# std_scaler = StandardScaler()
# std_scaler.fit(x_array)
# x_array = std_scaler.transform(x_array)
x_train, x_test, y_train, y_test = train_test_split(x_array, y_array, train_size=0.75, random_state=random_seed)
print(np.shape(x_train))
print(np.shape(y_train))
predict_, rmse_, r2_score_ = [], [], []
for num, model in enumerate(models):
print('\nmodel name:' + model_name[num])
model_start = time.time()
model.fit(x_train, y_train)
predict_.append(np.ravel(model.predict(x_test)))
r2_score_.append(r2_score(y_test, model.predict(x_test)))
rmse_.append(RMSE(y_test, model.predict(x_test)))
model_end = time.time()
model_tol = str_round(model_end - model_start)
print('R^2= ' + str_round(r2_score(y_test, model.predict(x_test)), 3))
print('RMSE= ' + str_round(math.sqrt(mean_squared_error(y_test, model.predict(x_test))), 3))
print('model running time=' + model_tol + 's')
# print(r2_score_)
# print(rmse_)
predict_ = np.array(predict_)
print(predict_[:, 6])
y_predict = np.average(predict_, weights=model_weight, axis=0)
print(y_predict)
print(y_test)
print(str_round(r2_score(y_test, y_predict), 3))
df.loc[i, 'soil_num'] = df.loc[i,'soil_num'][0:8]
d_rows = df[df['soil_num'].duplicated(keep=False)]
df.drop(d_rows.index,axis=0,inplace=True)
g_item = d_rows.groupby('soil_num').mean()
df = df.append(g_item,sort=False)
df.sort_values(by='soil_num',inplace=True,ascending=False)
df.drop(['soil_num'],axis=1,inplace=True)
# print(df)
X_array = np.array(df.loc[:,'point_1':'point_9'])
Y_array = np.array(df.loc[:,'TOC':'WT'])
# print(X_array)
scaler = StandardScaler()
scaler.fit(X_array)
X_train = scaler.transform(X_array)
for i in range(Y_array.shape[1]):
Y_train = Y_array[:,i]
rf = RandomForestRegressor(n_estimators=200,
criterion='mse', max_depth=9,
min_samples_split=2, min_samples_leaf=1,
min_weight_fraction_leaf=0.0, max_features=1,
max_leaf_nodes=None, min_impurity_decrease=0.0,
min_impurity_split=None, bootstrap=True, oob_score=True,
n_jobs=None, random_state=1, verbose=0, warm_start=False)
rf.fit(X_train,Y_train)
print('R^2 =',rf.score(X_train,Y_train))
print('RMSE = ',RMSE(Y_train,rf.predict(X_train)))
min_weight_fraction_leaf=0.0, max_features=1,
max_leaf_nodes=None, min_impurity_decrease=0.0,
min_impurity_split=None, bootstrap=True, oob_score=True,
n_jobs=None, random_state=1, verbose=0, warm_start=False)
# forest = ExtraTreesRegressor(n_estimators=100)
# forest = MLPRegressor(
# hidden_layer_sizes=(5, 3), activation='relu', solver='adam', alpha=0.0001, batch_size='auto',
# learning_rate='constant', learning_rate_init=0.001, power_t=0.5, max_iter=15000, shuffle=True,
# random_state=1, tol=0.0001, verbose=False, warm_start=False, momentum=0.9, nesterovs_momentum=True,
# early_stopping=False, beta_1=0.9, beta_2=0.999, epsilon=1e-08)
forest.fit(x, y)
ax.scatter(x_test[:, 0], y_test, color='teal', s=10, label='origin')
