写出EEMD+STL+XGBoost+LightGBM+ConvLSTM的读取csv数据的多输入单输出的时序训练及预测代码；其中，前80%数据为训练集，后20%为测试集。

使用EEMD STL XGBoost LightGBM ConvLSTM进行时序训练及预测代码如下：# 加载需要的包
import pandas as pd
from eemd import EEMD
from stl import STL
from xgboost import XGBRegressor
from lightgbm import LGBMRegressor
from keras.layers import ConvLSTM2D# 读取csv数据
data = pd.read_csv('data.csv')# 划分训练集和测试集
train_data = data.iloc[:80, :]
test_data = data.iloc[80:, :]# 使用EEMD分析
eemd_data = EEMD().fit_transform(train_data)# 使用STL分析
stl_data = STL().fit_transform(train_data)# 使用XGBoost训练
xgb_model = XGBRegressor().fit(eemd_data, train_data['target'])# 使用LightGBM训练
lgbm_model = LGBMRegressor().fit(stl_data, train_data['target'])# 使用ConvLSTM2D训练
model = Sequential()
model.add(ConvLSTM2D(filters=32, kernel_size=(3, 3),
input_shape=(train_data.shape[0], 1, train_data.shape[1], 1)))
model.add(Flatten())
model.add(Dense(1))
model.compile(loss='mse', optimizer='adam')
model.fit(train_data.values.reshape(train_data.shape[0], 1, train_data.shape[1], 1), train_data['target'], epochs=10, batch_size=32)# 预测
xgb_pred = xgb_model.predict(eemd_data)
lgbm_pred = lgbm_model.predict(stl_data)
convlstm_pred = model.predict(test_data.values.reshape(test_data.shape[0], 1, test_data.shape[1], 1))# 评估
xgb_score = xgb_model.score(eemd_data, train_data['target'])
lgbm_score = lgbm_model.score(stl_data, train_data['target'])
convlstm_score = model.evaluate(test_data.values.reshape(test_data.shape[0], 1, test_data.shape[1], 1), test_data['target'])