An Evaluation of Selection Method in the Classification of Scada Datasets Based on the Characteristics of the Data and Priority of Performance
Jaime Yeckle, Sherif Abdelwahed · 2017
Industrial control systems are becoming more complex in their architecture and design, and consequently many security concerns have arisen. To face these challenges, datasets of the systems are analyzed to detect intrusions. However, many techniques of preprocessing and classifiers are available with different approaches, which makes difficult the selection of suitable method. This work proposes a new approach to choose a preprocessing method and a classification algorithm based on the characteristics of the dataset and priority in the performance; the approach use two popular techniques of preprocessing and ten machine learning algorithms. We evaluate the method using three real datasets from different domains. The datasets are publicly available and can be used as a benchmark. The results show the feasibility of the new method of selection and the effectiveness of the use of machine-learning algorithms in SCADA datasets. This is a preliminary research and based on these initial results, a future work will propose practical implications for the design of Intrusion detection systems (IDS) on cyber security systems considering the data type as an enhancement to existing designs.