如下,还有生产者精度和用户精度 Class Reference Classified Number Producers Users Name Totals Totals Correct Accuracy Accuracy ---------- ---------- ---------- ------- --------- ----- Unclassified 0 0 0 --- --- forest 3 1 1 33.33% 100.00% grass 5 4 3 60.00% 75.00% yanshi 22 22 20 90.91% 90.91% ruotu 9 8 5 55.56% 62.50% nongzuowu 11 15 10 90.91% 66.67% river 0 0 0 --- --- Totals 50 50 39 Overall Classification Accuracy = 78.00% 生产则精度=正确分类数/某类别总数用户精度=正确分类数/(正确分类数+错分到某类别总数)当然是生产则精度和用户精度都大最好 当然是最大似然法最好,马氏距离、最小距离法都是最大似然法的简化。平行管道法也不如最大似然法。