SKlearn模型评估方法

准确率

1.accuracy_score

# 准确率
import numpy as np
from sklearn.metrics import accuracy_score
y_pred = [0, 2, 1, 3,9,9,8,5,8]
y_true = [0, 1, 2, 3,2,6,3,5,9]

accuracy_score(y_true, y_pred)
Out[127]: 0.33333333333333331

accuracy_score(y_true, y_pred, normalize=False)  # 类似海明距离，每个类别求准确后，再求微平均
Out[128]: 3

2.metrics

• 宏平均比微平均更合理，但也不是说微平均一无是处，具体使用哪种评测机制，还是要取决于数据集中样本分布
• 宏平均（Macro-averaging），是先对每一个类统计指标值，然后在对所有类求算术平均值。
• 微平均（Micro-averaging），是对数据集中的每一个实例不分类别进行统计建立全局混淆矩阵，然后计算相应指标。

from sklearn import metrics
metrics.precision_score(y_true, y_pred, average='micro')  # 微平均，精确率
Out[130]: 0.33333333333333331

metrics.precision_score(y_true, y_pred, average='macro')  # 宏平均，精确率
Out[131]: 0.375

metrics.precision_score(y_true, y_pred, labels=[0, 1, 2, 3], average='macro')  # 指定特定分类标签的精确率
Out[133]: 0.5

• 其中 average 参数有五种：(None, ‘micro’, ‘macro’, ‘weighted’, ‘samples’)
  召回率

metrics.recall_score(y_true, y_pred, average='micro')
Out[134]: 0.33333333333333331

metrics.recall_score(y_true, y_pred, average='macro')
Out[135]: 0.3125

F1

metrics.f1_score(y_true, y_pred, average='weighted')
Out[136]: 0.37037037037037035

F2

根据公式计算

from sklearn.metrics import precision_score, recall_score
def calc_f2(label, predict):
    p = precision_score(label, predict)
    r = recall_score(label, predict)
    f2_score = 5*p*r / (4*p + r)
    return f2_score

混淆矩阵

from sklearn.metrics import confusion_matrix
confusion_matrix(y_true, y_pred)

Out[137]:
array([[1, 0, 0, ..., 0, 0, 0],
       [0, 0, 1, ..., 0, 0, 0],
       [0, 1, 0, ..., 0, 0, 1],
       ...,
       [0, 0, 0, ..., 0, 0, 1],
       [0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 1, 0]])

分类报告

包含：precision/recall/fi-score/均值/分类个数

# 分类报告：precision/recall/fi-score/均值/分类个数
 from sklearn.metrics import classification_report
 y_true = [0, 1, 2, 2, 0]
 y_pred = [0, 0, 2, 2, 0]
 target_names = ['class 0', 'class 1', 'class 2']
 print(classification_report(y_true, y_pred, target_names=target_names))

输出

              precision    recall  f1-score   support

    class 0       0.67      1.00      0.80         2
    class 1       0.00      0.00      0.00         1
    class 2       1.00      1.00      1.00         2

avg / total       0.67      0.80      0.72         5

kappa score

• kappa score 是一个介于(-1, 1)之间的数. score>0.8 意味着好的分类；0 或更低意味着不好（实际是随机标签）

from sklearn.metrics import cohen_kappa_score
y_true = [2, 0, 2, 2, 0, 1]
y_pred = [0, 0, 2, 2, 0, 2]
cohen_kappa_score(y_true, y_pred)

• ROC 1.计算 ROC 值

import numpy as np
from sklearn.metrics import roc_auc_score
y_true = np.array([0, 0, 1, 1])
y_scores = np.array([0.1, 0.4, 0.35, 0.8])
roc_auc_score(y_true, y_scores)

• 2.ROC 曲线

y = np.array([1, 1, 2, 2])
scores = np.array([0.1, 0.4, 0.35, 0.8])
fpr, tpr, thresholds = roc_curve(y, scores, pos_label=2)

• 海明距离

from sklearn.metrics import hamming_loss
y_pred = [1, 2, 3, 4]
y_true = [2, 2, 3, 4]
hamming_loss(y_true, y_pred)
0.25

Jaccard 距离

import numpy as np
from sklearn.metrics import jaccard_similarity_score
y_pred = [0, 2, 1, 3,4]
y_true = [0, 1, 2, 3,4]
jaccard_similarity_score(y_true, y_pred)
0.5
jaccard_similarity_score(y_true, y_pred, normalize=False)
2

可释方差值（Explained variance score）

from sklearn.metrics import explained_variance_score
y_true = [3, -0.5, 2, 7]
y_pred = [2.5, 0.0, 2, 8]
explained_variance_score(y_true, y_pred)

平均绝对误差（Mean absolute error）

from sklearn.metrics import mean_absolute_error
y_true = [3, -0.5, 2, 7]
y_pred = [2.5, 0.0, 2, 8]
mean_absolute_error(y_true, y_pred)

均方误差（Mean squared error）

from sklearn.metrics import mean_squared_error
y_true = [3, -0.5, 2, 7]
y_pred = [2.5, 0.0, 2, 8]
mean_squared_error(y_true, y_pred)

中值绝对误差（Median absolute error）

from sklearn.metrics import median_absolute_error
y_true = [3, -0.5, 2, 7]
y_pred = [2.5, 0.0, 2, 8]
median_absolute_error(y_true, y_pred)

R 方值，确定系数

from sklearn.metrics import r2_score
y_true = [3, -0.5, 2, 7]
y_pred = [2.5, 0.0, 2, 8]
r2_score(y_true, y_pred)

参考文献

• python + sklearn ︱分类效果评估——acc、recall、F1、ROC、回归、距离
• sklearn 中的模型评估