R: Rearranging code / Loops for Confusion Matrix

Question

I am basically trying to work a confusion matrix with some prediction / response variables under a couple of conditions, but im not sure how to lay out the code.

for my prediction values, my current code is:

Prediction <- factor (rep(data$percentage <=40.0 || data$binary ==1,levels=c("TRUE","FALSE")))

but I am pretty sure this is incorrect because what I want to achieve is the following conditions:

• if percentage is <=40 then consider this as TRUE
• but if percentage is shown as NA then refer to data$binary column
• based on this, if data$binary = 1, then consider this as TRUE.

Any help would be much appreciated !

Answer 1

not to hard code, I worked directly on your data frame;

library(dplyr)

data %>%
mutate(logistic_column=rowSums(data)) %>% 
mutate(prediction=ifelse(is.na(logistic_column),
                         ifelse(binary==1,TRUE,FALSE),
                         ifelse(percentage<=40,TRUE,FALSE))) %>% 
select(prediction) %>% 
pull -> prediction