|
@@ -63,21 +63,6 @@ def ensemble_model_v2(X_train, y_train):
|
|
|
|
|
|
return ensemble_model
|
|
|
|
|
|
-def rfe_svm_model(X_train, y_train, n_components=1):
|
|
|
-
|
|
|
- # Cs = [0.001, 0.01, 0.1, 1, 10, 100, 1000]
|
|
|
- # gammas = [0.001, 0.01, 0.1, 1, 5, 10, 100]
|
|
|
- # param_grid = [{'estimator__C': Cs, 'estimator__gamma' : gammas}]
|
|
|
-
|
|
|
- gammas = [0.001, 0.01, 0.1]
|
|
|
- param_grid = [{'estimator__gamma' : gammas}]
|
|
|
-
|
|
|
- estimator = svm.SVC(kernel="linear")
|
|
|
- selector = RFECV(estimator, step=1, cv=4, verbose=0)
|
|
|
- clf = GridSearchCV(selector, param_grid, cv=5, verbose=1)
|
|
|
- clf.fit(X_train, y_train)
|
|
|
-
|
|
|
- return (clf.best_estimator_, clf.best_estimator_.support_)
|
|
|
|
|
|
|
|
|
def get_trained_model(choice, X_train, y_train):
|
|
@@ -89,7 +74,4 @@ def get_trained_model(choice, X_train, y_train):
|
|
|
return ensemble_model(X_train, y_train)
|
|
|
|
|
|
if choice == 'ensemble_model_v2':
|
|
|
- return ensemble_model_v2(X_train, y_train)
|
|
|
-
|
|
|
- if choice == 'rfe_svm_model':
|
|
|
- return rfe_svm_model(X_train, y_train)
|
|
|
+ return ensemble_model_v2(X_train, y_train)
|