As is always the way I stumbled across Tsallis entropy on SO whilst looking for something completely different. This soon lead me reading all sorts of interesting but terse academic papers.
I am unfortunately a mere layman and I still have one big unsolved question.
The key input to Tsallis entropy is a probability array.
What I don't understand is how do you get it out of a time-series ?
Allow me to give you a completely hypothetical example:
- I have a database with 10 year's worth of non-stationary data e.g. house prices,wind speeds, vehicle speeds etc.
- I want to explore it on various time frames (e.g. daily, monthly etc.)
- So conceptually I run through the 10 years of data with rolling-window of chosen delta to generate features, one of which will be a Tsallis value.
- What happens about feeding into Tsallis ? The individual numbers won't be of use ? log-diffs won't be of much use ? Where does this mythical array of probabilities come from ?