I made a Monte Carlo based hypothesis test, which starts comparing N values in M simulations to a confidence interval, giving me a binary N*M matrix. Then I calculate the percent of values equal to 1 for every simulation. So the final M values I'm going to use for the hypothesis test are between 0 and 1.
However, I can't fit the values to any distributions I know using Kolmogorov-Smirnov, even Box-Cox transformation (car package in R) fails to give me a normal distribution. Below are two pictures of the non-transformed distributions to help visualize.
What are the suggested distributions for my case study? Preferably those I can fit to any single mode or bi-modal data from my procedure and not only the example distribution.
In case pictures don't load here or you like Cross Validated forum, here's the link to the same question there:
https://stats.stackexchange.com/questions/613265/need-help-finding-a-corresponding-distribution
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R