将“ctree”对象中的“levels”属性提取到创建 Json 对象的函数中答案

【问题标题】：Extract the "levels" attribute from a "ctree" object into a Function that creates a Json object将“ctree”对象中的“levels”属性提取到创建 Json 对象的函数中
【发布时间】：2015-11-19 16:39:59
【问题描述】：

感谢以下问题： Converting toJson R object into a format that fits d3.js tree layout

我有一段很棒的代码可以为 d3.js 创建一个 Json 文件（见下文）

我知道如何从ctree 中提取分割点，如下所示：

library(party)
irisct <- ctree(Species ~ .,data = iris)

我只是不知道如何在下面的代码中实现它：

   #convert to Json that fits to d3.js layout####
get_ctree_parts <- function(x, ...)
{
  UseMethod("get_ctree_parts")
}

get_ctree_parts.BinaryTree <- function(x, ...)
{
  get_ctree_parts(attr(x, "tree"))
}

get_ctree_parts.SplittingNode <- function(x, ...)
{
  with(
    x,
    list(
      name = toString(nodeID),
      criteria=attr(psplit$splitpoint, "levels"),
      children   = list(get_ctree_parts(x$left),get_ctree_parts(x$right))

    )
  )
}

get_ctree_parts.TerminalNode <- function(x, ...)
{
  with(
    x,
    list(
      name     = paste(nodeID,
                       "weights",sum(weights),
                       "prediction",toString(paste("",toString(round(prediction,3)),"",sep=" ")),
                      # "criteria split",paste((attr(toString(psplit$splitpoint,levels)))), 
                       sep = " ")

    )
  )
}

toJSON(get_ctree_parts(irisct))

输出：

{"name":["1"],"criteria":{},"children":[{"name":["2 weights 50 prediction  1, 0, 0 "]},{"name":["3"],"criteria":{},"children":[{"name":["4"],"criteria":{},"children":[{"name":["5 weights 46 prediction  0, 0.978, 0.022 "]},{"name":["6 weights 8 prediction  0, 0.5, 0.5 "]}]},{"name":["7 weights 46 prediction  0, 0.022, 0.978 "]}]}]}

请注意，“标准”留空：{}，而我希望它们填充所有级别。

对此的任何帮助都会很棒！

【问题讨论】：

我认为paste((attributes(ct@tree$right$psplit$splitpoint)$levels)) 与此处无关

标签： json r

【解决方案1】：

我最终通过夹板节点得到了信息：

#convert to Json that fits to d3.js layout####
get_ctree_parts <- function(x, ...)
{
  UseMethod("get_ctree_parts")
}

get_ctree_parts.BinaryTree <- function(x, ...)
{
  get_ctree_parts(attr(x, "tree"))
}

get_ctree_parts.SplittingNode <- function(x, ...)
{
  with(
    x,
    list(
      name = toString(nodeID),
      variableNames=toString(x$psplit$variableName),
      criteria=toString(attributes(x$psplit$splitpoint)),
     # split   = paste(attributes(get_ctree_parts(x$left$psplit$splitpoint)),attributes(get_ctree_parts(x$right$psplit$splitpoint))),
      children   = list(get_ctree_parts(x$left),get_ctree_parts(x$right))

    )
  )
}

get_ctree_parts.TerminalNode <- function(x, ...)
{
  with(
    x,
    list(
      name     = paste(nodeID,
                       "weights",sum(weights),
                       "prediction",toString(paste("",toString(round(prediction,3)),"",sep=" ")),
                      # "criteria split",paste((attr(toString(psplit$splitpoint,levels)))), 
                       sep = " ")

    )
  )
}

【讨论】：