机器学习入门1

import warnings
warnings.filterwarnings(‘ignore’)
import os
import pandas as pd
import numpy as np
from pylab import *
matplotlib.rcParams[‘font.sans-serif’] = [‘SimHei’]
matplotlib.rcParams[‘axes.unicode_minus’] = False

一：分类

1、读入aviation数据集，去除重复值、空值。

data = pd.read_excel(‘aviation.xls’,index_col=‘MEMBER_NO’)
print(data.head())
data.dropna(inplace=True)
data.drop_duplicates(inplace=True)

print(data.info())

2、切片抽取10000样本，切片X和Y。

sample_data = data.sample(1000)
X = sample_data.iloc[:,:-1]
y = sample_data.iloc[:,-1]

print(X.shape)

3、使用5折交叉验证方法比较XGBOOST和随机森林的F1分值

from sklearn.model_selection import cross_val_score
from xgboost import XGBClassifier
from sklearn.ensemble import RandomForestClassifier
model1 = XGBClassifier()

model2 = RandomForestClassifier()

cv1 = cross_val_score(model1,X,y,scoring=‘f1’,cv=5)

cv2 = cross_val_score(model2,X,y,scoring=‘f1’,cv=5)

print(“XGBOOST和随机森林的F1分值分别为：{},{}”.format(cv1.mean(),cv2.mean()))

4、对上题中F1分值高的模型，进行网格搜索找到该模型的最优参数。

‘’‘XGBOOST的F1值较高’‘’

from sklearn.model_selection import GridSearchCV

param = {‘n_estimators’:[50,100,150],

‘subsample’:[0.4,0.6,0.8,1]}

grid = GridSearchCV(model1,param_grid=param,scoring=‘f1’,cv=10)

grid.fit(X,y)

print(grid.best_params_) #‘n_estimators’: 100, ‘subsample’: 1}

5、按照7：3划分数据集。

from sklearn.model_selection import train_test_split
X,y=data.iloc[:,:-1],data.iloc[:,-1]
X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=0.3,random_state=1)

6、使用训练集数据进行模型训练，对测试集数据进行预测，打印出混淆矩阵和结果报告，画出ROC曲线。

from sklearn.metrics import classification_report,confusion_matrix,roc_curve
model1.set_params(n_estimators=100,subsample=1)
model1.fit(X_train,y_train)

print(confusion_matrix(y_test,model1.predict(X_test)))

print(classification_report(y_test,model1.predict(X_test)))

#生成ROC曲线

预测的概率值—为0的概率和为1的概率两列

fpr, tpr, thresholds = roc_curve(y_test,model1.predict_proba(X_test)[:,1])
plt.plot(fpr,tpr)
plt.show()