def fit(self, X, y):
from mac import MAC
mac_ = MAC(th=self.th)
self.noise_est = mac_.fit_transform(X, y)
return
###Paths for testing only
#filename = r'G:\Shared drives\datasets\BCI\Competition IV\dataset 2a\Trials\NEW_22ch_A01.mat'
#filename = r'G:\My Drive\Students\vigomez\Code_A1_Application\data_4C\BCI_s02train.mat'
#filename = '..\data_4C\BCI_s01train.mat'
data = sio.loadmat(filename)
Xdata = data['X']
Xdata = np.transpose(Xdata, (2, 1, 0))
labels = data['labels'].reshape(-1, )
fs = 250
print('Loading', filename, 'with sampling frequency of', fs, 'Hz.')
# 2. Artifacts removal stage - Noise Estimation
from mac import MAC
mac_ = MAC(th=args.th)
noise_est = mac_.fit_transform(Xdata, labels)
Noise_sum = np.array([np.sum(noise) for noise in noise_est])
noise_index = np.where(Noise_sum != 0.0)[0]
Xdata = Xdata[noise_index]
Noise = noise_est[noise_index]
labels = labels[noise_index]
#3. Parameter Grid for AMK AF
param_dist = {
'embedding': randint(5, 10),
'eta': loguniform(1e-2, 0.5),
'epsilon': uniform(1e-1, 2),
'mu': uniform(1e-2, 1),
"Ka": randint(5, 15)
