【MATLAB第70期】基于MATLAB的LightGbm(LGBM)梯度增强决策树多输入单输出分类预测模型(全网首发,敬请期待)
(LGBM)是一种基于梯度增强决策树(GBDT)算法。
基于MATLAB的LightGbm即将研究测试上线。
下一个研究对象: ABCBOOST模型
一、效果展示
二、数据设置:
28输入1输出,2分类模型
三、预测结果
clike
[ 0] train l2 0.243524 train auc 0.768088 valid-1 l2 0.242963 valid-1 auc 0.755797
[ 1] train l2 0.239299 train auc 0.784627 valid-1 l2 0.239442 valid-1 auc 0.755071
[ 2] train l2 0.235559 train auc 0.797551 valid-1 l2 0.235933 valid-1 auc 0.777275
[ 3] train l2 0.231220 train auc 0.801937 valid-1 l2 0.231946 valid-1 auc 0.778678
[ 4] train l2 0.227275 train auc 0.802404 valid-1 l2 0.228124 valid-1 auc 0.781830
[ 5] train l2 0.223873 train auc 0.808435 valid-1 l2 0.225515 valid-1 auc 0.781387
[ 6] train l2 0.221039 train auc 0.812120 valid-1 l2 0.222806 valid-1 auc 0.789441
[ 7] train l2 0.217817 train auc 0.814248 valid-1 l2 0.220040 valid-1 auc 0.789095
[ 8] train l2 0.214815 train auc 0.816482 valid-1 l2 0.217652 valid-1 auc 0.789990
[ 9] train l2 0.211890 train auc 0.817649 valid-1 l2 0.214958 valid-1 auc 0.792046
[ 10] train l2 0.209273 train auc 0.819593 valid-1 l2 0.212286 valid-1 auc 0.795867
[ 11] train l2 0.206929 train auc 0.821450 valid-1 l2 0.210131 valid-1 auc 0.797472
[ 12] train l2 0.204648 train auc 0.823188 valid-1 l2 0.208102 valid-1 auc 0.799689
[ 13] train l2 0.202647 train auc 0.824870 valid-1 l2 0.206802 valid-1 auc 0.798560
[ 14] train l2 0.200606 train auc 0.826843 valid-1 l2 0.205296 valid-1 auc 0.798826
bestIteration: 12
rmse of prediction: 0.4584