from sklearn import datasets
# 过滤法选择特征，标准有方差，卡方检验，互信息，相关系数
# 方差选择
from sklearn.feature_selection import VarianceThreshold
#卡方：f_classif，F分布；chi2,卡方；f_regression,回归。
from sklearn.feature_selection import SelectKBest,f_classif,chi2,f_regression#按照个数保留特征
from sklearn.feature_selection import SelectPercentile#按照比例保留特征
#相关系数
from scipy.stats import pearsonr
#互信息
from sklearn import metrics as mr
#用法 mr.mutual_info_score(label,x)
#导入数据集
wine=datasets.load_wine()
x,y=wine.data,wine.target
#方差选择
#不能先将特征规划化，不然方差都为1
wine_x=x
selector=VarianceThreshold(1)
selector.fit(wine_x)
print('Variances is %s'%selector.variances_)
print('After transform is \n%s'%selector.transform(wine_x))
print('The surport is %s'%selector.get_support(True))#如果为True那么返回的是被选中的特征的下标
print('The surport is %s'%selector.get_support(False))#如果为FALSE那么返回的是布尔类型的列表，反应是否选中这列特征
#print('After reverse transform is \n%s'%selector.inverse_transform(selector.transform(wine_x)))
#方差太不好控制了，阀值难以取舍，并且数据规范化不能在方差取特征之前进行，不然方差很小，难以取出特征Variances is [6.55359730e-01 1.24100408e+00 7.48418003e-02 1.10900306e+01
2.02843328e+02 3.89489032e-01 9.92113512e-01 1.54016191e-02
3.25754248e-01 5.34425585e+00 5.19514497e-02 5.01254463e-01
9.86096010e+04]
The surport is [False True False True True False False False False True False False
True]
#chi2检验数组必须为非负数，标准化会产生负数
from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import MinMaxScaler
#规范化可以去除犅量的影响，提升模型效率
sc=MinMaxScaler()
#sc=MinMaxScaler()
wine_x=sc.fit_transform(x)
print('before transform:\n',wine_x)
selector2=SelectKBest(score_func=chi2,k=8)
selector2.fit(wine_x,y)
#chi2值
print('scores_:\n',selector2.scores_)
#p值
print('pvalues_:',selector2.pvalues_)
#特征编号，True返回数组索引，False返回各列特征选择情况，True为选中，False为未选中
print('selected index:',selector2.get_support(True))
#未进行数据规范化，选中特征索引为[ 1 3 4 5 6 9 11 12]
#规范化后，选中特征为[ 0 1 5 6 9 10 11 12]before transform:
[[0.84210526 0.1916996 0.57219251 ... 0.45528455 0.97069597 0.56134094]
[0.57105263 0.2055336 0.4171123 ... 0.46341463 0.78021978 0.55064194]
[0.56052632 0.3201581 0.70053476 ... 0.44715447 0.6959707 0.64693295]
...
[0.58947368 0.69960474 0.48128342 ... 0.08943089 0.10622711 0.39728959]
[0.56315789 0.36561265 0.54010695 ... 0.09756098 0.12820513 0.40085592]
[0.81578947 0.66403162 0.73796791 ... 0.10569106 0.12087912 0.20114123]]
scores_:
[ 9.45400066 8.11858304 0.9346697 3.31958248 1.64131331 9.40177031
16.05097687 5.34608154 3.9813523 12.45315712 8.44940247 16.67767772
18.79203469]
pvalues_: [8.85298728e-03 1.72612440e-02 6.26670214e-01 1.90178678e-01
4.40142537e-01 9.08722992e-03 3.27020260e-04 6.90419652e-02
1.36603030e-01 1.97620180e-03 1.46297049e-02 2.39049752e-04
8.30541840e-05]
selected index: [ 0 1 5 6 9 10 11 12]
#F检验检验数组可以为负值
from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import MinMaxScaler
#规范化可以去除犅量的影响，提升模型效率
#sc=StandardScaler()
#sc=MinMaxScaler()
#wine_x=sc.fit_transform(x)
print('before transform:\n',x)
selector3=SelectKBest(score_func=f_classif,k=8)
selector3.fit(x,y)
#chi2值
print('scores_:\n',selector3.scores_)
#p值
print('pvalues_:',selector3.pvalues_)
#特征编号，True返回数组索引，False返回各列特征选择情况，True为选中，False为未选中
print('selected index:',selector3.get_support(True))
#未进行数据规范化，选中特征索引为[ 0 1 5 6 9 10 11 12]
#标准化后，选中特征为[ 0 1 5 6 9 10 11 12]before transform:
[[1.423e+01 1.710e+00 2.430e+00 ... 1.040e+00 3.920e+00 1.065e+03]
[1.320e+01 1.780e+00 2.140e+00 ... 1.050e+00 3.400e+00 1.050e+03]
[1.316e+01 2.360e+00 2.670e+00 ... 1.030e+00 3.170e+00 1.185e+03]
...
