def print_score(no_nca_list, Y_train, Y_test):
""" The function prints the scores of the models and the prediction performance """
target_names = ["class 0", "class 1"]
no_nca = namedtuple("no_nca", [
"grid_score", "grid_params", "grid_confusion", "tr_report",
"te_report", "train_mat", "grid_matthews", "grid_train_confusion"
])
# Model comparison
grid_score = no_nca_list.fitted_grid.best_score_
grid_params = no_nca_list.fitted_grid.best_params_
# Training scores
grid_train_confusion = confusion_matrix(Y_train, no_nca_list.grid_train_Y)
grid_tr_report = class_re(Y_train,
no_nca_list.grid_train_Y,
target_names=target_names,
output_dict=True)
train_mat = matthews_corrcoef(Y_train, no_nca_list.grid_train_Y)
# Test metrics grid
grid_confusion = confusion_matrix(Y_test, no_nca_list.y_grid)
grid_matthews = matthews_corrcoef(Y_test, no_nca_list.y_grid)
grid_te_report = class_re(Y_test,
no_nca_list.y_grid,
target_names=target_names,
output_dict=True)
everything = no_nca(*[
grid_score, grid_params, grid_confusion, grid_tr_report,
grid_te_report, train_mat, grid_matthews, grid_train_confusion
])
return everything
def print_score(Y_test,
y_grid,
train_predicted,
Y_train,
test_index=None,
train_index=None,
mode=None):
""" The function prints the scores of the models and the prediction performance """
score_tuple = namedtuple("scores", [
"test_confusion", "tr_report", "te_report", "train_mat", "test_mat",
"train_confusion"
])
target_names = ["class 0", "class 1"]
# looking at the scores of those predicted by al 3 of them
if mode:
Y_test = Y_test.iloc[[
x for x in range(len(Y_test)) if x not in test_index
]]
Y_train = Y_train.iloc[[
x for x in range(len(Y_train)) if x not in train_index
]]
y_grid = [y_grid[x] for x in range(len(y_grid)) if x not in test_index]
train_predicted = [
train_predicted[x] for x in range(len(train_predicted))
if x not in train_index
]
# Training scores
train_confusion = confusion_matrix(Y_train, train_predicted)
train_matthews = matthews_corrcoef(Y_train, train_predicted)
# print(f"Y_train : {Y_train}, predicted: {train_predicted}")
tr_report = class_re(Y_train,
train_predicted,
target_names=target_names,
output_dict=True)
# Test metrics
test_confusion = confusion_matrix(Y_test, y_grid)
test_matthews = matthews_corrcoef(Y_test, y_grid)
te_report = class_re(Y_test,
y_grid,
target_names=target_names,
output_dict=True)
all_scores = score_tuple(*[
test_confusion, tr_report, te_report, train_matthews, test_matthews,
train_confusion
])
return all_scores
def print_score(fitted_grid, fitted_random, Y_test, y_random, y_grid,
random_train_predicted, grid_train_predicted, Y_train):
""" The function prints the scores of the models and the prediction performance """
target_names = ["class 0", "class 1"]
# Model comparison
grid_score = fitted_grid.best_score_
grid_params = fitted_grid.best_params_
random_score = fitted_random.best_score_
random_params = fitted_random.best_params_
# Training scores
random_train_confusion = confusion_matrix(Y_train, random_train_predicted)
grid_train_confusion = confusion_matrix(Y_train, grid_train_predicted)
g_train_matthews = matthews_corrcoef(Y_train, grid_train_predicted)
r_train_matthews = matthews_corrcoef(Y_train, random_train_predicted)
random_tr_report = class_re(Y_train,
random_train_predicted,
target_names=target_names,
output_dict=True)
grid_tr_report = class_re(Y_train,
grid_train_predicted,
target_names=target_names,
output_dict=True)
# Test metrics
random_confusion = confusion_matrix(Y_test, y_random)
random_matthews = matthews_corrcoef(Y_test, y_random)
random_te_report = class_re(Y_test,
y_random,
target_names=target_names,
output_dict=True)
grid_matthews = matthews_corrcoef(Y_test, y_grid)
grid_confusion = confusion_matrix(Y_test, y_grid)
grid_te_report = class_re(Y_test,
y_grid,
target_names=target_names,
output_dict=True)
return grid_score, grid_params, grid_confusion, random_tr_report, random_te_report, grid_te_report, grid_tr_report, \
grid_matthews, random_score, random_params, random_confusion, random_matthews, \
random_train_confusion, grid_train_confusion, r_train_matthews, g_train_matthews
def print_score(fitted_grid, Y_test, y_grid, grid_train_predicted, Y_train,
X_test):
""" The function prints the scores of the models and the prediction performance """
score_tuple = namedtuple("scores", [
"grid_score", "grid_params", "grid_confusion", "tr_report",
"te_report", "train_mat", "grid_matthews", "grid_train_confusion",
"grid_r2"
])
target_names = ["class 0", "class 1"]
# Model comparison
grid_score = fitted_grid.best_score_
grid_params = fitted_grid.best_params_
grid_r2 = fitted_grid.best_estimator_.score(X_test, Y_test)
# Training scores
grid_train_confusion = confusion_matrix(Y_train, grid_train_predicted)
g_train_matthews = matthews_corrcoef(Y_train, grid_train_predicted)
grid_tr_report = class_re(Y_train,
grid_train_predicted,
target_names=target_names,
output_dict=True)
# Test metrics
grid_confusion = confusion_matrix(Y_test, y_grid)
grid_matthews = matthews_corrcoef(Y_test, y_grid)
grid_te_report = class_re(Y_test,
y_grid,
target_names=target_names,
output_dict=True)
all_scores = score_tuple(*[
grid_score, grid_params, grid_confusion, grid_tr_report,
grid_te_report, g_train_matthews, grid_matthews, grid_train_confusion,
grid_r2
])
return all_scores