def best_poly_cross_validation(t, x, D=len(t), K=2, seed=1, istraining=False):
poly_Means = list()
poly_Stds = list()
tr_error_Means = list()
tr_error_Stds = list()
lamb = 0 # doesn't consider the ridge regression here (?)
lowest_order = 0 # will keep track the polynomial value with lowest MSE avg
lowest_mean = 9999999 #the lowest MSE avg from the lowest_order polynomial value
if (istraining == False):
for power in range(0, D + 1):
predictor_matrix = ols_main_vector.creates_predictor_matrix(
x, power)
temp_mean, temp_std = cross_validation(t, x, K, seed, lamb,
istraining)
poly_Means.append(temp_mean)
poly_Stds.append(temp_std)
if (lowest_mean > temp_mean):
lowest_mean = temp_mean
lowest_order = power
return poly_Means, poly_Stds, lowest_mean, lowest_order
else:
for power in range(0, D + 1):
predictor_matrix = ols_main_vector.creates_predictor_matrix(
x, power)
temp_mean, temp_std, temp_error = cross_validation(
t, x, K, seed, lamb, istraining)
poly_Means.append(temp_mean)
poly_Stds.append(temp_std)
tr_error_Means.append(statistics.mean(temp_error))
tr_error_Stds.append(statistics.mean(temp_error))
if (lowest_mean > temp_mean):
lowest_mean = temp_mean
lowest_order = power
return poly_Means, poly_Stds, lowest_mean, lowest_order, tr_error_Means, tr_error_Stds
def best_poly_cross_validation(t, x, D=None, K=2, seed=1, istraining=False):
if (D == None):
D = len(t)
validation_error_mean = list()
validation_error_std = list()
trainning_error_mean = list()
trainning_error_std = list()
lamb = 0
lowest_order = 0
lowest_mean = 9999999
for power in range(1, D + 1):
X = ols_main_vector.creates_predictor_matrix(x, power)
result = cross_validation(t, X, K, seed, lamb, istraining)
validation_mean = result[0]
validation_sdt = result[1]
validation_error_mean.append(validation_mean)
validation_error_std.append(validation_sdt)
if (lowest_mean > validation_mean):
lowest_mean = validation_mean
lowest_order = power
if (istraining == True):
trainning_mean = result[2]
trainning_std = result[3]
trainning_error_mean.append(trainning_mean)
trainning_error_std.append(trainning_std)
if (istraining == True):
return validation_error_mean, validation_error_std, lowest_mean, lowest_order, trainning_error_mean, trainning_error_std
else:
return validation_error_mean, validation_error_std, lowest_mean, lowest_order
def compare(master_trainning_x, master_trainning_t, master_testing_x,
master_testing_t, D, lamb):
result = list()
result.append(0)
for power in range(1, D + 1):
master_traiining_X = ols_main_vector.creates_predictor_matrix(
master_trainning_x, power)
w = ols_main_vector.ols_coefficent_prediction_lamda(
master_traiining_X, master_trainning_t, lamb)
t_hat = ols_main_vector.generate_predition_vector(master_testing_x, w)
error = cross.mean_squared_error(master_testing_t, t_hat)
result.append(error)
return result