You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In K-Gaussian models, P is a matrix N \times K, where N is the size of the input r-axis. The columns are individual Gaussians normalized to the unit area. After the current normalization step, the integral of each Gaussian is 1/K. This is at least inconsistent with the model description in the documentation. E.g. from https://jeschkelab.github.io/DeerLab/_autosummary/deerlab.dd_gauss2.html#deerlab.dd_gauss2 I conclude that the condition (a1 + a2 = 1) yields a normalized distribution which is not true in reality.
in dd_model.py, in the Section dd_gauss3, the variable notes is misspelled as ntoes. Because of this, the documentation on dd_gauss3 displays a LaTeX formula of dd_gauss2 (as it precedes dd_gauss3 in dd_models.py).
HKaras
added
bugfix
Patches something that isn't working
bug
Something isn't working
and removed
bugfix
Patches something that isn't working
labels
Jul 15, 2024
Dear all,
The function
dd_models._multigaussfun
has a misleading normalization step, in my opinionIn K-Gaussian models,
P
is a matrix N \times K, where N is the size of the inputr
-axis. The columns are individual Gaussians normalized to the unit area. After the current normalization step, the integral of each Gaussian is 1/K. This is at least inconsistent with the model description in the documentation. E.g. from https://jeschkelab.github.io/DeerLab/_autosummary/deerlab.dd_gauss2.html#deerlab.dd_gauss2 I conclude that the condition(a1 + a2 = 1)
yields a normalized distribution which is not true in reality.The code below tests the normalization
Thanks!
The text was updated successfully, but these errors were encountered: