO-RF. The SMO-RF approach identifies Probe, U2R, R2L, and DoS attacks more accurately than current IoT solutions.