def train():
data, _ = process_data.get_signal_values()
Y = process_data.get_crack_lengths()
X = metrics.fft_amp_sums(data) #yes
#X = metrics.correlation_coef(data)
i = metrics.correlation_coef(data) #yes
X = metrics.concatenate_data(X, i)
i = metrics.psd_height_sum(data) #maybe
X = metrics.concatenate_data(X, i)
i = metrics.xc_mean_bin1(data) #yes
X = metrics.concatenate_data(X, i)
X_train, Y_train, X_test, Y_test = process_data.remove_one_plate(X, Y)
X_train, Y_train, X_test, Y_test = process_data.flatten_data(
X_train, Y_train, X_test, Y_test)
scaler = RobustScaler().fit(np.concatenate((X_train, X_test)))
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)
Y_train = np.array(Y_train)
Y_test = np.array(Y_test)
lin_reg.fit(X_train, Y_train, X_test, Y_test)
def load_data():
data, _ = process_data.get_signal_values()
Y = process_data.get_crack_lengths()
X = metrics.fft_amp_sums(data)
# i = metrics.avg_peak_width(data)
# X = metrics.concatenate_data(X, i)
i = metrics.correlation_coef(data)
X = metrics.concatenate_data(X, i)
i = metrics.xc_mean_bin1(data)
X = metrics.concatenate_data(X, i)
return X, Y
def train(model):
data, _ = process_data.get_signal_values()
Y = process_data.get_crack_lengths()
X = metrics.fft_amp_sums(data) #yes
#X = metrics.correlation_coef(data)
# i = metrics.avg_peak_width(data) #no
# X = metrics.concatenate_data(X, i)
i = metrics.correlation_coef(data) #yes
X = metrics.concatenate_data(X, i)
# i = metrics.fft_amp_max(data) #no
# X = metrics.concatenate_data(X, i)
i = metrics.psd_height_sum(data) #maybe
X = metrics.concatenate_data(X, i)
i = metrics.xc_mean_bin1(data) #yes
X = metrics.concatenate_data(X, i)
# i = metrics.psd_max_height(data)
# X = metrics.concatenate_data(X, i)
X_train, Y_train, X_test, Y_test = process_data.remove_one_plate(X, Y)
X_train, Y_train, X_test, Y_test = process_data.flatten_data(
X_train, Y_train, X_test, Y_test)
scaler = RobustScaler().fit(np.concatenate((X_train, X_test)))
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)
Y_train = categorize(Y_train)
Y_test = categorize(Y_test)
model.fit(X_train, Y_train, X_test, Y_test)
def predict(model):
data, _ = process_data.get_signal_values()
Y = process_data.get_crack_lengths()
x1 = metrics.fft_amp_sums(data)
x2 = metrics.avg_peak_width(data)
X = metrics.concatenate_data(x1, x2)
X, Y, _, _ = process_data.flatten_data(X, Y)
predictions = model.predict(X, Y)
print("\nTrue | Predicted")
for t, p in zip(Y, predictions):
print(t, "-", p)
