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
For example: Within fitted.FPCA, Hmisc::approxExtrap(...) can probably be dropped in favour of stats::spline( method = "natural", ... ). The performance deterioration will probably be minimal but we will be able to drop the depend to Hmisc which is a big dependency. Similarly dependencies on pracma and/or MASS probably could be trimmed too.
The text was updated successfully, but these errors were encountered:
I agree with dropping redundant dependencies. We may port Hmisc::approxExtrap to fdapace instead of using spline smoothing for the sake of minimizing changes. pracma may also be dropped, but I think since everyone has MASS installed it should be less of concern.
I was working on matlab function interp in R.
I can contribute that code which is written in RcppArmadillo and verified. I think that it’s what you need.
You can find interp.cpp in url.
The test file is in here.
Hello ChingChuan, thanks for being interested to contribute to fdapace. fdapace doesn't depend on RccArmadillo so there is a bit of an issue of introducing a new dependency and potentially not dropping a single one. You are welcome to fork out fdapace and integrate your code, we will be happy to review the relevant pull request.
For example: Within
fitted.FPCA
,Hmisc::approxExtrap(...)
can probably be dropped in favour ofstats::spline( method = "natural", ... )
. The performance deterioration will probably be minimal but we will be able to drop the depend toHmisc
which is a big dependency. Similarly dependencies onpracma
and/orMASS
probably could be trimmed too.The text was updated successfully, but these errors were encountered: