Reproduce “The Random Forest Kernel and other kernels for big data from random partitions” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Alex Davies, Zoubin Ghahramani
The authors introduce a kernel construction that turns randomized partitions into similarities, measuring how often two points share a cluster. They demonstrate it with Random Forest and Fast Cluster kernels. In Gaussian-process regression on six UCI datasets using 200 partitions, both outperform standard kernels in test log-likelihood; mean-squared error improves less consistently. This does not establish superiority over Random Forest prediction.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.