LightGBM-Powered Solutions for Backdoor Malware Detection in SCADA Networks

Different backdoor malware as an issue is one of the most important ones in the domain of SCADA network security, as the presence of modern cyber threats is very diverse. This research compared LightGBM, a gradient boosting framework, to a number other machine learning algorithm for the identification of backdoor malware, these include Random Forest, Support Vector Machine (SVM), XGBoost and Neural Networks as a benchmark for comparison. The results of the study are measured against Kaggle’s network traffic, system call records, and file system interaction datasets; Performance parameters include accuracy, precision, recall, F1 Score, and the training time taken by each algorithm. It can observed that, LightGBM has highest accuracy, precision, recall and F1 Score than all the other five algorithms along with the lowest value of time required for training. Due to high performance and fast execution, LightGBM can be considered as an appropriate solution for the real-time malware detection in SCADA systems. LightGBM’s detection capabilities contrast similarly well along a resource efficiency gradient, as demonstrated by the research, which implies its potential as a tool for improving the cybersecurity situation in critical infrastructure.