def do_generate_logistic_simple_model(X_train, y_train, parameters):
model = LogisticRegression(random_state=my_constants.RANDOM_VALUE)
model_grid = GridSearchCV(model,
param_grid=parameters,
cv=3,
verbose=3,
n_jobs=3)
with ignore_warnings(category=ConvergenceWarning):
model_grid.fit(X_train, y_train)
file_operations.write_logs(FILENAME,
"Calculate logistic simple model" + model_grid)
return model_grid
def do_generate_rf_optimazed_model(X_train, y_train, parameters):
file_operations.write_logs(
FILENAME, 'Starting RF Grid Search with parameters:' + str(parameters))
model = RandomForestClassifier(random_state=my_constants.RANDOM_VALUE,
oob_score=True)
model_grid = GridSearchCV(model,
param_grid=parameters,
cv=3,
verbose=3,
n_jobs=3)
with ignore_warnings(category=ConvergenceWarning):
model_grid.fit(X_train, y_train)
file_operations.write_logs(FILENAME, 'RF Grid search completed')
return model_grid
def do_generate_metrics_logistic_simple_model(X_train, y_train, X_test, y_test,
grid):
file_operations.write_logs(FILENAME,
"do_generate_metrics_logistic_simple_model")
model = LogisticRegression(random_state=my_constants.RANDOM_VALUE)
file_operations.write_logs(FILENAME, "grid Best params")
file_operations.write_logs(FILENAME, str(grid.best_params_))
model.set_params(**grid.best_params_)
model.fit(X_train, y_train)
metrics = calculate_metrics(model, X_test, y_test)
file_operations.write_logs(
FILENAME, 'model params:' + str(model.get_params()) + " model score:" +
str(model.score))
file_operations.write_logs(
FILENAME, 'model grid.best_params_:' + str(model.get_params()) +
" grid.best_score_:" + str(grid.best_score_))
return model, metrics
def do_generate_lgbm_optimazed_model(X_train, y_train, parameters):
file_operations.write_logs(FILENAME,
'Starting LGBM Grid Search with parameters:')
file_operations.write_logs(FILENAME, str(parameters))
model = LGBMClassifier(random_state=0)
model = GridSearchCV(model,
param_grid=parameters,
cv=3,
verbose=3,
n_jobs=3)
model.fit(X_train, y_train)
file_operations.write_logs(FILENAME, "LGBM grid search completed")
return model
def do_generate_metrics_lgbm_optimazed_model(X_train, y_train, X_test, y_test,
grid):
file_operations.write_logs(FILENAME, "LGBM metrics calculation\n")
model = LGBMClassifier(random_state=0)
model.set_params(**grid.best_params_)
model.fit(X_train, y_train)
metrics = calculate_metrics(model, X_test, y_test)
file_operations.write_logs(
FILENAME, "Generated model params and results\n params:" +
str(model.get_params()) + "\nscore " +
str(model.score(X_test, y_test)))
file_operations.write_logs(
FILENAME, "Search grid best params and results\n params:" +
str(grid.best_params_) + "\nscore " + str(grid.best_score_))
return model, metrics
def do_generate_metrics_logistic_simple_model(X_train, y_train, X_test, y_test,
grid):
model = LogisticRegression(random_state=my_constants.RANDOM_VALUE)
file_operations.write_logs(
FILENAME,
"Calculate logistic simple model best params" + grid.best_params_)
model.set_params(**grid.best_params_)
model.fit(X_train, y_train)
metrics = calculate_metrics(model, X_test, y_test)
file_operations.write_logs(
FILENAME,
"model params" + model.get_params() + " scores:" + model.score)
file_operations.write_logs(
FILENAME,
"Grid params" + grid.best_params_ + " scores:" + grid.best_score_)
return model, metrics
def do_generate_metrics_rf_optimazed_model(X_train, y_train, X_test, y_test,
grid):
file_operations.write_logs(FILENAME, 'Starting metrics calculation')
model = RandomForestClassifier(random_state=my_constants.RANDOM_VALUE,
oob_score=True)
model.set_params(**grid.best_params_)
model.fit(X_train, y_train)
metrics = calculate_metrics(model, X_test, y_test)
file_operations.write_logs(
FILENAME, "Generated model params and results\n params:" +
str(model.get_params()) + "\nscore " +
str(model.score(X_test, y_test)))
file_operations.write_logs(
FILENAME, "Search grid best params and results\n params:" +
str(grid.best_params_) + "\nscore " + str(grid.best_score_))
return model, metrics
def predictions():
train_df = file_operations.read_data('processed', 'train.csv',
'PassengerId')
competition_df = file_operations.read_data('processed', 'test.csv',
'PassengerId')
X = train_df.loc[:, 'Age':].values.astype('float')
y = train_df['Survived'].ravel()
shape = X.shape
if shape[0] == 891 & shape[1] > 36:
file_operations.write_logs(
FILENAME,
"Dataset has " + shape[1] + " and right amount amount of rows")
file_operations.write_logs(FILENAME, 'Creating test and train dataset')
X_train, X_test, y_train, y_test = train_test_split(
X,
y,
test_size=my_constants.TEST_SIZE,
random_state=my_constants.RANDOM_VALUE)
# Linear base dummy model
file_operations.write_logs(FILENAME, 'Creating linear model')
base_model = create_base_model(X_train, y_train, X_test, y_test)
file_operations.write_logs(FILENAME,
"Metrics BaseModel: " + base_model['metrics'])
