使用sklearn对多分类的每个类别进行指标评价操作

灰太狼 2023-02-19 06:25 12阅读 0赞

更多python视频教程请到菜鸟教程https://www.piaodoo.com/

今天晚上，笔者接到客户的一个需要，那就是：对多分类结果的每个类别进行指标评价，也就是需要输出每个类型的精确率（precision），召回率（recall）以及F1值（F1-score）。

对于这个需求，我们可以用sklearn来解决，方法并没有难，笔者在此仅做记录，供自己以后以及读者参考。

我们模拟的数据如下：

> y\_true = \['北京', '上海', '成都', '成都', '上海', '北京', '上海', '成都', '北京', '上海'\]  
> y\_pred = \['北京', '上海', '成都', '上海', '成都', '成都', '上海', '成都', '北京', '上海'\]

其中y\_true为真实数据，y\_pred为多分类后的模拟数据。使用sklearn.metrics中的classification\_report即可实现对多分类的每个类别进行指标评价。

示例的Python代码如下：

# -*- coding: utf-8 -*-
    from sklearn.metrics import classification_report

y\_true = \[‘北京’, ‘上海’, ‘成都’, ‘成都’, ‘上海’, ‘北京’, ‘上海’, ‘成都’, ‘北京’, ‘上海’\]  
y\_pred = \[‘北京’, ‘上海’, ‘成都’, ‘上海’, ‘成都’, ‘成都’, ‘上海’, ‘成都’, ‘北京’, ‘上海’\]

t = classification\_report(y\_true, y\_pred, target\_names=\[‘北京’, ‘上海’, ‘成都’\])

print(t)

输出结果如下：

precision  recall f1-score  support
      北京    0.75   0.75   0.75     4
     上海    1.00   0.67   0.80     3
     成都    0.50   0.67   0.57     3

accuracy 0.70 10  
macro avg 0.75 0.69 0.71 10  
weighted avg 0.75 0.70 0.71 10

需要注意的是，输出的结果数据类型为str，如果需要使用该输出结果，则可将该方法中的output\_dict参数设置为True，此时输出的结果如下:

{‘北京': {‘precision': 0.75, ‘recall': 0.75, ‘f1-score': 0.75, ‘support': 4},
    ‘上海': {‘precision': 1.0, ‘recall': 0.6666666666666666, ‘f1-score': 0.8, ‘support': 3},
    ‘成都': {‘precision': 0.5, ‘recall': 0.6666666666666666, ‘f1-score': 0.5714285714285715, ‘support': 3},
    ‘accuracy': 0.7,
    ‘macro avg': {‘precision': 0.75, ‘recall': 0.6944444444444443, ‘f1-score': 0.7071428571428572, ‘support': 10},
    ‘weighted avg': {‘precision': 0.75, ‘recall': 0.7, ‘f1-score': 0.7114285714285715, ‘support': 10}}

使用confusion\_matrix方法可以输出该多分类问题的混淆矩阵，代码如下：

from sklearn.metrics import confusion_matrix
    y_true = ['北京', '上海', '成都', '成都', '上海', '北京', '上海', '成都', '北京', '上海']
    y_pred = ['北京', '上海', '成都', '上海', '成都', '成都', '上海', '成都', '北京', '上海']
    print(confusion_matrix(y_true, y_pred, labels = ['北京', '上海', '成都']))

输出结果如下：

[[2 0 1]
     [0 3 1]
     [0 1 2]]

为了将该混淆矩阵绘制成图片，可使用如下的Python代码：

# -*- coding: utf-8 -*-
    # author: Jclian91
    # place: Daxing Beijing
    # time: 2019-11-14 21:52

from sklearn.metrics import confusion\_matrix  
import matplotlib.pyplot as plt  
import matplotlib as mpl

# 支持中文字体显示, 使用于Mac系统 #

zhfont=mpl.font\_manager.FontProperties(fname="/Library/Fonts/Songti.ttc")

classes = \[‘北京’, ‘上海’, ‘成都’\]  
confusion = confusion\_matrix(y\_true, y\_pred)

# 绘制热度图 #

plt.imshow(confusion, cmap=plt.cm.Greens)  
indices = range(len(confusion))  
plt.xticks(indices, classes, fontproperties=zhfont)  
plt.yticks(indices, classes, fontproperties=zhfont)  
plt.colorbar()  
plt.xlabel(‘y\_pred’)  
plt.ylabel(‘y\_true’)

# 显示数据 #

for first\_index in range(len(confusion)):  
for second\_index in range(len(confusion\[first\_index\])):  
plt.text(first\_index, second\_index, confusion\[first\_index\]\[second\_index\])

# 显示图片 #

plt.show()

