def run_random_forest(train_data,train_labels,test_data,test_labels,weights):
print("Running Random Forest...")
clf = RandomForestClassifier(n_estimators=100, max_depth=8, random_state=42)
clf = clf.fit(train_data, train_labels)
preds = clf.predict(test_data)
preds[preds<0.6] = 0
preds[preds>=0.6] = 1
print("Random Forest Classifier:")
print("Accuracy: ", clf.score(test_data,test_labels))
print("AMS score: ", helper.calc_ams(weights.astype(np.float32),test_labels,preds))
print("Precision: ", helper.precision(test_labels,preds))
print("Recall: ", helper.recall(test_labels,preds))
return preds
def run_decision_tree(train_data,train_labels,test_data,test_labels,weights):
print("Running Decision Tree...")
clf = DecisionTreeClassifier(random_state=0)
clf = clf.fit(train_data, train_labels)
preds = clf.predict(test_data)
preds[preds<0.6] = 0
preds[preds>=0.6] = 1
print("Decision Tree Classifier:")
print("Accuracy: ", clf.score(test_data,test_labels))
print("AMS score: ", helper.calc_ams(weights.astype(np.float32),test_labels,preds))
print("Precision: ", helper.precision(test_labels,preds))
print("Recall: ", helper.recall(test_labels,preds))
return preds
def run_gradient_boosting(train_data,train_labels,test_data,test_labels,weights):
print("Running Gradient Boosting...")
clf = GradientBoostingClassifier(n_estimators=100, learning_rate=1.0,max_depth=6, random_state=0)
clf = clf.fit(train_data, train_labels)
preds = clf.predict(test_data)
preds[preds<0.6] = 0
preds[preds>=0.6] = 1
print("Gradient Boosting Classifier:")
print("Accuracy: ", clf.score(test_data,test_labels))
print("AMS score: ", helper.calc_ams(weights.astype(np.float32),test_labels,preds))
print("Precision: ", helper.precision(test_labels,preds))
print("Recall: ", helper.recall(test_labels,preds))
return preds
def run_gnb(train_data,train_labels,test_data,test_labels,weights):
print("Running Gaussian Naive Bayes...")
clf = GaussianNB()
clf = clf.fit(train_data, train_labels)
preds = clf.predict_proba(test_data)[:,1]
preds[preds<0.6] = 0
preds[preds>=0.6] = 1
print("Gaussian Naive Bayes:")
print("Accuracy: ", clf.score(test_data,test_labels))
print("AMS score: ", helper.calc_ams(weights.astype(np.float32),test_labels,preds))
print("Precision: ", helper.precision(test_labels,preds))
print("Recall: ", helper.recall(test_labels,preds))
return preds
def run_lr(train_data,train_labels,test_data,test_labels,weights):
print("Running Logistic Regression...")
clf = LogisticRegression(random_state=0, solver='lbfgs')
clf = clf.fit(train_data, train_labels)
preds = clf.predict_proba(test_data)[:,1]
preds[preds<0.6] = 0
preds[preds>=0.6] = 1
print("Logistic Regression:")
print("Accuracy: ", clf.score(test_data,test_labels))
print("AMS score: ", helper.calc_ams(weights.astype(np.float32),test_labels,preds))
print("Precision: ", helper.precision(test_labels,preds))
print("Recall: ", helper.recall(test_labels,preds))
return preds
def merge_jet_num(preds_0, preds_1, preds_2_3, test_0_labels, test_1_labels,
test_2_3_labels, test_0_weights, test_1_weights,
test_2_3_weights):
preds = np.concatenate((preds_0, preds_1, preds_2_3))
labels = np.concatenate((test_0_labels, test_1_labels, test_2_3_labels))
weights = np.concatenate(
(test_0_weights, test_1_weights, test_2_3_weights))
print(preds.shape)
print(labels.shape)
print(weights.shape)
print("jet-num-merged: ")
print("Accuracy: ", accuracy_score(labels, preds))
print("AMS score: ",
helper.calc_ams(weights.astype(np.float32), labels, preds))
print("Precision: ", helper.precision(labels, preds))
print("Recall: ", helper.recall(labels, preds))