def obtain_concatenate_segments_fft(inputData):
X = []
Y = []
for i in range(len(inputData.result.arrays)):
count = len(inputData.result.arrays[i].array_data) / 2
print("length is :" + str(count))
for j in range(int(count)):
valid = True
band_fft_avg1 = get_average_band_power(
inputData.result.arrays[i].array_data[j * 2],
inputData.result.arrays[i].array_validate[j * 2])
band_fft_avg2 = get_average_band_power(
inputData.result.arrays[i].array_data[j * 2 + 1],
inputData.result.arrays[i].array_validate[j * 2 + 1])
x1 = np.asarray(getList(band_fft_avg1))
x2 = np.asarray(getList(band_fft_avg2))
if x1[0] == 0 or x2[0] == 0:
valid = False
if valid:
X.append(np.append(x1, x2))
Y.append(inputData.result.arrays[i].name)
return pd.DataFrame(X), Y
def obtain_TESPAR_S_FFT_features(inputData, encoding, selected_symbols=32):
X = []
Y = []
for i in range(len(inputData.result.arrays)):
for j in range(len(inputData.result.arrays[i].array_data)):
valid = True
band_fft_avg = get_average_band_power(
inputData.result.arrays[i].array_data[j],
inputData.result.arrays[i].array_validate[j])
arr = []
x1 = np.asarray(getList(band_fft_avg))
x2 = np.asarray(
encoding.get_s(
inputData.result.arrays[i].array_data[j],
inputData.result.arrays[i].array_validate[j])).ravel()
if x1[0] == 0:
valid = False
arr.extend(x1)
arr.extend(x2)
# arr.append(x2)
# print(arr)
if valid:
X.append(arr)
Y.append(inputData.result.arrays[i].name)
return pd.DataFrame(X), Y
def obtain_FFT_features_labels(inputData):
X = []
Y = []
for i in range(len(inputData.result.arrays)):
for j in range(len(inputData.result.arrays[i].array_data)):
valid = True
band_fft_avg = get_average_band_power(
inputData.result.arrays[i].array_data[j],
inputData.result.arrays[i].array_validate[j])
# band_fft_max = Welch.get_max_band_power(inputData.result.arrays[i].array_data[j],
# inputData.result.arrays[i].array_validate[j])
x1 = np.asarray(getList(band_fft_avg))
if x1[0] == 0:
valid = False
if valid:
X.append(np.asarray(getList(band_fft_avg)))
Y.append(inputData.result.arrays[i].name)
return pd.DataFrame(X), Y
delta = []
beta = []
theta = []
gamma_hi = []
gamma_lo = []
alpha = []
fast = []
FFT = Welch(doas, [10], ['light'], ['stimulus'], ['all'])
FFT_data, data_validate = FFT.get_data()
for i in range(len(FFT_data)):
# dataset, channels, levels, segment, trial, orientation
# FFT = Welch(doas, [1], ['deep'], ['stimulus'], [i], ['all'])
band_fft_avg = get_average_band_power(FFT_data[i], data_validate[i])
band_fft_max = get_max_band_power(FFT_data[i], data_validate[i])
# delta_avg = delta_avg + band_fft_max['Delta']
# beta_avg = beta_avg + band_fft_max['Beta']
# theta_avg = theta_avg + band_fft_max['Theta']
# alpha_avg = alpha_avg + band_fft_max['Alpha']
# gamma_hi_avg = gamma_hi_avg + band_fft_max['GammaHigh']
# gamma_lo_avg = gamma_lo_avg + band_fft_max['GammaLow']
# fast_avg = fast_avg + band_fft_max['Fast']
if band_fft_avg['Delta'] != 0:
delta.append(band_fft_max['Delta'])
if band_fft_avg['Theta'] != 0:
theta.append(band_fft_max['Theta'])
if band_fft_avg['Beta'] != 0:
beta.append(band_fft_max['Beta'])
def obtain_concatenate_segments_FFT_TESPAR_A(inputData, encoding):
X = []
Y = []
for i in range(len(inputData.result.arrays)):
count = len(inputData.result.arrays[i].array_data) / 2
print("length is :" + str(count))
for j in range(int(count)):
valid = True
arr1 = []
arr2 = []
band_fft_avg1 = get_average_band_power(
inputData.result.arrays[i].array_data[j * 2],
inputData.result.arrays[i].array_validate[j * 2])
x1 = np.asarray(getList(band_fft_avg1))
t1 = np.asarray(
encoding.get_a(
inputData.result.arrays[i].array_data[j * 2],
inputData.result.arrays[i].array_validate[j * 2])).ravel()
arr1.extend(x1)
arr1.extend(t1)
band_fft_avg2 = get_average_band_power(
inputData.result.arrays[i].array_data[j * 2 + 1],
inputData.result.arrays[i].array_validate[j * 2 + 1])
x2 = np.asarray(getList(band_fft_avg2))
t2 = np.asarray(
encoding.get_s(
inputData.result.arrays[i].array_data[j * 2 + 1],
inputData.result.arrays[i].array_validate[j * 2 +
1])).ravel()
arr2.extend(x2)
arr2.extend(t2)
if x1[0] == 0 or x2[0] == 0:
valid = False
# valid = True
# band_fft_avg1 = get_average_band_power(inputData.result.arrays[i].array_data[j * 2],
# inputData.result.arrays[i].array_validate[j * 2])
# band_fft_avg2 = get_average_band_power(inputData.result.arrays[i].array_data[j * 2 + 1],
# inputData.result.arrays[i].array_validate[j * 2 + 1])
# x1 = np.asarray(getList(band_fft_avg1))
# x2 = np.asarray(getList(band_fft_avg2))
# if x1[0] == 0 or x2[0] == 0:
# valid = False
# array1 = np.asarray(encoding.get_a(inputData.result.arrays[i].array_data[j * 2],
# inputData.result.arrays[i].array_validate[j * 2])).ravel()
# array2 = np.asarray(encoding.get_a(inputData.result.arrays[i].array_data[j * 2 + 1],
# inputData.result.arrays[i].array_validate[j * 2 + 1])).ravel()
# arr = []
# arr.extend(np.append(array1, array2))
# arr.extend(np.append(x1, x2))
if valid:
X.append(np.append(arr1, arr2))
Y.append(inputData.result.arrays[i].name)
return pd.DataFrame(X), Y