From @bob-carpenter on January 12, 2015 19:20
Jan Scholz (via e-mail) suggested implementing the general form of interpolation used in SciPy.
https://github.com/scipy/scipy/blob/v0.15.0/scipy/ndimage/interpolation.py#L230
It should be able to handle 2D and 3D structures. Jan says the linear form (order 1) should be enough for his purposes.
For background, BUGS provides an interp.lin function for the 1D case:
http://www.mrc-bsu.cam.ac.uk/wp-content/uploads/manual14.pdf
The goal would be to get the derivatives working properly in a way that encapsulates the index-fiddling relative to the user and takes care of error conditions.
This would be a nice standalone project.
Copied from original issue: stan-dev/stan#1214
From @bob-carpenter on January 12, 2015 19:20
Jan Scholz (via e-mail) suggested implementing the general form of interpolation used in SciPy.
https://github.com/scipy/scipy/blob/v0.15.0/scipy/ndimage/interpolation.py#L230
It should be able to handle 2D and 3D structures. Jan says the linear form (order 1) should be enough for his purposes.
For background, BUGS provides an interp.lin function for the 1D case:
http://www.mrc-bsu.cam.ac.uk/wp-content/uploads/manual14.pdf
The goal would be to get the derivatives working properly in a way that encapsulates the index-fiddling relative to the user and takes care of error conditions.
This would be a nice standalone project.
Copied from original issue: stan-dev/stan#1214