"Processed ORNL-MSU Dataset for Intrusion Detection (Binary and Multi-Class Classification)"

"This dataset is derived from the publicly available ORNL\u2013MSU intrusion detection dataset and is prepared to support reproducible research in machine learning-based intrusion detection systems. It includes processed data used in our experimental study for both binary and multi-class classification tasks.The dataset consists of 30 sub-datasets in total: 15 for binary classification and 15 for multi-class classification. Each sub-dataset is provided with predefined training and testing splits (90% training and 10% testing, using a fixed random state for reproducibility). Additionally, the selected feature subsets for each dataset are included, enabling direct model training without the need for computationally expensive feature selection procedures.Preprocessing steps applied to the data include label encoding (normal = 0, attack = 1 for binary tasks), removal of missing and infinite values, feature scaling using MinMaxScaler, and independent feature selection for each dataset.The original ORNL\u2013MSU dataset remains publicly accessible, while this repository provides only the processed and structured data to facilitate efficient benchmarking, fair comparison, and reproducibility of intrusion detection models."