生成的混淆矩阵图片如下：

![20200611143523.jpg][]

**补充知识：****python Sklearn实现xgboost的二分类和多分类**

**二分类：**

train2.txt的格式如下：

![20200611143733.jpg][]

import numpy as np
    import pandas as pd
    import sklearn
    from sklearn.cross_validation import train_test_split,cross_val_score
    from xgboost.sklearn import XGBClassifier
    from sklearn.metrics import precision_score,roc_auc_score

min\_max\_scaler = sklearn.preprocessing.MinMaxScaler(feature\_range=(-1,1))  
resultX = \[\]  
resultY = \[\]  
with open("./train\_data/train2.txt",‘r’) as rf:  
train\_lines = rf.readlines()  
for train\_line in train\_lines:  
train\_line\_temp = train\_line.split(",")  
train\_line\_temp = map(float, train\_line\_temp)  
line\_x = train\_line\_temp\[1:-1\]  
line\_y = train\_line\_temp\[-1\]  
resultX.append(line\_x)  
resultY.append(line\_y)

X = np.array(resultX)  
Y = np.array(resultY)  
X = min\_max\_scaler.fit\_transform(X)  
X\_train,X\_test, Y\_train, Y\_test = train\_test\_split(X,Y,test\_size=0.3)

xgbc = XGBClassifier()  
xgbc.fit(X\_train,Y\_train)  
pre\_test = xgbc.predict(X\_test)

auc\_score = roc\_auc\_score(Y\_test,pre\_test)  
pre\_score = precision\_score(Y\_test,pre\_test)

print(“xgb\_auc\_score:”,auc\_score)  
print(“xgb\_pre\_score:”,pre\_score)

多分类：有19种分类其中正常0，异常1~18种。数据格式如下：

![20200611143842.jpg][]

# -*- coding:utf-8 -*-
    from sklearn import datasets
    from sklearn.multiclass import OneVsRestClassifier
    from sklearn.svm import LinearSVC
    from sklearn.cross_validation import train_test_split,cross_val_score
    from sklearn.svm import SVC
    from sklearn.linear_model import LogisticRegression
    from xgboost.sklearn import XGBClassifier
    import sklearn
    import numpy as np
    from sklearn.preprocessing import OneHotEncoder
    from sklearn.metrics import precision_score,roc_auc_score
    min_max_scaler = sklearn.preprocessing.MinMaxScaler(feature_range=(-1,1))

resultX = \[\]  
resultY = \[\]  
with open("…/train\_data/train\_multi\_class.txt",‘r’) as rf:  
train\_lines = rf.readlines()  
for train\_line in train\_lines:  
train\_line\_temp = train\_line.split(",")  
train\_line\_temp = map(float, train\_line\_temp) \# 转化为浮点数  
line\_x = train\_line\_temp\[1:-1\]  
line\_y = train\_line\_temp\[-1\]  
resultX.append(line\_x)  
resultY.append(line\_y)

X = np.array(resultX)  
Y = np.array(resultY)

\#fit\_transform(partData)对部分数据先拟合fit，找到该part的整体指标，如均值、方差、最大值最小值等等（根据具体转换的目的），然后对该partData进行转换transform，从而实现数据的标准化、归一化等等。。  
X = min\_max\_scaler.fit\_transform(X)  
\#通过OneHotEncoder函数将Y值离散化成19维，例如3离散成000000···100

Y = OneHotEncoder(sparse = False).fit\_transform(Y.reshape(-1,1))  
X\_train,X\_test, Y\_train, Y\_test = train\_test\_split(X,Y,test\_size=0.2)

model = OneVsRestClassifier(XGBClassifier(),n\_jobs=2)  
clf = model.fit(X\_train, Y\_train)

pre\_Y = clf.predict(X\_test)  
test\_auc2 = roc\_auc\_score(Y\_test,pre\_Y)\#验证集上的auc值  
print (“xgb\_muliclass\_auc:”,test\_auc2)

以上这篇使用sklearn对多分类的每个类别进行指标评价操作就是小编分享给大家的全部内容了，希望能给大家一个参考，也希望大家多多

茂名论坛https://www.hnthzk.com/

化州橘红http://www.sfkyty.com/

茂名论坛http://www.nrso.net/

源码搜藏网http://www.cntkd.net/

茂名市高级技工学校(茂名一技)http://www.szsyby.net/茂名一技https://www.mmbbu.com/

[20200611143523.jpg]: https://img.jbzj.com/file_images/article/202006/20200611143523.jpg
[20200611143733.jpg]: https://img.jbzj.com/file_images/article/202006/20200611143733.jpg
[20200611143842.jpg]: https://img.jbzj.com/file_images/article/202006/20200611143842.jpg