Scikit-learn（python）学习笔记 （不定期更新）

1.XGboost在python中安装
首先 需要从作者github上clone并且编译C++版本（Ubuntu）

git clone --recursive https://github.com/dmlc/xgboost
cd xgboost; make -j4

然后进入目录中python-packge下

sudo python setup.py install

安装完后正常导入就好

import xgboost as xgb

2.使用XGBClassifier

from xgboost import XGBClassifier
from sklearn.metrics import f1_score
from sklearn.metrics import recall_score

xgbc = XGBClassifier()
xgbc.fit(train_inf, train_target)

predict = xgbc.predict(test_inf) #返回预测结果
xgbc.score(test_inf, test_target) #获得平均准确率
print f1_score(test_target, predict) #获得F1值
print recall_score(test_target, predict) #获得召回率

3.使用随机森林

#随机森林的使用
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import f1_score
from sklearn.metrics import recall_score
from sklearn.metrics import zero_one_loss

clf = RandomForestClassifier(n_estimators=10)
clf = clf.fit(train_inf, train_target) #训练

clf_predict = clf.predict(test_inf) #返回预测结果

print clf.score(test_inf, test_target) #返回准确率
print f1_score(test_target, predict) #返回F1值
print recall_score(test_target, predict) #返回召回率
print zero_one_loss(test_target, predict) #返回0-1损失

4.使用KNN

#KNN的使用
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import f1_score
from sklearn.metrics import recall_score
from sklearn.metrics import zero_one_loss

neigh = KNeighborsClassifier(n_neighbors=3)
neigh.fit(train_inf, train_target) #训练

neigh_predict = neigh.predict(test_inf)  #返回预测结果

print neigh.score(test_inf, test_target) #返回准确率
print f1_score(test_target, neigh_predict) #返回F1值
print recall_score(test_target, neigh_predict) #返回召回率
print zero_one_loss(test_target, neigh_predict) #返回0-1损失

5.使用朴素贝叶斯

#朴素贝叶斯
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import f1_score
from sklearn.metrics import recall_score
from sklearn.metrics import zero_one_loss

Gbayes = GaussianNB()
Gbayes = Gbayes.fit(train_inf, train_target)   #训练

Gbayes_predict = Gbayes.predict(test_inf)  #返回预测结果

print Gbayes.score(test_inf, test_target) #返回准确率
print f1_score(test_target, Gbayes_predict) #返回F1值
print recall_score(test_target, Gbayes_predict) #返回召回率
print zero_one_loss(test_target, Gbayes_predict) #返回0-1损失

1. 地方