模型框架默认错误（Terms，newdata，na.action = na.action，xlev = object$xlevels）：因子X有新水平。

Question

我做了一次逻辑回归：

 EW <- glm(everwrk~age_p + r_maritl, data = NH11, family = "binomial")

此外，我希望能针对每个r_maritl的级别预测everwrk。 r_maritl包括以下级别：

levels(NH11$r_maritl)
 "0 Under 14 years" 
 "1 Married - spouse in household" 
 "2 Married - spouse not in household"
 "3 Married - spouse in household unknown" 
 "4 Widowed"                               
 "5 Divorced"                             
 "6 Separated"                             
 "7 Never married"                        
 "8 Living with partner"  
 "9 Unknown marital status"  

于是我就这么做了：

predEW <- with(NH11,
expand.grid(r_maritl = c( "0 Under 14 years", "1 Married - 
spouse in household", "2 Married - spouse not in household", "3 Married - 
spouse in household unknown", "4 Widowed", "5 Divorced", "6 Separated", "7 
Never married", "8 Living with partner", "9 Unknown marital status"),
age_p = mean(age_p,na.rm = TRUE)))

cbind(predEW, predict(EW, type = "response",
                        se.fit = TRUE, interval = "confidence",
                        newdata = predEW))

问题是我得到了以下回复：

``` 在 model.frame.default(Terms, newdata, na.action = na.action, xlev = object$xlevels) 中出错：factor r_maritl has new levels 0 Under 14 years, Married - spouse in household unknown ```

样本数据：

str(NH11$age_p)
num [1:33014] 47 18 79 51 43 41 21 20 33 56 ...

str(NH11$everwrk)
Factor w/ 2 levels "2 No","1 Yes": NA NA 2 NA NA NA NA NA 2 2 ...

str(NH11$r_maritl)
Factor w/ 10 levels "0 Under 14 years",..: 6 8 5 7 2 2 8 8 8 2 ...

- Hadsga

---

能否提供一些样例数据？例如 mtcars，目前我无法重现你的问题。另外，你的数据集 NH11 中是否存在未使用的因子？ - coffeinjunky

Answer 1

简而言之，您的因子中可能存在某些水平，在数据中没有被表示出来，从而使得模型中使用的因子被删除。事后看来这并不令人惊讶，因为您无法预测这些水平的响应。尽管如此，R 并没有像自动生成 NA 值一样为您提供便利。您可以通过在构建预测框架时使用 levels(droplevels(NH11$r_maritl)) 或等效的 EW$xlevels$r_maritl 来解决这个问题。

一个可复制的示例:

maritl_levels <- c( "0 Under 14 years", "1 Married - spouse in household", 
  "2 Married - spouse not in household", "3 Married - spouse in household unknown", 
  "4 Widowed", "5 Divorced", "6 Separated", "7 Never married", "8 Living with partner", 
 "9 Unknown marital status")
set.seed(101)
NH11 <- data.frame(everwrk=rbinom(1000,size=1,prob=0.5),
                 age_p=runif(1000,20,50),
                 r_maritl = sample(maritl_levels,size=1000,replace=TRUE))

让我们创造一个缺失的关卡：

NH11 <- subset(NH11,as.numeric(NH11$r_maritl) != 3)

拟合模型：

EW <- glm(everwrk~r_maritl+age_p,data=NH11,family=binomial)
predEW <- with(NH11,
  expand.grid(r_maritl=levels(r_maritl),age_p=mean(age_p,na.rm=TRUE)))
predict(EW,newdata=predEW)

成功！

Error in model.frame.default(Terms, newdata, na.action = na.action, xlev = object$xlevels) : factor r_maritl has new levels 2 Married - spouse not in household

错误：在model.frame.default（Terms，newdata，na.action = na.action，xlev = object $ xlevels）中，因子r_maritl有新的2个级别，已婚-配偶不在家庭内。

predEW <- with(NH11,
           expand.grid(r_maritl=EW$xlevels$r_maritl,age_p=mean(age_p,na.rm=TRUE)))
predict(EW,newdata=predEW)

- Ben Bolker

---

我真的非常怀疑[版本差异]。这是所有核心R功能，已经存在了几十年，并由数千个用户使用。我的先验非常强烈地认为这是OP的一个细微的拼写错误。 - Ben Bolker

酷，我今天学到了新东西。（我不认为我以前遇到过这个特殊的问题。） - Ben Bolker

那个命令不会做你想要的事情。使用 droplevels()。 - Ben Bolker

Answer 2

非常感谢您的答案，我也遇到了新层次的同样问题。 我已经在我的代码中进行了以下更改。

1. 我使用了 data.frame() 并将其替换为 expand.grid() 函数。
2. 在平均函数中，我将 na.rm=TRUE 添加到变量后面。
3. 将 factor(1:2) 替换为 glmoutput$xlevels$variablename

这个解决方案起作用了!

- Vipin Saini