K_FOLD = 10
subjects = range(4, 6)
subjects_set = set(subjects)
accuracies = {
"GNB": np.zeros((len(subjects), K_FOLD)),
"SVM": np.zeros((len(subjects), K_FOLD)),
"LDA": np.zeros((len(subjects), K_FOLD))
}
for subject in subjects:
print("========= Subject: ", subject)
# Load data
print("Loading data ...")
left_data_file = f"{DATA_FOLDER}/{subject}-left.csv"
right_data_file = f"{DATA_FOLDER}/{subject}-right.csv"
data = read_eeg_file(left_data_file, right_data_file, TIME_LENGTH,
TIME_WINDOW, EPOCH_SIZE)
# Pre-processing
print("Pre-processing ...")
print("Applying 5º order Butterworth bandpass filter (7-30 Hz)")
b, a = signal.butter(5, [7, 30], btype="bandpass", fs=FS)
data.left_data = signal.filtfilt(b, a, data.left_data, axis=1)
data.right_data = signal.filtfilt(b, a, data.right_data, axis=1)
print("Spatial-filtering ...")
data.X = np.concatenate((data.left_data, data.right_data))
csp = CSP(average_trial_covariance=True, n_components=CSP_COMPONENTS)
csp.fit(data.left_data, data.right_data)
data.Z = np.array([csp.project(x) for x in data.X])
print("D: ", D)
subjects = range(1, 6)
accuracies = {"SVM": np.zeros(len(subjects)), "LDA": np.zeros(len(subjects))}
# Create the Filter Bank
filter_bank = FilterBank(4, 40, 4)
for subject in subjects:
print("========= Subject: ", subject)
# Load training data
print("Loading training data ...")
left_data_file = f"{DATA_FOLDER}/left-hand-training-subject-{subject}.csv"
right_data_file = f"{DATA_FOLDER}/right-hand-training-subject-{subject}.csv"
training_data = read_eeg_file(left_data_file, right_data_file, TIME_LENGTH,
TIME_WINDOW, EPOCH_SIZE)
# Load test data
print("Loading test data ...")
left_data_file = f"{DATA_FOLDER}/left-hand-test-subject-{subject}.csv"
right_data_file = f"{DATA_FOLDER}/right-hand-test-subject-{subject}.csv"
test_data = read_eeg_file(left_data_file, right_data_file, TIME_LENGTH,
TIME_WINDOW, EPOCH_SIZE, False)
# Apply the filter bank to the training data
training_data.left_data = filter_bank.apply(training_data.left_data)
training_data.right_data = filter_bank.apply(training_data.right_data)
# Compute the spatial and spectral covariances
spatio_spectral_covariances = SpatioSpectralCovariancesEstimation(
training_data)
