Get rid of logy, logZ transformation

harrispopgen / mushi

[mu]tation [s]pectrum [h]istory [i]nference

https://harrispopgen.github.io/mushi/

MIT License

24 stars 6 forks source link

Get rid of logy, logZ transformation #34

Closed kamdh closed 4 years ago

wsdewitt commented 5 years ago

Yeah we're currently fitting logy instead of y. Fitting on y didn't work very well because 1. effective population size varies over orders of magnitude, and the spline regularizers seemed not to like this, 2. the integrand contains factors like exp(1 / y), which blows up at zero.

kamdh commented 4 years ago

I think we can close this issue for now on logy. We aren't using logZ as far as I know.