基于XGBoost的数据回归预测

导入包

from sklearn.model_selection import train_test_split
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.model_selection import GridSearchCV #网格搜索
import matplotlib.pyplot as plt#可视化
from sklearn.model_selection import KFold
from sklearn.metrics import r2_score
from sklearn.metrics import mean_squared_error
from scipy.stats import gaussian_kde
import xgboost as xgb
from sklearn.metrics import mean_absolute_error
from sklearn.preprocessing import MinMaxScaler
import pickle

import warnings
warnings.filterwarnings("ignore")

 导入数据

dataset = pd.read_excel('../dataset.xlsx')
dataset = dataset.drop('Unnamed: 0',axis=1)
dataset = dataset.dropna(axis=0)

十折交叉验证

# 十折交叉验证封装方法
def kfold_val(dataset,n_estimators,max_depth,learning_rate,min_child_weight):
    kf = KFold(n_splits=10,shuffle=True,random_state=True)
    for train_i,test_i in kf.split(dataset):
        x_train = dataset.iloc[train_i,5:]
        y_train=dataset.iloc[train_i,4]
        x_test=dataset.iloc[test_i,5:]
        y_test = dataset.iloc[test_i,4]
        model=xgb.XGBRegressor(n_estimators=n_estimators,max_depth =max_depth,learning_rate=learning_rate,min_child_weight=min_child_weight)  
        model.fit(x_train,y_train)
        y_pred = model.predict(x_test)
        mse=mse + mean_squared_error(y_pred,y_test)
        r2 =r2 + r2_score(y_pred,y_test)
    r2 = r2/10
    rmse = np.sqrt(mse)
    print('r2',r2)
    print('rmse',rmse)

kfold_val(dataset,1600,10,0.1,14)

可以确定好大范围后进行网格调参，最后进行画图

网格调参代码

parameters = {'n_estimators':[......],
              'max_depth': [......],
              'learning_rate':[.......],
              'min_child_weight':[......]
             }
model = xgb.XGBRegressor()
gsearch = GridSearchCV(model, param_grid=parameters, cv=2)
gsearch = gsearch.fit(x_train,y_train)
print("Best score: %0.3f" % gsearch.best_score_)
print("Best parameters set:")
best_parameters = gsearch.best_estimator_.get_params()
for param_name in sorted(parameters.keys()):
    print("\t%s: %r" % (param_name, best_parameters[param_name]))