多分类：准确率(Accuracy) 精准率(Precision) 召回率(Recall) F1值(F1_score)		预测
		风险级别1	风险级别2	风险级别3	风险级别4
实际	风险级别1	T11	F12	F13	F14
	风险级别2	F21	T22	F23	F24
	风险级别3	F31	F32	T33	F34
	风险级别4	F41	F42	F43	T44
$模型的准确率 (Accuracy) = \sum_{i}^{4} \sum_{j}^{4} T_{i j} / (\sum_{i}^{4} \sum_{j}^{4} T_{i j} + \sum_{i}^{4} \sum_{j}^{4} F_{i j})$ $风险级别 1 的精准率 (Precision) = T_{11} / (T_{11} + \sum_{i}^{4} F_{i 1})$ $风险级别 1 的召回率 (Recall) = T_{11} / (T_{11} + \sum_{j}^{4} F_{1 j})$ $F1_score = \frac{2 * Precision * Recall}{Precision + Recall}$