In sbalci/histopathology-template: Template of R Codes Used in Histopathology Research.

impute

Codes for missing data and impute.^[See childRmd/_10impute.Rmd file for other codes]

• Multiple imputation support in Finalfit
  https://www.datasurg.net/2019/09/25/multiple-imputation-support-in-finalfit/

• Missing data
  https://finalfit.org/articles/missing.html

Missing Data

Plot missing data

visdat::vis_miss(mydata)

visdat::vis_miss(airquality,
                 cluster = TRUE)

visdat::vis_miss(airquality,
         sort_miss = TRUE)

impute continious

# https://cran.r-project.org/web/packages/dlookr/vignettes/transformation.html

income <- dlookr::imputate_na(carseats, Income, US, method = "rpart")
income
attr(income,"var_type")
attr(income,"method")
attr(income,"na_pos")
attr(income,"type")
attr(income,"message")
attr(income,"success")
attr(income,"class")

summary(income)

plot(income)

carseats %>%
  mutate(Income_imp = dlookr::imputate_na(carseats, Income, US, method = "knn")) %>%
  group_by(US) %>%
  summarise(orig = mean(Income, na.rm = TRUE),
    imputation = mean(Income_imp))

impute categorical

library(mice)
urban <- dlookr::imputate_na(carseats, Urban, US, method = "mice")
urban 
summary(urban)
plot(urban)

impute outlier

price <- dlookr::imputate_outlier(carseats, Price, method = "capping")
price
summary(price)
plot(price)

carseats %>%
  mutate(Price_imp = dlookr::imputate_outlier(carseats, Price, method = "capping")) %>%
  group_by(US) %>%
  summarise(orig = mean(Price, na.rm = TRUE),
    imputation = mean(Price_imp, na.rm = TRUE))

transform

min -max

carseats %>% 
  mutate(Income_minmax = dlookr::transform(carseats$Income, method = "minmax"),
    Sales_minmax = dlookr::transform(carseats$Sales, method = "minmax")) %>% 
  select(Income_minmax, Sales_minmax) %>% 
  boxplot()

skewness

dlookr::find_skewness(carseats)

dlookr::find_skewness(carseats, index = FALSE)

dlookr::find_skewness(carseats, value = TRUE)

dlookr::find_skewness(carseats, value = TRUE, thres = 0.1)

log

Advertising_log = transform(carseats$Advertising, method = "log")
# Advertising_log <- transform(carseats$Advertising, method = "log+1")
head(Advertising_log)
summary(Advertising_log)
plot(Advertising_log)

binning

bin <- dlookr::binning(carseats$Income)
bin <- binning(carseats$Income, nbins = 4,
              labels = c("LQ1", "UQ1", "LQ3", "UQ3"))
binning(carseats$Income, nbins = 5, type = "equal")
binning(carseats$Income, nbins = 5, type = "pretty")
binning(carseats$Income, nbins = 5, type = "kmeans")
binning(carseats$Income, nbins = 5, type = "bclust")

bin
summary(bin)
plot(bin)



carseats %>%
 mutate(Income_bin = dlookr::binning(carseats$Income)) %>%
 group_by(ShelveLoc, Income_bin) %>%
 summarise(freq = n()) %>%
 arrange(desc(freq)) %>%
 head(10)

optimal binning

bin <- dlookr::binning_by(carseats, "US", "Advertising")
bin
summary(bin)
attr(bin, "iv") # information value 
attr(bin, "ivtable") # information value table

plot(bin, sub = "bins of Advertising variable")

standardize

# https://cran.r-project.org/web/packages/exploreR/vignettes/exploreR.html

(regressResults <- exploreR::masslm(iris,
                                   "Sepal.Length",
                                   ignore = "Species")
)

exploreR::massregplot(iris, "Sepal.Length", ignore = "Species")

(stand.Petals <- exploreR::standardize(iris,
                                      c("Petal.Width", "Petal.Length"))
)

data transformation report

carseats %>%
  dlookr::transformation_report(target = US)

carseats %>%
  dlookr::transformation_report(target = US, output_format = "html", 
    output_file = "transformation.html")

inspectdf

inspectdf::inspect_na(starwars)

inspectdf::inspect_na(starwars) %>% inspectdf::show_plot()

inspectdf::inspect_na(star_1, star_2)

inspectdf::inspect_na(star_1, star_2) %>% inspectdf::show_plot()

sbalci/histopathology-template documentation built on Nov. 14, 2024, 2:09 a.m.