Question

我根据here中的示例使用R构建了一个条件推理树分析。我正在运行的代码如下：

fit <- ctree(Status~Amount.Converted + Campaign.Type + Region,
         data = d_c)

我不明白的是当我查看'fit'值时统计数据的含义。我尝试查看帮助和文件here，但我无法理解标准和统计意味着什么。以下是输出的示例：

> fit

     Conditional inference tree with 10 terminal nodes

Response:  Status 
Inputs:  Amount.Converted, Campaign.Type, Region 
Number of observations:  5822 

1) Region == {FR, IT, N. Africa, Rest of Africa, Turkey}; criterion = 1, statistic = 261.91
  2) Amount.Converted <= 12486.57481; criterion = 1, statistic = 42.64
    3) Region == {IT, N. Africa, Rest of Africa, Turkey}; criterion = 1, statistic = 50.451
      4) Region == {N. Africa, Rest of Africa}; criterion = 0.99, statistic = 13.683
        5)*  weights = 66 
      4) Region == {IT, Turkey}
        6)*  weights = 843 
    3) Region == {FR}
      7)*  weights = 908 
  2) Amount.Converted > 12486.57481
    8) Amount.Converted <= 39800; criterion = 0.998, statistic = 11.339
      9)*  weights = 9 
    8) Amount.Converted > 39800
      10)*  weights = 11 
1) Region == {Benelux, cis, East EU, DE, SP, Mid. East, Nordics, UK}
  11) Campaign.Type == {Rental}; criterion = 1, statistic = 49.144
    12) Region == {Benelux, UK}; criterion = 1, statistic = 22.721
      13)*  weights = 25 
    12) Region == {DE, SP}
      14)*  weights = 37 
  11) Campaign.Type == {Retail}
    15) Region == {Benelux, East EU, SP, Mid. East}; criterion = 1, statistic = 37.237
      16) Amount.Converted <= 2734.76; criterion = 0.986, statistic = 8.023
        17)*  weights = 1179 
      16) Amount.Converted > 2734.76
        18)*  weights = 311 
    15) Region == {cis, DE, Nordics, UK}
      19)*  weights = 2433

有人可以解释标准和统计数据的含义。我理解的权重是落在一个节点上的样本数。

条件推理树意义上的统计值

0 个答案: