Hello,

I am trying to replicate Figure 3 from this paper here that uses SG++, but cannot find the code steps and procedure to achieve such figure. This paper is where I was first introduced to SG++ and want to implement this for my work.

https://cims.nyu.edu/~pehersto/preprints/sgde_siam.pdf

@xantares @elfeck @jschueller @stoiver @MichaelReh

I would like the sparse grid coordinates and their associated densities. That way I can understand how to use it in my dataset. I have data of size 2638 samples by 10 features. If I create a meshgrid by splicing each dimension into 10 equal width binning, then the problem becomes impossible to do standard kernel density estimation on (10 ^10 total hypercube bins). But with the SG++ sparse grids, I am hoping to reduce that and only obtain densities on relevant grids and their coordinates. For example, grid position [10,5,2,3,2,1,1,2,1,5] = 0.66245 , etc ...

Any help or guidance would be greatly apprecieted.