ax.scatter(x_test[:, 0], forest.predict(x_test), color='chocolate', s=3, label='predict')
ax.set_xticks([])
ax.text(0.54, 0.80,
'r^2=' + str(round(r2_score(y_test, forest.predict(x_test)), 3)) +
'\nrmse=' + str(round(RMSE(y_test, forest.predict(x_test)), 3)) , transform=ax.transAxes, fontsize=8,bbox=dict(boxstyle="round,pad=0.2", fc="w", ec="gray", lw=1,alpha=0.5)
)
ax.legend(fontsize=8, loc='upper left')
n = n + 1
# pca = PCA(n_components=2)
# pca = PCA(n_components='mle', copy=True, whiten=False)
# # n_components 保留特征数，‘mle’为自动选择。copy将原始训练数据复制。whiten白化，使得每个特征具有相同方差
# pca.fit()
# pca.fit_transform()
# pca.transform()
# pca.inverse_transform()
# plt.savefig('D:\\Desktop\\PCA.jpg', dpi=400)
plt.show()
]
plt.figure('Normalize', figsize=(9.6, 5.4), dpi=100)
# plt.suptitle('Normalize Methods', fontsize=30)
def str_round(text, decimal=3):
return str(round(text, decimal))
for num, method in enumerate(methods):
start_time = time.time()
method.fit(x_train, y_train)
prediction = method.predict(x_test)
r2 = str_round(r2_score(y_test, prediction), 3)
rmse = str_round(RMSE(y_test, prediction), 3)
plt.subplot(2, 4, num + 1)
plt.title(method_name[num], fontsize=15)
plt.scatter(x_label, y_test, c='chocolate', s=10, label='origin')
plt.scatter(x_label, prediction, c='teal', s=3, label='predict')
plt.text(0.13,
37,
'r^2=' + r2 + '\nrmse=' + rmse,
fontsize=8,
bbox=dict(boxstyle="round,pad=0.2",
fc="w",
ec="gray",
lw=1,
alpha=0.8))
plt.xticks([]), plt.yticks(fontsize=6)
p_x_test.append(nine_point_average(i))
p_x_test = np.array(p_x_test)
clf1 = clf
clf1.fit(x_train, y_train)
clf2 = clf
clf2.fit(p_x_train, y_train)
plt.figure(figsize=(9.6, 5.4), dpi=100)
plt.subplot(2, 1, 1)
plt.title('origin')
plt.scatter(x_test[:, 0], y_test, s=10, c='chocolate', label='origin')
plt.scatter(x_test[:, 0], clf1.predict(x_test), s=3, c='teal', label='predict')
plt.text(
2.75, 33, 'r^2=' + str(round(r2_score(y_test, clf1.predict(x_test)), 3)) +
'\nrmse=' + str(round(RMSE(y_test, clf1.predict(x_test)), 3)))
plt.legend(fontsize=8)
plt.xticks([]), plt.yticks(fontsize=6)
plt.subplot(2, 1, 2)
plt.title('tranform')
plt.scatter(p_x_test[:, 0], y_test, s=10, c='chocolate', label='origin')
plt.scatter(p_x_test[:, 0],
clf2.predict(p_x_test),
s=3,
c='teal',
label='predict')
plt.text(
2.75, 33,
'r^2=' + str(round(r2_score(y_test, clf2.predict(p_x_test)), 3)) +
'\nrmse=' + str(round(RMSE(y_test, clf2.predict(p_x_test)), 3)))
'''
plt.figure('Outlier Test', figsize=(9.6, 3.8), dpi=200)
# plt.suptitle('Elliptic Envelope contamination', fontsize=20)
for num, i in enumerate(list([0.01, 0.03, 0.05, 0.1, 0.2, 0.5])):
cov = EllipticEnvelope(random_state=1, contamination=i)
cov.fit(np.hstack([X_array, Y_array.reshape(-1, 1)]))
index = cov.predict(np.hstack([X_array, Y_array.reshape(-1, 1)]))
X_valid, X_invalid = error_wipe(X_array, index)
Y_valid, Y_invalid = error_wipe(Y_array, index)
reg1 = LinearRegression()
