def select_classify():
return [
naive(),
tree(criterion="entropy"),
knn(n_neighbors=8, weights='uniform', metric="manhattan"),
mlp(hidden_layer_sizes=(128, ),
alpha=0.01,
activation='tanh',
solver='sgd',
max_iter=300,
learning_rate='constant',
learning_rate_init=0.001)
]
def predictNaiveBayes():
if not os.path.isfile("sentiment/naive_model.plk"):
start = time.time()
#MultinomialNB Pipeline
clf = Pipeline([
('vect', vectorizer),
('clf', naive(alpha=1.0, fit_prior=True))
], verbose=True)
parameters={
'vect__ngram_range': [(1, 1), (1, 2), (1,3), (1,4), (1,5)],
'clf__fit_prior': (True, False),
'clf__alpha': (1.0, 0.1, 0.5, 2.0, .25, 0.75, 0.002),
}
gs_clf = GridSearchCV(clf, parameters, cv=5, iid=False, n_jobs=-1)
gs_clf.fit(docs_train, y_train)
print(gs_clf.best_params_)
y_predicted = gs_clf.predict(docs_test)
print("End.......... total=%.2f s" % (start - time.time()))
# Print the classification report
print(metrics.classification_report(y_test, y_predicted,
target_names=dataset.target_names))
cm = metrics.confusion_matrix(y_test, y_predicted)
print(cm)
plt.matshow(cm, cmap=plt.cm.jet)
#plt.show()
joblib.dump(gs_clf, "naive_model.plk")
return gs_clf
else:
return joblib.load("sentiment/naive_model.plk")
rkf = RepeatedKFold(n_splits=10, n_repeats=10, random_state=1967)
#Loop para percorrer os folds
for train, test in rkf.split(data, target):
X_train = data[train]
X_test = data[test]
y_train = target[train]
y_test = target[test]
for i in range(10):
if (i < 3):
clf.append(knn(n_neighbors=2))
elif (i >= 3 and i < 8):
clf.append(naive())
else:
clf.append(
mlp(solver='sgd',
momentum=0.8,
hidden_layer_sizes=(150),
learning_rate='constant',
learning_rate_init=0.1,
max_iter=500,
random_state=870))
meta = naive()
sclf = StackingClassifier(classifiers=[
clf[0], clf[0], clf[1], clf[2], clf[3], clf[4], clf[5], clf[6], clf[7],
clf[8], clf[9]
data = sismica.iloc[:, :-1].values
target = sismica.iloc[:, -1].values
#Gerando a divisão
kfold = StratifiedKFold(n_splits=10)
#Loop para percorrer os folds
for train, test in kfold.split(data, target):
X_train = data[train]
X_test = data[test]
y_train = target[train]
y_test = target[test]
model = naive()
model.fit(X_train, y_train)
y_pred = model.predict(X_test)
print(y_pred)
cm = confusion_matrix(y_test, y_pred)
acc = str(accuracy_score(y_test, y_pred))
print(cm)
print(acc)
# Fill Nan Categorical values with Unknown age
df['Embarked'] = df['Embarked'].fillna("Unknown")
df['Cabin'] = df['Cabin'].fillna("Unknown")
x = df.loc[:, sel_cols].values
y = df.iloc[:, 1].values
# change Age from ['Male', 'Female'] to [0, 1] through sklearn label encoder
x[:, 2] = label_encoder_x.fit_transform(x[:, 2])
x[:, -1] = label_encoder_x.fit_transform(x[:, -1])
x[:, -2] = label_encoder_x.fit_transform(x[:, -2])
x[:, 6] = label_encoder_x.fit_transform(x[:, 6])
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2, random_state=0)
gbm = naive().fit(x_train, y_train)
predictions = gbm.predict(x_test)
# Print accuracy
accuracy = accuracy_score(y_test, predictions)
print("Accuracy: %.2f%%" % (accuracy * 100.0))
confusion_matrix(y_test, predictions)
df2 = pd.read_csv("/Users/apple/Desktop/Vision/Python_Workspace/dataScience/MachineHack/Titanic/data/titanic/train.csv")
# Fill Nan values with average age
df2['Age'] = df2['Age'].fillna((df2['Age'].mean()))
# Fill Nan Categorical values with Unknown age
df2['Embarked'] = df2['Embarked'].fillna("Unknown")
df2['Cabin'] = df2['Cabin'].fillna("Unknown")
X_train = data[train]
X_test = data[test]
y_train = target[train]
y_test = target[test]
for i in range(20):
clf.append(
MLPClassifier(solver='sgd',
momentum=0.8,
hidden_layer_sizes=(150),
learning_rate='constant',
learning_rate_init=0.1,
max_iter=500,
random_state=870))
meta = naive()
sclf = StackingClassifier(classifiers=[
clf[0], clf[0], clf[1], clf[2], clf[3], clf[4], clf[5], clf[6], clf[7],
clf[8], clf[9], clf[10], clf[11], clf[12], clf[13], clf[14], clf[15],
clf[16], clf[17], clf[18], clf[19]
],
meta_classifier=meta)
sclf.fit(X_train, y_train)
y_pred = sclf.predict(X_test)
print(y_pred)
print(y_test)