我有一个 .csv 文件，我将其导入 R 以进行一些计算。令我困惑的是我收到的错误消息如下：
Warning Message:
In mean.default(lace$TOTAL.LACE) :
argument is not numeric or logical: returning NA
我觉得这很奇怪的原因是因为：
> mode(lace$TOTAL.LACE)
[1] "numeric"
所以我完全不知道为什么会这样。
我正在导入的数据如下所示：
ENCOUNTER | MRN | AGE | NAME | LACE.DAYS.SCORE | LACE.ACUTE.IP.SCORE | LACE.ER.SCORE | LACE.COMORBID.SCORE | TOTAL.LACE | FAILURE | LOS | DAYS.TO.FAILURE
123 | 123 | 50 | DOE | 6 | 3 | 0 | 1 | 10 | 0 | 0 | 0
LOS 列是停留时间，只有在 Failure 为 1 时才为非零。Failure 为二进制 1 或 0。Days.To.Failure 也仅在 Failure 为 1 时为非零。
这是我的全部代码：
# Load the requried libraries
library(aod)
library(ggplot2)
library(knitr)
lace <- read.csv("lace for R.csv")
head(lace)
summary(lace)
# This dataset has a binary response called FAILURE. There are four
# predictors of this, Length of Stay, Acute Admission, Comorbidity Score,
# ER visits in the last 6 months. There is also a total.lace coloumn that
# is simply the sum of the four scores
sapply(lace, sd)
# We will no make a two-way contingency table of categorical outcome
# and predictors we want to make sure there are not 0 cells
xtabs(~FAILURE + TOTAL.LACE, data = lace)
# We now need to convert TOTAL.LACE to a factor to indicate that
# it should be treated as a categorical variable
lace$TOTAL.LACE <- factor(lace$TOTAL.LACE) <-- pos. problem but mode() says numeric
mylogit <- glm(FAILURE ~ TOTAL.LACE, data = lace, family = "binomial")
summary(mylogit)
# We will here use confint to get the confidence intervals of our data
confint(mylogit)
# if there are warnings this will show them
warnings()
# Default confint.default
confint.default(mylogit)
# Wald Test of the TOTAL.LACE (rank)
wald.test(b = coef(mylogit), Sigma = vcov(mylogit), Terms = 2:15)
# Odds Ratios only
exp(coef(mylogit))
# odds ratios and 95% CI
exp(cbind("Odds Ratio" = coef(mylogit), confint(mylogit)))
更新
我被要求运行 dput(lace$TOTAL.LACE) 这是示例输出：
8L, 7L, 5L, 7L, 6L, 6L, 6L, 6L, 4L, 7L, 4L, 5L, 7L, 7L, 6L, 3L,
...
.Label = c("3", "4", "5", "6", "7", "8", "9", "10",
"11", "12", "13", "14", "15", "16", "17", "18"), class = "factor")
还要求 str(lace)，这里是示例输出：
'data.frame': 17044 obs. of 12 variables:
$ ENCOUNTER : int ...
$ MRN : int ...
$ AGE : int 74 87 74 57 52 60 42 84 75 79 ...
$ NAME : Factor w/ 11637 levels "doe",..: 3945...
$ LACE.DAYS.SCORE : int 6 6 6 6 6 6 6 6 6 6 ...
$ LACE.ACUTE.IP.SCORE: int 3 3 3 3 3 3 3 3 3 3 ...
$ LACE.ER.SCORE : int 0 0 0 0 0 0 0 0 0 0 ...
$ LACE.COMORBID.SCORE: int 1 0 2 0 0 0 0 0 2 1 ...
$ TOTAL.LACE : Factor w/ 16 levels "3","4","5","6",..: 8 7 9 7 7 7 7 ...
$ FAILURE : int 0 0 0 0 0 0 0 0 0 0 ...
$ LOS : int 0 0 0 0 0 0 0 0 0 0 ...
$ DAYS.TO.FAILURE : int 0 0 0 0 0 0 0 0 0 0 ...
谢谢你，