Can anyone explain how to apply geometric distribution on a real dataset? I know there's a formula but I need to estimate the prob of success first. My problem is I want to build a model using such concept to predict the likelihood of attempt x to have the first response of the whole n attempts. To be specific, say what's the probability that I get an intial response after sending the xth message? Two factors are unknowns here - prob of success & which of the n attempts would lead me to the first success (getting a response). Or if you have any other mathematical approach you could provide, I'm very much willing to listen esp if it can be addressed by any machine learning algo using python.
My theory on estimating the prob of success is by getting the
example would be to focus on the 3rd attempt (thought this should be unknown as well)
(number of obs getting responses (success) on attempt 3)/total number of attempts regardless of success in attempt 3.