使用随机森林匹配

Question

作为随机森林世界的新手，我有一个（基本）问题。继赵，苏，婷婷和范（2016）之后，我正在尝试以下程序：

For b = 1, ..., B, do

   Using all data, grow a tree with treatment as output and all other covariates as inputs.

   Compute % of treated observations s_i for the terminal node to which i belongs.

   Update S_i = S_i + s_i

end

prop_score = S_i / B

要运行此算法，我使用randomForest函数：

 model1 = randomForest(treatment ~ ., data = dataset, keep.forest = T, ntree = 100, na.action = na.omit)

话虽如此，您能告诉我如何获取每个 i 属于某个终端节点的信息吗？还有，任何推荐的包处理相当大的数据？除非我将树木数量限制为少量，否则randomForest会冻结。