def naive_bayes_k(k, sequence_origin='DairyDB', primers_origin='DairyDB', taxonomy_level: int = 1,
selected_primer: str = 'V4',
model_preprocessing='Computing frequency of {}-mer (ATCG) in every sequence', test_size=0.2):
"""
Apply Naive Bayes model on a set of sequence preprocessed data.
:return:
"""
model_preprocessing = model_preprocessing.format(k)
X_train, X_test, y_train, y_test = ETL_NB_k_mer(k=k,
sequence_origin=sequence_origin,
primers_origin=primers_origin,
taxonomy_level=taxonomy_level,
selected_primer=selected_primer)
GNB = GaussianNB()
y_pred = GNB.fit(X_train, y_train).predict(X_test)
test_size, prop_main_class, accuracy = main_stats_model(y_train=y_train,
y_test=y_test,
y_pred=y_pred,
model_name='Naive Bayes - NB({})'.format(k),
model_parameters=GNB.get_params(),
model_preprocessing=model_preprocessing,
sequence_origin=sequence_origin,
primers_origin=primers_origin,
taxonomy_level=taxonomy_level,
selected_primer=selected_primer,
test_size=test_size)
return test_size, prop_main_class, accuracy
def random_forest_k_default(
k=4,
sequence_origin='DairyDB',
primers_origin='DairyDB',
taxonomy_level: int = 1,
selected_primer: str = 'V4',
model_preprocessing='Computing frequency of {}-mer (ATCG) in every sequence',
test_size=0.2,
max_depth=None,
n_estimators=None):
"""
Apply Random Forest model on a set of sequence preprocessed data.
:return:
"""
model_preprocessing = model_preprocessing.format(k)
X_train, X_test, y_train, y_test = ETL_RF_k_mer(
k=k,
sequence_origin=sequence_origin,
primers_origin=primers_origin,
taxonomy_level=taxonomy_level,
selected_primer=selected_primer)
if max_depth is None:
if taxonomy_level >= 5:
max_depth = 10
elif taxonomy_level >= 3 and selected_primer == 'sequence' and sequence_origin == '':
max_depth = 20
else:
max_depth = 50
if n_estimators is None:
n_estimators = 200
RF = RandomForestClassifier(bootstrap=False,
min_samples_leaf=1,
min_samples_split=2,
max_features=min(50, 4**k),
n_estimators=n_estimators,
max_depth=max_depth,
n_jobs=-1) # 30 for max_depth is not backed-up
y_pred = RF.fit(X_train, y_train).predict(X_test)
test_size, prop_main_class, accuracy = main_stats_model(
y_train=y_train,
y_test=y_test,
y_pred=y_pred,
model_name='RF_{}'.format(k),
model_parameters=RF.get_params(),
model_preprocessing=model_preprocessing,
sequence_origin=sequence_origin,
primers_origin=primers_origin,
taxonomy_level=taxonomy_level,
selected_primer=selected_primer,
test_size=test_size,
k=k,
feature_importances=RF.feature_importances_)
return RF, test_size, prop_main_class, accuracy
def xgboost_k_grid_search_cv(
k=4,
param_grid=None,
sequence_origin='DairyDB',
primers_origin='DairyDB',
taxonomy_level: int = 1,
selected_primer: str = 'V4',
model_preprocessing='Computing frequency of {}-mer (ATCG) in every sequence',
test_size=0.2):
"""
Apply Random Forest model on a set of sequence preprocessed data.
:return:
"""
model_preprocessing = model_preprocessing.format(k)
X_train, X_test, y_train, y_test = ETL_k_mer(
k=k,
sequence_origin=sequence_origin,
primers_origin=primers_origin,
taxonomy_level=taxonomy_level,
selected_primer=selected_primer)
XGB = XGBClassifier()
grid_search = GridSearchCV(estimator=XGB,
param_grid=param_grid,
cv=3,
n_jobs=8,
verbose=2)
grid_search.fit(X_train, y_train)
XGB_opt = grid_search.best_estimator_
y_pred = XGB_opt.fit(X_train, y_train).predict(X_test)
test_size, prop_main_class, accuracy = main_stats_model(
y_train=y_train,
y_test=y_test,
y_pred=y_pred,
model_name='XGB_CV_{}'.format(k),
model_parameters=grid_search.best_params_,
model_preprocessing=model_preprocessing,
sequence_origin=sequence_origin,
primers_origin=primers_origin,
taxonomy_level=taxonomy_level,
selected_primer=selected_primer,
test_size=test_size,
feature_importances=XGB_opt.feature_importances_,
k=k,
save_csv=True,
xgb_model=XGB_opt,
save_model=True)
return test_size, prop_main_class, accuracy
def xgboost_k_default(
k=4,
sequence_origin='DairyDB',
primers_origin='DairyDB',
taxonomy_level: int = 1,
selected_primer: str = 'V4',
model_preprocessing='Computing frequency of {}-mer (ATCG) in every sequence',
test_size=0.2):
"""
Apply Random Forest model on a set of sequence preprocessed data.
