Feat normal-score-transform for DSI - Githubissues

pypest / pyemu

python modules for model-independent uncertainty analyses, data-worth analyses, and interfacing with PEST(++)

BSD 3-Clause "New" or "Revised" License

168 stars 94 forks source link

Feat normal-score-transform for DSI #520

Open rhugman opened 1 week ago

rhugman commented 1 week ago

Allows for users to specify normal-score transformation and/or log-transformation of observation data when calling pyemu.EnDS.prep_for_dsi().

Notes:

included python native 1D interpolation Class in helpers with the sole purpose of avoiding from scipy.interpolate import interp1d. I stress tested against scipy's interp1d. Differences start to creep in when extrapolating far from the original x-values. It may be worth using scipy instead (or someone smarter than me taking a crack at it...).

coveralls commented 1 week ago

coverage: 79.309% (-0.01%) from 79.322% when pulling 8bece65564a6d1c1f2c3da931c0fb1e33944cbf2 on rhugman:feat_normtransf into ca8cb77474655c3f1ce1e08e48128c4ebb1626b8 on pypest:develop.