fix double calculation of L-skewness

GeoscienceAustralia / tcrm

A statistical-parametric model for assessing wind hazard from tropical cyclones

http://geoscienceaustralia.github.io/tcrm

Other

80 stars 52 forks source link

fix double calculation of L-skewness #129

Closed kieranricardo closed 2 years ago

kieranricardo commented 2 years ago

This PR fixes a bug where the L-skewness was calculated incorrectly. Previously the L-skewness was calculated to be L3 / (L2 * L2) instead of L3 / L2.

wcarthur commented 2 years ago

@kieranricardo Needs the unit tests for Utilities.evd to be updated: https://github.com/GeoscienceAustralia/tcrm/runs/5194979895?check_suite_focus=true#step:4:348

kieranricardo commented 2 years ago

@wcarthur I've changed the test to assess whether our fitting procedure can correctly reproduce a distribution from a sample generated using scipy - and with the new bug fix this is done correctly!