第十一章：XGBoost （1）直接调用xgb的库

一、直接调用xgb的库

1.调用xgb的库

导入数据

from sklearn.datasets import load_boston
from sklearn.model_selection import KFold,cross_val_score as CVS,train_test_split as TTS
from sklearn.metrics import mean_squared_error as MSE
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd

data=load_boston()

X=data.data
y=data.target

Xtrain,Xtest,Ytrain,Ytest=TTS(X,y,test_size=0.3,random_state=420)

pd.DataFrame(Xtrain)

导入XGBoost库

#xgb实现法
import xgboost as xgb

#使用类Dmatrix读取数据
dtrain = xgb.DMatrix(Xtrain,Ytrain)  #特征矩阵和标签都进行一个输入
dtest = xgb.DMatrix(Xtest,Ytest)
#非常遗憾无法打开来查看，所以通常都是先读到pandas里面查看之后再放到DMatrix中

#写明参数，
param = {'verbosity':0  #2是输出
         ,'objective':'reg:linear'
         ,"eta":0.1}

num_round=180  #相当于 n_estimators

#类train，可以直接导入的参数是训练数据，树的数量，其他参数都需要通过params来导入
bst = xgb.train(param, dtrain, num_round)

#接口predict
preds=bst.predict(dtest)

MSE(Ytest,preds)
#6.876827553497432

from sklearn.metrics import r2_score
r2_score(Ytest,preds)  #输出的是r的平方，类似score（Xtest,ytest）的结果
#0.9260984369386971

2.xgb.cv(param1, dfull, num_round,n_fold)

有了xgboost.cv这个工具，我们的参数调整就容易多了。这个工具可以让我们直接看到参数如何影响了模型的泛化能
力。接下来，我们将重点讲解如何使用xgboost.cv这个类进行参数调整。

参数，数据，迭代次数，折数
看看类xgb.cv生成了什么结果？随着树不断增加，我们的模型效果如何增加
（训练后的，rmse的均值；训练后的，rmse的方差；测试均值，测试方差）
180次迭代，所以是180棵树

import xgboost as xgb

#为了便捷，使用全数据
dfull = xgb.DMatrix(X,y)

#设定参数
param1 = {'silent':True,'obj':'reg:linear',"gamma":0}
num_round = 180
n_fold=5  #相当于sklearn中的KFold，进行5折交叉验证

#参数，数据，迭代次数，折数
cvresult1 = xgb.cv(param1, dfull, num_round,n_fold)

#看看类xgb.cv生成了什么结果？随着树不断增加，我们的模型效果如何增加
#（训练后的，rmse的均值；训练后的，rmse的方差；测试均值，测试方差）
#180次迭代，所以是180棵树
cvresult1

plt.figure(figsize=(20,5))
plt.grid()
plt.plot(range(1,181),cvresult1.iloc[:,0],c="red",label="train,gamma=0")
plt.plot(range(1,181),cvresult1.iloc[:,2],c="orange",label="test,gamma=0")
plt.legend()
plt.show()


param1 = {'silent':True,'obj':'reg:linear',"gamma":0,"eval_metric":"mae"}
cvresult1 = xgb.cv(param1, dfull, num_round,n_fold)
plt.figure(figsize=(20,5))
plt.grid()
plt.plot(range(1,181),cvresult1.iloc[:,0],c="red",label="train,gamma=0")
plt.plot(range(1,181),cvresult1.iloc[:,2],c="orange",label="test,gamma=0")
plt.legend()
plt.show()
#从这个图中，我们可以看出什么？
#怎样从图中观察模型的泛化能力？
#从这个图的角度来说，模型的调参目标是什么？

