Offline Features & Models (Pre- & Post-Calibration)

Offline data of both the pre- and post-calibration offline pipeline.

Download the "Offline_Data.zip" zip file, to get access to all data at once.

"v1": first and only version

Note:
Labels: "4sec_sampled_labels.mat" - used for the Base Models and also later for the Calibration. (for feature set splitting into train, val, test) -> resulting in "4sec_train_labels.mat", "4sec_val_labels.mat", "4sec_test_labels.mat".

"AutoPipeline": 1000, 2000, 3000 and 4000 samples (4-sec epochs) as training data

• Base Models: ("training samples, feature configuration, 4-sec, preprocessing type, dataset name, model type")

"AutoCalibration": 1000, 2000, 3000 and 4000 training samples with respective 10, 25, 20, 25 and 30% Calibration Data (new target cross-data percentage of the training data on a base model)

• Offline Calibrated Base Models (Base Models calibrated with other Base Model Data (not subject data!). See: "_w*" --> Calibrated "w" other Base Model.