# Event similarity based on probability distance measure

Hеllо еvеrybоdy!

It became interesting to me to code the following case: I have some spatio temporal events, and they have uncertain boundaries (for example "John visited England in April" and "Harry was in London on 1 May"). Here we have uncertain location (England and London), so we dont know exacltly at which certain point of England or London it occured, and time is uncertain as well. And the problem is to figure out What the level of probability that given events with uncertain boundries are similar? (kind of correlation or clustering events)

I made a literature research and found some algorithms, for example to use ptobability distribution function, or calculte probability and compare it with predifined epsilon value.

Dear friends who a pro in the field of fuzzy sets and probability theory please just give me an idea or links how to deal with that problem.

If you don't have a closed form solution (try computing overlap of polygons and time - this should be fairly easy to do) you can always do sampling. But sampling from polygons is probably harder than the analytical approach...

The idea of sampling is as follows: you draw a random sample satisfying the conditions of A and then check if it also satisfies the conditions of B. Repeat this a million times and you get a naive estimate of similarity.

You can do the same with distances: draw a sample from each, compute their distances, repeat until you get an average of the desired precision.