A benchmarking analysis of artificial intelligence in handling data imbalance and DDoS assaults in WSN

A Wireless Sensor Network (WSN) network comprises multiple sensor nodes that collect sensitive data. These nodes connect with one another and provide information to the central station. During these communications, malicious nodes employ Distributed-Denial-of-Service (DDoS) assaults, resulting in server disruptions, monetary loss, and unnecessary stress on IT experts. To classify these assaults, we employed the text-based dataset “Wireless Sensor Network Blackhole Flooding Selective Forwarding (WSNBFSF)”. They are primarily collected from Kaggle’s websites and are publicly accessible. We use data encoding, linear regression to deal with null values, principal component analysis to deal with features, and Synthetic Minority Oversampling Technique (SMOTE) to balance data; the WSN-DS dataset was also used to compare the efficacy of the model. For the classification of DDoS attacks, various machine learning, ensemble learning, and deep learning models are utilized. The evaluation is based on the WSNBFSF dataset. The results indicate that Decision Tree outperforms all other classification methods with a 99.96% accuracy rate. Additional evaluation of the results highlights the future application potential of this method.