sklearn分类任务性能度量——以MNIST为例的精度、查准率、查全率、F1、ROC、AUC
https://blog.csdn.net/lly1122334/article/details/89294642
python + sklearn ︱分类效果评估——acc、recall、F1、ROC、回归、距离
https://blog.csdn.net/sinat_26917383/article/details/75199996#averageNone_micro_macro_weighted_samples2_52
sklearn中的模型评估
http://d0evi1.com/sklearn/model_evaluation/
scikit-learn官网
https://scikit-learn.org/stable/modules/model_evaluation.html