Hello,
I have a time series dataset of approximately 3000 observations, which are split fairly even 52 % to 50 % of 0's to 1's.
My general goal is to find what is the probability of the next outcome in the time series to be 0 or 1.
First, I thought that if I find the maximum number of consecutive sequences of 1's and 0's I could determine the probability of the change in the next outcome. For example, if my maximum consecutive sequence of 1's is 10, so 1/10 is 0.10 %. SO for every next 1, I have in my time series, the probability of reverse (to get 0 next) increases by 10 %. But when I thought about it I found a couple of problems I am not sure how to solve.
1 - The increase of the probability should not be linear, judging from the nature of the data. I am not sure how to determine the power of exponential growth I need to use for every next consecutive outcome.
2 - Not sure how to determine the frequencies and their probabilities. What I mean by that is, that a sequence of frequent change (010110100101) is far more likely from the probability of having long consecutive sequences (11111111110110). not sure how to quantitively measure the probabilities.
Namely, I am interested if someone can help with the statistics, but any python code will be greatly appreciated as well.
I have a time series dataset of approximately 3000 observations, which are split fairly even 52 % to 50 % of 0's to 1's.
My general goal is to find what is the probability of the next outcome in the time series to be 0 or 1.
First, I thought that if I find the maximum number of consecutive sequences of 1's and 0's I could determine the probability of the change in the next outcome. For example, if my maximum consecutive sequence of 1's is 10, so 1/10 is 0.10 %. SO for every next 1, I have in my time series, the probability of reverse (to get 0 next) increases by 10 %. But when I thought about it I found a couple of problems I am not sure how to solve.
1 - The increase of the probability should not be linear, judging from the nature of the data. I am not sure how to determine the power of exponential growth I need to use for every next consecutive outcome.
2 - Not sure how to determine the frequencies and their probabilities. What I mean by that is, that a sequence of frequent change (010110100101) is far more likely from the probability of having long consecutive sequences (11111111110110). not sure how to quantitively measure the probabilities.
Namely, I am interested if someone can help with the statistics, but any python code will be greatly appreciated as well.
