Sklearn笔记：度量和评分
——————————————————————————————————————————————————————
原文：3.3. Metrics and scoring: quantifying the quality of predictions — scikit-learn 0.22.2 documentation
主要函数概览
对应的数学公式(来源于：周志华《机器学习》)
- 准确率
\[准确率： {\rm{acc}}uracy = \sum\limits_{i = 1}^n {I({y_{true}} = {y_{pred}})}
\]
- Brier分数
\[Brie{r_{score}} = \frac{1}{N}\sum\limits_{t = 1}^N {({y_{true,t}} - {y_{pred,t}})}
\]
- 查准率、召回率、F指数
\[precision:P = \frac{{TP}}{{TP + FP}}{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {P_{{\rm{macro}}}} = \frac{1}{n}\sum\limits_{i = 1}^n {{P_i}}
\]
\[recall:R = \frac{{TP}}{{TP + FN}}{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {R_{macro}} = \frac{1}{n}\sum\limits_{i = 1}^n {{R_i}}
\]
\[{F_\beta }{\rm{ = }}\frac{{(1 + {\beta ^2})*P*R}}{{({\beta ^2}*P + R)}}{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {F_{\beta {\rm{ - macro}}}} = \frac{{(1 + {\beta ^2})*{P_{macro}}*{R_{macro}}}}{{({\beta ^2}*{P_{macro}} + {R_{macro}})}}
\]
- ROC
- 聚类指标
- 其它