reg1.fit(X_valid, Y_valid)
reg2 = LinearRegression()
reg2.fit(X_array, Y_array)
reg1.rmse = RMSE(Y_valid,reg1.predict(X_valid))
reg2.rmse = RMSE(Y_array,reg2.predict(X_array))
print('reg1', reg1.score(X_valid, Y_valid))
print('reg2', reg2.score(X_array, Y_array))
print(all_np(index))
print(random.sample(range(100), 10))
plt.subplot(2, 3, num + 1)
plt.title('contamination = ' + str(i), fontsize=10)
plt.scatter(X_valid[:, 48], Y_valid, color='chocolate', s=10)
plt.scatter(X_invalid[:, 48], Y_invalid, color='teal', s=15)
plt.xticks([]), plt.yticks(fontsize=6)
plt.text(0.16, 38, '原始R^2 = ' + str(round(reg2.score(X_array, Y_array), 3)) + '\n去除异常值后R^2 = ' + str(
round(reg1.score(X_valid, Y_valid), 3)) + '\n原始RMSE=' + str(round(reg2.rmse,3)) + '\n去除异常值后RMSE=' + str(round(reg1.rmse,3)), fontsize=6)
plt.savefig('D:\Desktop\hyper\EllipticEnvelope contamination.jpeg', dpi=400)
# plt.subplot(2, 3, num + 1)
# plt.title('nu = ' + str(i), fontsize=10)
# plt.scatter(X_valid[:, 48], Y_valid, color='chocolate', s=10)
# plt.scatter(X_invalid[:, 48], Y_invalid, color='teal', s=15)
# plt.xticks([]), plt.yticks(fontsize=6)
# plt.text(0.16, 38, '原始R^2 = ' + str(round(reg2.score(X_array, Y_array), 3)) + '\n去除异常值后R^2 = ' + str(
# round(reg1.score(X_valid, Y_valid), 3)) + '\n原始RMSE=' + str(round(reg2.rmse, 3)) + '\n去除异常值后RMSE=' + str(
# round(reg1.rmse, 3)), fontsize=6)
#
# plt.savefig('D:\Desktop\hyper\OneClassSVM contamination.jpeg', dpi=400)
# plt.show()
one = OneClassSVM(kernel='rbf', degree=3, gamma='auto', coef0=0.0, tol=0.001, nu=0.3, shrinking=True, cache_size=200, verbose=False, max_iter=-1, random_state=1)
one.fit(X_array)
index = one.predict(X_array)
print(all_np(index))
X_valid, X_invalid = error_wipe(X_array, index)
Y_valid, Y_invalid = error_wipe(Y_array, index)
reg = LinearRegression()
reg.fit(X_valid,Y_valid)
print(str(round(reg.score(X_valid,Y_valid),3)))
print(str(round(RMSE(Y_valid,reg.predict(X_valid)),3)))
plt.figure('OneClassSVM',figsize=(4.8,5.4),dpi=100)
plt.title('OneClassSVM',fontsize=20)
plt.scatter(X_array[:,48],Y_array,color='teal', s=10)
plt.scatter(X_valid[:,48],Y_valid,color='chocolate', s=15)
plt.savefig('D:\Desktop\hyper\OneClassSVM.jpg',dpi=100)
plt.show()
print(name)
try:
regressor.fit(x_train, y_train)
except ValueError:
p_x_train, p_x_test, p_y_train, p_y_test = continuous_to_multiclass(
x_train, x_test, y_train, y_test)
regressor.fit(p_x_train, p_y_train)
error = False
try:
score = regressor.score(p_x_train, p_y_train)
except AttributeError:
score = False
print(AttributeError)
try:
rmse = RMSE(p_y_test, regressor.predict(p_x_test))
except AttributeError:
rmse = False
print(AttributeError)
else:
error = True
try:
score = regressor.score(x_train, y_train)
except AttributeError:
score = False
print(AttributeError)
try:
rmse = RMSE(y_test, regressor.predict(x_test))
except AttributeError:
rmse = False