[1.327e+01 4.280e+00 2.260e+00 ... 5.900e-01 1.560e+00 8.350e+02]
[1.317e+01 2.590e+00 2.370e+00 ... 6.000e-01 1.620e+00 8.400e+02]
[1.413e+01 4.100e+00 2.740e+00 ... 6.100e-01 1.600e+00 5.600e+02]]
scores_:
[135.07762424 36.94342496 13.3129012 35.77163741 12.42958434
93.73300962 233.92587268 27.57541715 30.27138317 120.66401844
101.31679539 189.97232058 207.9203739 ]
pvalues_: [3.31950380e-36 4.12722880e-14 4.14996797e-06 9.44447294e-14
8.96339544e-06 2.13767002e-28 3.59858583e-50 3.88804090e-11
5.12535874e-12 1.16200802e-33 5.91766222e-30 1.39310496e-44
5.78316836e-47]
selected index: [ 0 1 5 6 9 10 11 12]
#F检验检验数组可以为负值
from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import MinMaxScaler
#规范化可以去除犅量的影响，提升模型效率
#sc=StandardScaler()
#sc=MinMaxScaler()
#wine_x=sc.fit_transform(x)
print('before transform:\n',x)
#使用保留百分比个数，这里保留50%
selector4=SelectPercentile(score_func=f_classif,percentile=50)
selector4.fit(x,y)
#chi2值
print('scores_:\n',selector4.scores_)
#p值
print('pvalues_:',selector4.pvalues_)
#特征编号，True返回数组索引，False返回各列特征选择情况，True为选中，False为未选中
print('selected index:',selector4.get_support(True))
#未进行数据规范化，选中特征索引为[ 0 1 5 6 9 10 11 12]
#标准化后，选中特征为[ 0 1 5 6 9 10 11 12]before transform:
[[1.423e+01 1.710e+00 2.430e+00 ... 1.040e+00 3.920e+00 1.065e+03]
[1.320e+01 1.780e+00 2.140e+00 ... 1.050e+00 3.400e+00 1.050e+03]
[1.316e+01 2.360e+00 2.670e+00 ... 1.030e+00 3.170e+00 1.185e+03]
...
[1.327e+01 4.280e+00 2.260e+00 ... 5.900e-01 1.560e+00 8.350e+02]
[1.317e+01 2.590e+00 2.370e+00 ... 6.000e-01 1.620e+00 8.400e+02]
[1.413e+01 4.100e+00 2.740e+00 ... 6.100e-01 1.600e+00 5.600e+02]]
scores_:
[135.07762424 36.94342496 13.3129012 35.77163741 12.42958434
93.73300962 233.92587268 27.57541715 30.27138317 120.66401844
101.31679539 189.97232058 207.9203739 ]
pvalues_: [3.31950380e-36 4.12722880e-14 4.14996797e-06 9.44447294e-14
8.96339544e-06 2.13767002e-28 3.59858583e-50 3.88804090e-11
5.12535874e-12 1.16200802e-33 5.91766222e-30 1.39310496e-44
5.78316836e-47]
selected index: [ 0 6 9 10 11 12]
C:\ProgramData\Anaconda3\lib\site-packages\scipy\stats\stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval
#相关系数可以计算两种，特征之间的相关系数，特征与目标的相关系数
#特征之间的相关系数，用pandas的corr可以计算
import pandas as pd
import matplotlib.pyplot as plt
col=['Alcohol','Malic acid','Ash','Alcalinity of ash','Magnesium','Total phenols','Flavanoids','Nonflavanoid phenols',
'Proanthocyanins','Color intensity','Hue','OD280/OD315 of diluted wines','Proline']
wine_df=pd.DataFrame(x,columns=col)
colormap=plt.cm.viridis
import seaborn as sns
plt.figure(figsize=(8,8))
plt.title('Pearson Correlation of Features',y=1.05,size=15)
sns.heatmap(wine_df.corr(),linewidths=0.1,vmax=1,square=True,cmap=colormap,linecolor='white',annot=True)
plt.show()
#目标值与特征间相关系数，取前最大的k项
from scipy.stats import pearsonr
corr=[]
p=[]
for i in wine_df.columns:
re=list(pearsonr(wine_df[i],y))
corr_value,p_value=re[0],re[1]
corr.append(abs(corr_value))
p.append(p_value)
corr=np.array(corr)
p=np.array(p)
chose=np.argsort(-corr)[:8]
print(chose)
print(np.sort(chose))
#结果为[ 3 5 6 7 8 10 11 12[ 6 11 5 12 10 3 8 7]
[ 3 5 6 7 8 10 11 12]
#互信息
Mi={}
x_T=x.T
for i in range(len(x[0])):
Mi[i]=mr.mutual_info_score(x.T[i],y)
X_MIF = [each for each in Mi if Mi[each]>=sorted(Mi.values(),reverse=True)[8]]
print('选择的特征对应列序号：',X_MIF)
结果为：[0, 1, 5, 6, 8, 9, 10, 11, 12]选择的特征对应列序号： [0, 1, 5, 6, 8, 9, 10, 11, 12]
