机器学习项目--- 天池风险评估代码
2019-02-26 本文已影响0人
longsan0918
# -*- coding: utf-8 -*-
# @Time : 2019/2/25 4:13 PM
# @Author : scl
# @Email : 1163820757@qq.com
# @File : tianchipy.py
# @Software: PyCharm
import pandas as pd
import time
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier
from sklearn.metrics import f1_score,recall_score,precision_score,accuracy_score
'''
比赛链接：https://tianchi.aliyun.com/competition/information.htm?spm=5176.11165320.5678.2.45056f7cm4yHGQ&raceId=231631
'''
df = pd.read_csv('./datas/train.csv')
print(df.head())
df.info()
# 提取x y
Y = df['Label']
X = df.drop(['ID','V_Time','Label'],1,inplace=False)
print(X.head())
# 分割数据
x_train, x_test,y_train,y_test = train_test_split(X,Y,test_size=0.3,random_state=12)
print('训练样本数据数量:{}'.format(x_train.shape[0]))
print('测试样本数据数量:{}'.format(x_test.shape[1]))
print(y_train.value_counts())
t1 = time.time()
# 模型训练
lr = DecisionTreeClassifier(class_weight={0:1,1:5},random_state=12,max_depth=3)
lr.fit(x_train,y_train)
print('训练耗时%.4f{}s'.format(time.time()-t1))
# 训练数据集上的指标
print('----- 训练数据集上的指标 ----- ')
train_predict = lr.predict(x_train)
print('得分%.4f'%(f1_score(y_train,train_predict)))
print('召回率%.4f'%(recall_score(y_train,train_predict)))
print('精准率%.4f'%(precision_score(y_train,train_predict)))
print('准确率%.4f'%(accuracy_score(y_train,train_predict)))
print('---- 测试数据集上的指标 ------ ')
test_predict = lr.predict(x_test)
print('得分%.4f'%(f1_score(y_test,test_predict)))
print('召回率%.4f'%(recall_score(y_test,test_predict)))
print('精准率%.4f'%(precision_score(y_test,test_predict)))
print('准确率%.4f'%(accuracy_score(y_test,test_predict)))
# 加载预测数据集
prdf = pd.read_csv('./datas/pred.csv')
prX = prdf.drop(['ID','V_Time'],1, inplace=False)
print(prX.head())
## 输出预测结果 写入文件
result = pd.DataFrame()
result['ID'] = prdf['ID']
result['Label'] = lr.predict(prX)
result.to_csv('./datas/out/result.csv',index=False)
print('Done!!!')
'''
log 输出结果
/anaconda3/envs/mlenvment/bin/python3.7 /Users/long/Desktop/ml_worksapce/MlGitHubCode/MlWorkSpacePrj/项目/天池风险识别项目/tianchipy.py
ID V_Time V1 ... V29 V30 Label
0 254359 156699 -0.935008 ... 0.042881 0.771583 0
1 244959 152554 2.039188 ... 0.002934 -1.225918 0
2 79483 58048 -0.377984 ... 0.013761 -3.124638 0
3 164477 116748 1.985660 ... 0.006732 -2.550306 0
4 184542 126292 1.930330 ... 0.000068 0.139250 0
[5 rows x 33 columns]
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 100000 entries, 0 to 99999
Data columns (total 33 columns):
ID 100000 non-null int64
V_Time 100000 non-null int64
V1 100000 non-null float64
V2 100000 non-null float64
V3 100000 non-null float64
V4 100000 non-null float64
V5 100000 non-null float64
V6 100000 non-null float64
V7 100000 non-null float64
V8 100000 non-null float64
V9 100000 non-null float64
V10 100000 non-null float64
V11 100000 non-null float64
V12 100000 non-null float64
V13 100000 non-null float64
V14 100000 non-null float64
V15 100000 non-null float64
V16 100000 non-null float64
V17 100000 non-null float64
V18 100000 non-null float64
V19 100000 non-null float64
V20 100000 non-null float64
V21 100000 non-null float64
V22 100000 non-null float64
V23 100000 non-null float64
V24 100000 non-null float64
V25 100000 non-null float64
V26 100000 non-null float64
V27 100000 non-null float64
V28 100000 non-null float64
V29 100000 non-null float64
V30 100000 non-null float64
Label 100000 non-null int64
dtypes: float64(30), int64(3)
memory usage: 25.2 MB
V1 V2 V3 ... V28 V29 V30
0 -0.935008 0.820946 1.067777 ... 0.051355 0.042881 0.771583
1 2.039188 -0.264982 -1.235053 ... -0.055856 0.002934 -1.225918
2 -0.377984 0.917614 1.714673 ... -0.407508 0.013761 -3.124638
3 1.985660 -0.752667 -1.669258 ... -0.049041 0.006732 -2.550306
4 1.930330 2.563490 -4.759537 ... 0.109086 0.000068 0.139250
[5 rows x 30 columns]
训练样本数据数量:70000
测试样本数据数量:30
0 69789
1 211
Name: Label, dtype: int64
训练耗时%.4f1.0837130546569824s
----- 训练数据集上的指标 -----
得分0.8723
召回率0.8578
精准率0.8873
准确率0.9992
---- 测试数据集上的指标 ------
得分0.8000
召回率0.7865
精准率0.8140
准确率0.9988
V1 V2 V3 ... V28 V29 V30
0 -2.389003 0.508246 0.955227 ... -0.221525 0.035281 -2.741180
1 1.324636 0.095398 -0.105591 ... -0.015829 0.002406 -7.049453
2 -0.083895 0.543350 -0.244593 ... 0.086205 0.112603 -1.469655
3 2.027800 -0.080261 -1.167567 ... -0.068097 0.011066 -0.322347
4 -0.621263 -0.400795 1.599899 ... -0.167883 0.049452 -2.212872
[5 rows x 30 columns]
Done!!!
'''
