R：在回归方程中添加求和

Question

我试图在 R 中编写这个公式，其中 i = 类别的每个值（类别可以是 1 2 3 或 4）

这是我的代码尝试，但 R 打印此错误消息：
lm(类别 ~ (年份 * 州 * 区) + 年 + 州 + 区 + 中的错误： 形式参数“数据”与多个实际参数匹配

我正在尝试创建求和，因此我必须在数据后添加多个参数，是否有另一种方法来编写求和以避免错误消息？ 我在网上查了一下，但找不到类似的东西，我猜很少会在回归中添加求和。 预先感谢您的帮助

ID <-  c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,
         17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,
         33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48)
year <- c(1980,1980,1980,1980,1980,1980,1980,1980,1980,1980,1980,1980,1980,1980,1980,1980,
      1981,1981,1981,1981,1981,1981,1981,1981,1981,1981,1981,1981,1981,1981,1981,1981,
      1982,1982,1982,1982,1982,1982,1982,1982,1982,1982,1982,1982,1982,1982,1982,1982)
 state <-     c("NY","NY","NY","NY","NY","NY","NY","NY","CA","CA","CA","CA","CA","CA","CA","CA",
      "NY","NY","NY","NY","NY","NY","NY","NY","CA","CA","CA","CA","CA","CA","CA","CA",
      "NY","NY","NY","NY","NY","NY","NY","NY","CA","CA","CA","CA","CA","CA","CA","CA")
district <- c(1,1,1,1,2,2,2,2,1,1,1,1,2,2,2,2,
          1,1,1,1,2,2,2,2,1,1,1,1,2,2,2,2,
          1,1,1,1,2,2,2,2,1,1,1,1,2,2,2,2)
quantity <- c(100,200,45,87,65,32,94,52,67,72,14,53,28,94,12,41,
          10,20,45,87,65,32,8,52,67,1,14,53,28,94,12,41,
          1000,2000,45,87,9,32,94,5,6,7,1,5,2,9,1,4)
category <- c(1,2,3,4,1,2,3,4,1,2,3,4,1,2,3,4,
   1,2,3,4,1,2,3,4,1,2,3,4,1,2,3,4,
   1,2,3,4,1,2,3,4,1,2,3,4,1,2,3,4)

df <- data.frame(ID,year,state,district,quantity,category)

df$year <- as.factor(df$year)
df$state <- as.factor(df$state)
df$district <- as.factor(df$district)
df$category <- as.factor(df$category)

print(df)

# force regression baseline values 
relevel(df$year, ref = '1981')
relevel(df$district, ref = '2')

# r1 is when y = 1
r1 <- lm( category ~ (year*state*district) +  
           quantity + district + state + year,
           data = subset(df, year == 1980)
      +
        (year*state*district) +  
         quantity + district + state + year,
         data = subset(df, year == 1981)
      +
        (year*state*district) +  
         quantity + district + state + year,
         data = subset(df, year == 1980)
     )
summary(r1)

# r2 is when y = 2
r2 <- lm( category ~ (year*state*district) +  
        year + state + district + quantity,
      data = subset(df, year == 1980)
      +
        (year*state*district) +  
        year + state + district + quantity,
      data = subset(df, year == 1981)
      +
        (year*state*district) +  
        year + state + district + quantity,
      data = subset(df, year == 1980)
)
summary(r2)

then r3 and r4

Answer 1

这里有几个问题：

• 开头附近求和的行需要将系数与每一项相乘

• year[i, y] 是什么意思？ 没有定义。

• 线性回归不适用于分类响应。 假设类别是无序的，我们可以使用多项逻辑回归。

• 交互通常要求也包括所有低阶交互。

也许你想要这个：

library(nnet)

fm <- multinom(category ~ year/(district * state) + district + state + quantity, df)
summary(fm)

fm

属于“multinom”类对象，具有以下方法：

methods(class = "multinom")
## [1] add1        anova       coef        confint     drop1       extractAIC 
##  [7] logLik      model.frame predict     print       summary     vcov       
## see '?methods' for accessing help and source code

有关解释，请参阅https://stats.oarc.ucla.edu/r/dae/multinomial-logistic-regression/