无法使用 R 对数据进行分类

Question

我尝试对我的数据进行分类，以便制作逻辑回归模型。我对 R 很陌生，为了我的学习而学习它。我已经使用了我在多个示例中看到的这段代码，但没有任何东西经过并保持不变。也没有报错。

ds <- read.csv("adult.csv")
colnames(ds)<- c("age","workclass","responsenum","education","education_years","marital_status","occupation","familyrole", "race","sex", "capital_gain", "capital_loss", "hours_per_week","country", "income") 

ds$workclass <- as.character(ds$workclass)
ds$workclass[ds$workclass == "Without-pay" | ds$workclass == "Never-worked"] <- "Jobless"
ds$workclass[ds$workclass == "State-gov" | ds$workclass == "Local-gov"]  <- "govt" 
ds$workclass[ds$workclass == "Self-emp-inc" | ds$workclass == "Self-emp-not-inc"]  <- "Self-employed"

当我之后使用 table() 时，我仍然从中提取旧名称。

有人知道出了什么问题吗？

dput(head(ds)) 的输出=

structure(list(age = c(50L, 38L, 53L, 28L, 37L, 49L), workclass = c(" Self-emp-not-inc", 
" Private", " Private", " Private", " Private", " Private"), 
    responsenum = c(83311L, 215646L, 234721L, 338409L, 284582L, 
    160187L), education = c(" Bachelors", " HS-grad", " 11th", 
    " Bachelors", " Masters", " 9th"), education_years = c(13L, 
    9L, 7L, 13L, 14L, 5L), marital_status = c(" Married-civ-spouse", 
    " Divorced", " Married-civ-spouse", " Married-civ-spouse", 
    " Married-civ-spouse", " Married-spouse-absent"), occupation = c(" Exec-managerial", 
    " Handlers-cleaners", " Handlers-cleaners", " Prof-specialty", 
    " Exec-managerial", " Other-service"), familyrole = c(" Husband", 
    " Not-in-family", " Husband", " Wife", " Wife", " Not-in-family"
    ), race = c(" White", " White", " Black", " Black", " White", 
    " Black"), sex = c(" Male", " Male", " Male", " Female", 
    " Female", " Female"), capital_gain = c(0L, 0L, 0L, 0L, 0L, 
    0L), capital_loss = c(0L, 0L, 0L, 0L, 0L, 0L), hours_per_week = c(13L, 
    40L, 40L, 40L, 40L, 16L), country = c(" United-States", " United-States", 
    " United-States", " Cuba", " United-States", " Jamaica"), 
    income = c(" <=50K", " <=50K", " <=50K", " <=50K", " <=50K", 
    " <=50K")), row.names = c(NA, 6L), class = "data.frame")

Answer 1

您的数据有前导空格，因此 " Self-imp-not-inc" 永远不会匹配 "Self-emp-not-inc"。

想法：

您可以 trim leading/trailing 所有类似字符串的列中的空格。

str(ds, list.len = 4)
# 'data.frame': 6 obs. of  15 variables:
#  $ X39          : int  50 38 53 28 37 49
#  $ State.gov    : chr  " Self-emp-not-inc" " Private" " Private" " Private" ...
#  $ X77516       : int  83311 215646 234721 338409 284582 160187
#  $ Bachelors    : chr  " Bachelors" " HS-grad" " 11th" " Bachelors" ...
#   [list output truncated]
ischr <- sapply(ds, is.character)
ischr
#           X39     State.gov        X77516     Bachelors           X13 Never.married  Adm.clerical Not.in.family 
#         FALSE          TRUE         FALSE          TRUE         FALSE          TRUE          TRUE          TRUE 
#         White          Male         X2174            X0           X40 United.States        X..50K 
#          TRUE          TRUE         FALSE         FALSE         FALSE          TRUE          TRUE 
ds[ischr] <- lapply(ds[ischr], trimws)
str(ds, list.len = 4)
# 'data.frame': 6 obs. of  15 variables:
#  $ X39          : int  50 38 53 28 37 49
#  $ State.gov    : chr  "Self-emp-not-inc" "Private" "Private" "Private" ...
#  $ X77516       : int  83311 215646 234721 338409 284582 160187
#  $ Bachelors    : chr  "Bachelors" "HS-grad" "11th" "Bachelors" ...
#   [list output truncated]

或者，您可以在所有模式中添加空格，例如：

ds$workclass[ds$workclass == " Without-pay" | ds$workclass == " Never-worked"] <- " Jobless"
ds$workclass[ds$workclass == " State-gov" | ds$workclass == " Local-gov"]  <- " govt" 
ds$workclass[ds$workclass == " Self-emp-inc" | ds$workclass == " Self-emp-not-inc"]  <- " Self-employed"

（我认为这不是最好的方法，所以我假设您选择使用 trimws 作为第一个项目符号。）

您可以使用 %in%:

来简化您的一些 |

    ds$workclass[ds$workclass %in% c("Without-pay", "Never-worked")] <- "Jobless"
    ds$workclass[ds$workclass %in% c("State-gov", "Local-gov")]  <- "govt" 
    ds$workclass[ds$workclass %in% c("Self-emp-inc", "Self-emp-not-inc")]  <- "Self-employed"

您还可以创建各种字典，将某些内容翻译成其他内容。例如，

translations <- read.csv(header = TRUE, text = "
src,tgt
Without-pay,Jobless
Never-worked,Jobless
State-gov,govt
Local-gov,govt
Self-emp-inc,Self-employed
Self-emp-not-inc,Self-employed")

ds$State.gov
# [1] "Self-emp-not-inc" "Private"          "Private"          "Private"          "Private"          "Private"         
ifelse(ds$State.gov %in% translations$src, translations$tgt[ match(ds$State.gov, translations$src) ], ds$State.gov)
# [1] "Self-employed" "Private"       "Private"       "Private"       "Private"       "Private"      
ds$State.gov <- ifelse(ds$State.gov %in% translations$src,
                       translations$tgt[ match(ds$State.gov, translations$src) ],
                       ds$State.gov)

（此技术也可以作为 merge 或 dplyr::*_join 操作来实现，但目前我认为这可能比您需要的更复杂。）

使用这种类似字典的翻译的主要优点（在我看来）是它最容易查看、理解和维护。

无法使用 R 对数据进行分类

Can't categorize data using R

model

r