本例子是测试一些数据分析模型的R值,R值越接近1,表明该模型越适合分析该数据集.
本例子是在集成开发环境Aptana Studio 3 中创建 一个dataAnaly ,然后创建modelTest.py调用modelChose.py中的函数;在modelTest.py中需要import modelChose
格式:from 模块名 import 函数名1,函数名2....
![](https://images.cnblogs.com/OutliningIndicators/ContractedBlock.gif)
''' Created on 2015-1-19 @author: xuzhengzhu ''' #input files import xlrd,openpyxl import pandas as pd from sklearn import cross_validation from dataAnaly import modelChose from sklearn.metrics import r2_score import numpy as np file=pd.ExcelFile('e:\\report.xlsx') data=file.parse('Sheet1') n=len(data) #init data x=data[['myjg','tjg']] y=data['byjg'] models=['linear_model.SGDRegressor','GradientBoostingRegressor','RandomForestRegressor','AdaBoostRegressor','BaggingRegressor','linear_model.LinearRegression','linear_model.LogisticRegression','svm.svr','svm.NuSVR'] m=len(models) k=10 R2=np.zeros(k) z=2 count=0 modelCount=0 #lookup get model object for modelCount in range(m-1): clf=modelChose.modelChose(models[modelCount]) R2=np.zeros(k) count=0 #lookup folds for train_index,test_index in cross_validation.KFold(n-z,n_folds=k): x_train,x_test=x.ix[train_index],x.ix[test_index] y_train,y_test=y[train_index],y[test_index] clf.fit(x_train,y_train) y_predict=clf.predict(x_test); r2=r2_score(y_test,y_predict) #print 'computed %d time(s) and R square is:%f ' %(count+1,r2) R2[count]=r2 count+=1 print 'model choose is :',models[modelCount],'the mean of R2 is :',np.mean(R2) y_validation = clf.predict(x.ix[(n-z):n]) r2_val=r2_score(y.ix[(n-z):n],y_validation) print 'model choose is :',models[modelCount],'the validation ser R square is :%f ',r2_val #print pd.DataFrame({'y_true':y.ix[(n-z):n,],'y_validation':y_validation}) modelCount+=1
![](https://images.cnblogs.com/OutliningIndicators/ContractedBlock.gif)
''' Created on 2015-1-19 @author: xuzhengzhu ''' from sklearn.ensemble import BaggingRegressor from sklearn.ensemble import AdaBoostRegressor from sklearn.ensemble import RandomForestRegressor from sklearn.ensemble import GradientBoostingRegressor from sklearn import linear_model from sklearn.svm import SVR from sklearn.svm import NuSVR def modelChose(modelName): if(cmp(modelName,'linear_model.SGDRegressor')==0): #print modelName clf = linear_model.SGDRegressor() return clf elif (cmp(modelName,'GradientBoostingRegressor')==0): #print modelName clf = GradientBoostingRegressor() return clf elif (cmp(modelName,'RandomForestRegressor')==0): #print modelName clf = RandomForestRegressor() return clf elif (cmp(modelName,'AdaBoostRegressor')==0): #print modelName clf = AdaBoostRegressor() return clf elif (cmp(modelName,'BaggingRegressor')==0): #print modelName clf = BaggingRegressor() return clf elif (cmp(modelName,'linear_model.LinearRegression')==0): #print modelName clf = linear_model.LinearRegression() return clf elif (cmp(modelName,'linear_model.LogisticRegression')==0): #print modelName clf = linear_model.LogisticRegression() return clf elif (cmp(modelName,'svm.svr')==0): #print modelName clf = SVR() return clf elif (cmp(modelName,'svm.NuSVR')==0): #print modelName clf = NuSVR() return clf else: #print modelName,count,'dddd',models[count] return 1
测试结果:
model choose is : linear_model.SGDRegressor the mean of R2 is : -4.40149514377e+158
model choose is : linear_model.SGDRegressor the validation ser R square is :%f -1.69950873171e+175
model choose is : GradientBoostingRegressor the mean of R2 is : 0.06842532769
model choose is : GradientBoostingRegressor the validation ser R square is :%f -0.706828939678
model choose is : RandomForestRegressor the mean of R2 is : 0.0656454293629
model choose is : RandomForestRegressor the validation ser R square is :%f -1.62440546968
model choose is : AdaBoostRegressor the mean of R2 is : 0.0678670360111
model choose is : AdaBoostRegressor the validation ser R square is :%f -0.743162901308
model choose is : BaggingRegressor the mean of R2 is : 0.0913739612188
model choose is : BaggingRegressor the validation ser R square is :%f -1.11141498216
model choose is : linear_model.LinearRegression the mean of R2 is : 0.0976952970181
model choose is : linear_model.LinearRegression the validation ser R square is :%f -15.3631379961
model choose is : linear_model.LogisticRegression the mean of R2 is : -0.224099722992
model choose is : linear_model.LogisticRegression the validation ser R square is :%f 0.588585017836
model choose is : svm.svr the mean of R2 is : -0.243679440381
model choose is : svm.svr the validation ser R square is :%f -1.21033155027