基于多源环境变量和随机森林模型的江西省耕地土壤pH值空间预测

钟骁勇, 李洪义, 郭冬艳, 谢模典, 赵婉如, 胡碧峰

Spatial distribution prediction of soil pH in arable land of Jiangxi Province based on multi-source environmental variables and the random forest model

ZHONG Xiaoyong, LI Hongyi, GUO Dongyan, XIE Modian, ZHAO Wanru, HU Bifeng

表2 RF模型中节点分裂次数和决策树数量的筛选

Tab.2 Optimization of critical parameters for RF model

组别 mtry ntree 训练集r 验证集r 1组 2 300 0.969 0.551 600 0.970 0.558 900 0.970 0.556 1 200 0.970 0.559 1 500 0.970 0.555 2组 4 300 0.970 0.567 600 0.970 0.574 900 0.970 0.573 1 200 0.971 0.573 1 500 0.971 0.572 3组 6 300 0.969 0.579 600 0.969 0.578 900 0.970 0.579 1 200 0.970 0.580 1 500 0.970 0.580 4组 8 300 0.969 0.582 600 0.970 0.582 900 0.970 0.583 1 200 0.970 0.582 1 500 0.969 0.581