关键词
随机森林分类器
5折交叉验证
ROC曲线
AUC
可视化
import matplotlib.pylab as plt
from scipy import interp
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import roc_curve,auc
from sklearn.model_selection import StratifiedKFold
import matplotlib.patches as patches
import numpy as np
import pandas as pd
# 导入数据,查看特征名称,标签数量和数据信息
data = pd .read_csv('./voice.csv')
print(data.columns)
label_value_count = data.label.value_counts()
print(label_value_count)
print(data.info())
Index(['meanfreq', 'sd', 'median', 'Q25', 'Q75', 'IQR', 'skew', 'kurt',
'sp.ent', 'sfm', 'mode', 'centroid', 'meanfun', 'minfun', 'maxfun',
'meandom', 'mindom', 'maxdom', 'dfrange', 'modindx', 'label'],
dtype='object')
male 1584
female 1584
Name: label, dtype: int64
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3168 entries, 0 to 3167
Data columns (total 21 columns):
meanfreq 3168 non-null float64
sd 3168 non-null float64
median 3168 non-null float64
Q25 3168 non-null float64
Q75 3168 non-null float64
IQR 3168 non-null float64
skew 3168 non-null float64
kurt 3168 non-null float64
sp.ent 3168 non-null float64
sfm 3168 non-null float64
mode 3168 non-null float64
centroid 3168 non-null float64
meanfun 3168 non-null float64
minfun 3168 non-null float64
maxfun 3168 non-null float64
meandom 3168 non-null float64
mindom 3168 non-null float64
maxdom 3168 non-null float64
dfrange 3168 non-null float64
modindx 3168 non-null float64
label 3168 non-null object
dtypes: float64(20), object(1)
memory usage: 519.9+ KB
None
可以看到label变量是对象(字符型),所以我们需要将它转换成数值型
# 转换规则 : male = 1, female = 0
dict = {'label':{'male':1,'female':0}}
data.replace(dict,inplace = True)
x = data.loc[:, data.columns != 'label']
y = data.loc[:,'label']
创建随机森林分类器并进行五折交叉验证
在k重交叉验证中,会将数据集划分为k个子集,重复k次。每次k子集中的一个子集用于测试集,而其他子集用于训练集。然后计算所有k个试验的平均准确度。
random_state = np.random.RandomState(0)
clf = RandomForestClassifier(random_state=random_state)
cv = StratifiedKFold(n_splits=5,shuffle=False)
在ROC曲线中,x轴是假阳性率,y轴是真阳性率
如果曲线图中的曲线靠近左上角则测试更准确。
Roc曲线得分高低取决于auc,即预测得分曲线下的计算区域
我们希望auc越接近于1,这代表预测结果越精确
fig1 = plt.figure(figsize=[12,12])
ax1 = fig1.add_subplot(111,aspect = 'equal')
ax1.add_patch(
patches.Arrow(0.45,0.5,-0.25,0.25,width=0.3,color='green',alpha = 0.5)
)
ax1.add_patch(
patches.Arrow(0.5,0.45,0.25,-0.25,width=0.3,color='red',alpha = 0.5)
)
tprs = []
aucs = []
mean_fpr = np.linspace(0,1,100)
i = 1
for train,test in cv.split(x,y):
prediction = clf.fit(x.iloc[train],y.iloc[train]).predict_proba(x.iloc[test])
fpr, tpr, t = roc_curve(y[test], prediction[:, 1])
tprs.append(interp(mean_fpr, fpr, tpr))
roc_auc = auc(fpr, tpr)
aucs.append(roc_auc)
plt.plot(fpr, tpr, lw=2, alpha=0.3, label='ROC fold %d (AUC = %0.2f)' % (i, roc_auc))
i= i+1
plt.plot([0,1],[0,1],linestyle = '--',lw = 2,color = 'black')
mean_tpr = np.mean(tprs, axis=0)
mean_auc = auc(mean_fpr, mean_tpr)
plt.plot(mean_fpr, mean_tpr, color='blue',
label=r'Mean ROC (AUC = %0.2f )' % (mean_auc),lw=2, alpha=1)
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('ROC')
plt.legend(loc="lower right")
plt.text(0.32,0.7,'More accurate area',fontsize = 12)
plt.text(0.63,0.4,'Less accurate area',fontsize = 12)
plt.show()
在本例中,ROC用于二分类。ROC主要用于二进制类,实际上也可以用于多分类。
喜欢
分享
or