This repository contains an implementation of the Compositional Kernel Search algorithm for automatic structure discovery in time series and regression data using Gaussian Processes.
This is an implementation of the algorithms described in:
"Structure Discovery in Nonparametric Regression through Compositional Kernel Search"
David Duvenaud, James Robert Lloyd, Roger Grosse, Joshua B. Tenenbaum, Zoubin Ghahramani
Proceedings of the 30th International Conference on Machine Learning (ICML), 2013
Paper: https://arxiv.org/abs/1302.4922
compositional_kernel_search.py- Main implementation of the CKS algorithmexamples.py- Example usage and demonstrations
