Question

I am trying to reproduce Naive Bayes nomogram as given in Nomograms for Visualization of Naive Bayesian Classifier by Mozina. It is a great visualization for looking at Bayes probabilities. I have been searching and trying various things, but no luck. (I am unable to put all the points on one row for a column.) I’ve computed probabilities and put them in a data frame called df

structure(list(.id = c("outlook", "outlook", "outlook", "windy", 
"windy"), variablevalue = structure(c(1L, 2L, 3L, 5L, 6L), .Label = c("sunny", 
"overcast", "rainy", "'All'", "FALSE", "TRUE"), class = "factor"), 
prob = c(0.222222222222222, 0.444444444444444, 0.333333333333333, 
0.666666666666667, 0.333333333333333)), .Names = c(".id", 
"variablevalue", "prob"), row.names = c(1L, 3L, 5L, 11L, 13L), class = "data.frame")

Here’s how the chart would like (this chart is all cut and paste):

Answer 1

Does this work?

ggplot(df, aes(prob,.id,label=variablevalue)) + 
  geom_text() + 
  xlim(c(0,1))