Closed calvinmccarter closed 11 months ago
Hi Calvin,
Our method does not perform density estimation. It could be possible to implement it following the approach used in the zero-shot classification, but it would be computationally intensive (n_t=10, n_z=10, 250 values = 250 x 10 x 10 xgboost predictions) and I'm not sure what would be the advantage over using standard kernel-based density estimation.
Alexia
Hi Alexia, thanks -- makes sense!
This method worked really well on my data! Is it possible to perform density estimation with ForestDiffusion?