Timezone: »

 
Probabilistic structure discovery in time series data
David Janz · Brooks Paige · Thomas Rainforth · Jan-Willem van de Meent

Sat Dec 10 02:40 AM -- 03:00 AM (PST) @
in Towards an Artificial Intelligence for Data Science »

Existing methods for structure discovery in time series data construct interpretable, compositional kernels for Gaussian process regression models. While the learned Gaussian process model provides posterior mean and variance estimates, typically the structure is learned via a greedy optimization procedure. This restricts the space of possible solutions and leads to over-confident uncertainty estimates. We introduce a fully Bayesian approach, inferring a full posterior over structures, which more reliably captures the uncertainty of the model.

Author Information

David Janz (University of Cambridge)
Brooks Paige (Alan Turing Institute / University of Cambridge)
Thomas Rainforth (University of Oxford)
Jan-Willem van de Meent (University of Amsterdam and Northeastern University)

More from the Same Authors