Probability algorithm: Finding probable correct item in a list (e.g John, John, Jon)
- by Andrew White
Hi,
Take for example the list (L):
John, John, John, John, Jon
We are to presume one item is to be correct (e.g. John in this case), and give a probability it is correct.
First (and good!) attempt: MostFrequentItem(L).Count / L.Count (e.g. 4/5 or 80% likelihood)
But consider the cases:
John, John, Jon, Jonny
John, John, Jon, Jon
I want to consider the likelihood of the correct item being John to be higher in the first list! I know I have to count the SecondMostFrequent Item and compare them.
Any ideas? This is really busting my brain!
Thx,
Andrew