:return:
"""
model_preprocessing = model_preprocessing.format(k)
X_train, X_test, y_train, y_test = ETL_k_mer(
k=k,
sequence_origin=sequence_origin,
primers_origin=primers_origin,
taxonomy_level=taxonomy_level,
selected_primer=selected_primer)
XGB = XGBClassifier(silent=0, eta=0.3, max_depth=3, n_estimators=100)
y_pred = XGB.fit(X_train, y_train).predict(X_test)
test_size, prop_main_class, accuracy = main_stats_model(
y_train=y_train,
y_test=y_test,
y_pred=y_pred,
model_name='XGB_{}'.format(k),
model_parameters=XGB.get_params(),
model_preprocessing=model_preprocessing,
sequence_origin=sequence_origin,
primers_origin=primers_origin,
taxonomy_level=taxonomy_level,
selected_primer=selected_primer,
test_size=test_size,
k=k,
feature_importances=XGB.feature_importances_,
xgb_model=XGB,
save_model=True,
save_tree=20)
del XGB, X_train, X_test, y_train, y_test, y_pred
return test_size, prop_main_class, accuracy
def random_forest_k_grid_search_cv(
k=5,
param_grid=None,
sequence_origin='DairyDB',
primers_origin='DairyDB',
taxonomy_level: int = 1,
selected_primer: str = 'V4',
model_preprocessing='Computing frequency of {}-mer (ATCG) in every sequence',
test_size=0.2):
"""
Apply Random Forest model on a set of sequence preprocessed data.
:return:
"""
model_preprocessing = model_preprocessing.format(k)
if param_grid is None:
# Number of trees in random forest
n_estimators = [200] # Checked as often the best option
# Number of features to consider at every split
max_features = ['auto'] # Checked as best option
# Maximum number of levels in tree
max_depth = [
None
] # Checked as best option -> Due to memory errors, limiting at 30
# Minimum number of samples required to split a node
min_samples_split = [
2
] # Instead of 2, 5, 10 because of unbalanced classes
# Minimum number of samples required at each leaf node
min_samples_leaf = [
1
] # Instead of 1, 2, 4 because of unbalanced classes
# Method of selecting samples for training each tree
bootstrap = [False] # Checked as best option
# Create the random grid
param_grid = {
'n_estimators': n_estimators,
'max_features': max_features,
'max_depth': max_depth,
'min_samples_split': min_samples_split,
'boostrap': bootstrap
}
X_train, X_test, y_train, y_test = ETL_RF_k_mer(
k=k,
sequence_origin=sequence_origin,
primers_origin=primers_origin,
taxonomy_level=taxonomy_level,
selected_primer=selected_primer)
RF = RandomForestClassifier()
grid_search = GridSearchCV(estimator=RF,
param_grid=param_grid,
cv=3,
n_jobs=2,
verbose=1)
grid_search.fit(X_train, y_train)
RF_opt = grid_search.best_estimator_
y_pred = RF_opt.fit(X_train, y_train).predict(X_test)
test_size, prop_main_class, accuracy = main_stats_model(
y_train=y_train,
y_test=y_test,
y_pred=y_pred,
model_name='RF_CV_{}'.format(k),
model_parameters=grid_search.best_params_,
model_preprocessing=model_preprocessing,
sequence_origin=sequence_origin,
primers_origin=primers_origin,
taxonomy_level=taxonomy_level,
selected_primer=selected_primer,
test_size=test_size,
feature_importances=RF_opt.feature_importances_,
save_model=True,
rf_model=RF_opt,
k=k,
save_csv=True)
return test_size, prop_main_class, accuracy