机器学习Note：性能度量
参考西瓜书，仅当做自己学习笔记使用。
1. 精度 vs 错误率
- 精度: a c c = 1 n ∑ i = 1 n 1 { f ( x i ) = y i } = ∫ x 1 f ( x ) = y p ( x ) d x acc=\frac{1}{n}\sum\limits_{i=1}^n1_{\{f(x_i)=y_i\}}=\int_{x}1_{f(x)=y}p(x)\mathrm{d}x acc=n1?i=1∑n?1{f(xi?)=yi?}?=∫x?1f(x)=y?p(x)dx
- 错误率: e r r o r ? r a t e = 1 n ∑ i = 1 n 1 { f ( x i ) ≠ y i } = ∫ x 1 { f ( x ) ≠ y } p ( x ) d x \mathrm{error\ rate}=\frac{1}{n}\sum\limits_{i=1}^n1_{\{f(x_i)\ne y_i\}}=\int_x1_{\{f(x)\ne y\}}p(x)\mathrm{d}x error?rate=n1?i=1∑n?1{f(xi?)?=yi?}?=∫x?1{f(x)?=y}?p(x)dx
2. 准确率precision vs. 召回率recall
- p r e c i s i o n = T P T P + F P precision=\frac{TP}{TP+FP} precision=TP+FPTP?
- r e c a l l = T P T P + F N recall=\frac{TP}{TP+FN} recall=TP+FNTP?
- F 1 = 2 ? p r e c i s i o n ? r e c a l l p r e c i s i o n + r e c a l l F1=\frac{2\cdot precision\cdot recall}{precision+recall} F1=precision+recall2?precision?recall?
- 如果有多个精确率结果和召回率结果，则有macro结果
- m a c r o ? p r e c i s i o n = 1 n ∑ i = 1 n p r e c i s i o n i \mathrm{macro\ precision}=\frac{1}{n}\sum\limits_{i=1}^nprecision_i macro?precision=n1?i=1∑n?precisioni?
- m a c r o ? r e c a l l = 1 n ∑ i = 1 n r e c a l l i \mathrm{macro\ recall}=\frac{1}{n}\sum\limits_{i=1}^nrecall_i macro?recall=n1?i=1∑n?recalli?