Dataset for: Detection and Classification of Physiological Activation States Through Biosignal Analysis Using Machine Learning
Year: 2026
Type: dataset
Abstract: This dataset contains all the data required for reproducing the experimentation of the four experiments presented in the paper entitled "Detection and Classification of Physiological Activation States Through Biosignal Analysis Using Machine Learning" published in MDPI Sensors journal. More precisely, the dataset contains the following files: EXP1_EXP2_features.mat: this is the matlab .mat matrix with the features extracted from the physiological signals for experiments 1 and 2. For experiment 1, the features used come only from the ECG. For experiment 2, they also include those extracted from the respiration signal. Each row also contains the corresponding classification label for that time window or instance. PER_SUBEJCT_WESAD_raw_signals_norm_allSubjects.mat: this file is required to run the deep-learning part of the experimentation for both experiments 3 and 4. It contains the standardized raw signals (ECG and RESP) for all the 15 subjects. Due to its 1GB weight and the impossibility to upload it to Zenodo (because of unknown reasons), this file has been split into 9 compressed .zip files. To get it, please gather together the 9 compressed .zip files and extract them. All_ECR_9.csv: this file contains the 9 handcrafted features extracted from the physiological signals that are used for the training of the hybrid MST-Conformer architecture of experiment 3. Each row also contains the corresponding classification label for that time window or instance. EXP4_features.mat: this is the matlab .mat matrix with the features extracted from the physiological signals for the traditional machine learning part of experiment 4. Each row also contains the corresponding classification label for that time window or instance. All_ECR_27.csv: this file contains the 27 handcrafted features extracted from the physiological signals that are used for the training of the hybrid MST-Conformer architecture of experiment 4. Each row also contains the corresponding classification label for that time window or instance.
Language: English
Topic: Emotion and Mood Recognition
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7221583348
