I'm sorry if this is already answered! I don't know the terminology.
If I have a sequence of categorical observations:
{a: 100, b: 20, c: 30}
I can calculate percentage for each category as category / sum(all categories). This would give me:
{a: 0.666.., b: 0.133.., c: 0.2}
What I am unclear on is what the calculation sum(all categories) / category would be. Which would produce:
{a: 1.5, b: 7.5, c: 5}
What do you call this? I'd think 'inverse frequency', or something like that. Searching for that wasn't helpful.