Closed segasai closed 4 years ago
Let me also push one more commit before merging, which ensures that nan's are returned when input is outside 0,1 range
I think I've concluded my push request now. (edited). I've added a couple of other commits with tests. One compares means/stddevs with integrals of pdf, and another compares the PDF integrals with the CDF values
Hi,
I've implemented the PPF (percentile point function; inverse of CDF) for all the distributions. They are useful for example for sampling (i.e. nested sampling where you need to convert your prior into uniform from zero to one). I've also added tests that they work as inverse as CDFs.
Sergey