3.分类算法使用

from sklearn.datasets import load_breast_cancer
#xgb实现法
import xgboost as xgb

data2 = load_breast_cancer()

x2 = data2.data
y2 = data2.target

dfull2 = xgb.DMatrix(x2,y2)

param1 = {'silent':True,'obj':'binary:logistic',"gamma":0,"nfold":5}
param2 = {'silent':True,'obj':'binary:logistic',"gamma":2,"nfold":5}
num_round = 100

from time import time
import datetime

time0 = time()
cvresult1 = xgb.cv(param1, dfull2, num_round,metrics=("error"))
print(datetime.datetime.fromtimestamp(time()-time0).strftime("%M:%S:%f"))

time0 = time()
cvresult2 = xgb.cv(param2, dfull2, num_round,metrics=("error")) 
print(datetime.datetime.fromtimestamp(time()-time0).strftime("%M:%S:%f"))

plt.figure(figsize=(20,5))
plt.grid()
plt.plot(range(1,101),cvresult1.iloc[:,0],c="red",label="train,gamma=0")
plt.plot(range(1,101),cvresult1.iloc[:,2],c="orange",label="test,gamma=0")
plt.plot(range(1,101),cvresult2.iloc[:,0],c="green",label="train,gamma=2")
plt.plot(range(1,101),cvresult2.iloc[:,2],c="blue",label="test,gamma=2")
plt.legend()
plt.show()

综合调参：使用 xgb.cv调参

#默认情况下的一组线
param1 = {'verbosity':0 #并非默认
         ,'obj':'reg:linear'#并非默认
         ,"subsample":1
         ,"max_depth":6
         ,"eta":0.3
         ,"gamma":0
         ,"lambda":1
         ,"alpha":0
         ,"colsample_bytree":1
         ,"colsample_bylevel":1
         ,"colsample_bynode":1
         ,"nfold":5}
num_round = 300

time0 = time()
cvresult1 = xgb.cv(param1, dfull, num_round)
print(datetime.datetime.fromtimestamp(time()-time0).strftime("%M:%S:%f"))

fig,ax = plt.subplots(1,figsize=(15,8))
ax.set_ylim(top=5)
ax.grid()
ax.plot(range(1,301),cvresult1.iloc[:,0],c="red",label="train,original")
ax.plot(range(1,301),cvresult1.iloc[:,2],c="orange",label="test,original")
# ax.legend(fontsize="xx-large")
# plt.show()

#先把不默认的先统一
param2 = {'verbosity':0 #并非默认
         ,'obj':'reg:linear'
         ,"max_depth":2   #过拟合，调小2
         ,"eta":0.05
          ,"gamma":0     #等于5，没影响运行时间长，10也没影响，20，有一点，所以这个调没啥用
          ,"lambda":1  #调小过拟合严重，调大也没啥用
          ,"alpha":0   #调大没啥用，过大还影响
          ,"colsample_bytree":1
          ,"colsample_bylevel":1
          ,"colsample_bynode":1
         ,"nfold":5}

param3 = {'verbosity':0 #并非默认
          ,'obj':'reg:linear'
           ,"max_depth":2   #过拟合，调小3，与上面对比，看到没有2好，在换成1看看，不好
          ,"eta":0.05
          ,"gamma":0
          ,"lambda":1
          ,"alpha":1
          ,"colsample_bytree":1
         ,"colsample_bylevel":1
         ,"colsample_bynode":0.5
          ,"nfold":5}

time0 = time()
cvresult2 = xgb.cv(param2, dfull, num_round)
print(datetime.datetime.fromtimestamp(time()-time0).strftime("%M:%S:%f"))
time0 = time()
cvresult3 = xgb.cv(param3, dfull, num_round)
print(datetime.datetime.fromtimestamp(time()-time0).strftime("%M:%S:%f"))
ax.plot(range(1,301),cvresult2.iloc[:,0],c="green",label="train,last")
ax.plot(range(1,301),cvresult2.iloc[:,2],c="blue",label="test,last")
ax.plot(range(1,301),cvresult3.iloc[:,0],c="gray",label="train,this")
ax.plot(range(1,301),cvresult3.iloc[:,2],c="pink",label="test,this")
ax.legend(fontsize="xx-large")
plt.show()

xgb.CV输出的结果