sm = SMOTE(k_neighbors=args.kneighbors,
random_state=args.randomseed,
n_jobs=-1)
x_resampled, y_resampled = sm.fit_sample(X, y)
# after over sampleing 读取分类信息并返回数量
np_resampled_y = np.asarray(np.unique(y_resampled, return_counts=True))
df_resampled_y = pd.DataFrame(np_resampled_y.T, columns=['Class', 'Sum'])
print("\nNumber of samples after over sampleing:\n{0}\n".format(
df_resampled_y))
# 初始化 classifier
clf = CascadeForestClassifier(random_state=args.randomseed)
print("\nClassifier parameters:")
print(clf.get_params())
print("\nSMOTE parameters:")
print(sm.get_params())
print("\n")
# 使用SMOTE后数据进行训练
clf.fit(x_resampled, y_resampled)
# 预测测试集
y_pred = clf.predict(X_test)
# 输出测试集统计结果
if (num_categories > 2):
model_evaluation(num_categories, y_test, y_pred)
else:
bi_model_evaluation(y_test, y_pred)
end_time = time.time() # 程序结束时间
print("\n[Finished in: {0:.6f} mins = {1:.6f} seconds]".format(
((end_time - start_time) / 60), (end_time - start_time)))
# Binarización de características categóricas.
# Usamos M.todense() para ver los datos en tamaño normal. Si no, se guardan en formato COOmatrix
OHE = preprocessing.OneHotEncoder(
categorical_features = categoricalAttributes,
handle_unknown = 'ignore'
).fit( X_train )
X_train = OHE.transform(X_train).todense()
X_test = OHE.transform(X_test).todense()
# Equilibrado de representación de cada clase
sm = SMOTE( ratio = 'minority', random_state = seed, k_neighbors = 3 )
Xres, Yres = sm.fit_sample( X_train, y_train )
print ("Tras el equilibrado con SMOTE:{}".format(sm.get_params()))
for i in np.unique( y_train ):
print( "Número de instancias en la clase {}: {} {}"
.format( i, len( np.where( y_train == i )[0] ), len( np.where( y_test == i )[0] ) )
)
# Creación y ajuste del modelo de aprendizaje Random Forest
# para la selección de instancais y características
rfc = RandomForestClassifier(
random_state=seed,
n_estimators = 50,
n_jobs = -1,
max_depth = 30,
min_samples_leaf = 10,