file_operations.get_submission_file(base_model['model'],
'01_base_model.csv', competition_df)
# Logistic regression model
file_operations.write_logs(FILENAME, 'Creating logistic simple model')
lg_simple_model = create_logistic_simple_model(X_train, y_train, X_test,
y_test)
file_operations.write_logs(
FILENAME, "Metrics lg_simple_model: " + lg_simple_model['metrics'])
file_operations.get_submission_file(lg_simple_model['model'],
'02_lg_model.csv', competition_df)
# Logistic regression model with hyp optimization
file_operations.write_logs(FILENAME, 'Creating logistic optimazed model')
lg_optimazed_model = create_logistic_optimazed_model(
X_train, y_train, X_test, y_test)
file_operations.write_logs(
FILENAME,
"Metrics lg_optimazed_model: " + lg_optimazed_model['metrics'])
file_operations.get_submission_file(lg_optimazed_model['model'],
'03_lg_model_optimized.csv',
competition_df)
# print('Creating rf model')
# rf_model_scaled = create_rf_optimized_model(X_train, y_train, X_test, y_test)
# print("Metrics rf_model_scaled: ", rf_model_scaled['metrics'])
# file_operations.get_submission_file(rf_model_scaled['model'], '04_rf_model_optimized.csv', competition_df)
#
# print('Creating lgbm model')
# lgbm_model = create_lgbm_optimized_model(X_train, y_train, X_test, y_test)
# print("Metrics lgbm_model: ", lgbm_model['metrics'])
# file_operations.get_submission_file(lgbm_model['model'], '04_lgbm_model_optimized.csv', competition_df)
#
# print('Creating SVC scaled model')
# svc_model = create_svc_optimized_model(X_train, y_train, X_test, y_test)
# file_operations.get_submission_file(svc_model['model'], '05_svc_model_optimized_scaled.csv', competition_df)
# Feature standarization
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)
# linear base dummy model
file_operations.write_logs(FILENAME, 'Creating dummy scaled model')
base_model_scaled = create_base_model(X_train_scaled, y_train,
X_test_scaled, y_test)
file_operations.write_logs(
FILENAME, "Metrics base_model_scaled: " + base_model_scaled['metrics'])
file_operations.get_submission_file_with_standardization(
base_model_scaled['model'], '01_base_model_scaled.csv', scaler,
competition_df)
# Logistic regression model
file_operations.write_logs(FILENAME,
'Creating logitic optimazed scaled model')
lg_simple_model_scaled = create_logistic_simple_model(
X_train, y_train, X_test, y_test)
file_operations.write_logs(
FILENAME,
"Metrics lg_simple_model_scaled: " + lg_simple_model_scaled['metrics'])
file_operations.get_submission_file_with_standardization(
lg_simple_model_scaled['model'], '02_lg_model_scaled.csv', scaler,
competition_df)
# Logistic regression model with hyp optimization
print('Creating logitic optimazed scaled model')
lg_optimazed_model_scaled = create_logistic_optimazed_model(
X_train, y_train, X_test, y_test)
print("Metrics lg_optimazed_model_scaled: ",
lg_optimazed_model_scaled['metrics'])
file_operations.get_submission_file_with_standardization(
lg_optimazed_model_scaled['model'], '03_lg_model_optimized_scaled.csv',
scaler, competition_df)
def predictions():
train_df = file_operations.read_data('processed', 'train.csv',
'PassengerId')
competition_df = file_operations.read_data('processed', 'test.csv',
'PassengerId')
X = train_df.loc[:, 'Age':].values.astype('float')
y = train_df['Survived'].ravel()
shape = X.shape
if shape[0] == 891 & shape[1] > 36:
file_operations.write_logs(FILENAME, "Dataset has ", shape[1],
" and right amount amount of rows")
file_operations.write_logs(FILENAME, 'Creating test and train dataset')
X_train, X_test, y_train, y_test = train_test_split(
X,
y,
test_size=my_constants.TEST_SIZE,
random_state=my_constants.RANDOM_VALUE)
# Linear base dummy model
file_operations.write_logs(FILENAME, 'Creating linear model')
base_model = create_base_model(X_train, y_train, X_test, y_test)
file_operations.write_logs(FILENAME, "Metrics base_model: ")
file_operations.write_logs(FILENAME, str(base_model['metrics']))
file_operations.get_submission_file(base_model['model'],
'01_base_model.csv', competition_df)
file_operations.write_logs(FILENAME, 'Creating lgbm model')
lgbm_model = create_lgbm_optimized_model(X_train, y_train, X_test, y_test)
file_operations.write_logs(FILENAME, "Metrics lgbm_model: ")
file_operations.write_logs(FILENAME, str(lgbm_model['metrics']))
file_operations.get_submission_file(lgbm_model['model'],
'04_lgbm_model_optimized.csv',
competition_df)
# Feature standarization
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)
file_operations.write_logs(FILENAME, 'Creating lgbm scaled model')
lgbm_model_scaled = create_lgbm_optimized_model(X_train_scaled, y_train,
X_test_scaled, y_test)
file_operations.write_logs(FILENAME, "Metrics lgbm_model_scaled: ")
file_operations.write_logs(FILENAME, str(lgbm_model_scaled['metrics']))
file_operations.get_submission_file(str(lgbm_model_scaled['model']),
'04_lgbm_model_optimized_scaled.csv',
competition_df)