The Automatic Statistician: A Relational Perspective
This work addresses a specific challenge in time-series modeling for domains like finance and economics, offering an incremental improvement over existing methods.
The authors tackled the problem of learning less informative composite covariance kernels from single time-series datasets in Gaussian Processes by proposing two relational kernel learning methods that model multiple datasets to find shared causes of changes. They demonstrated that these methods achieve more accurate regression models on real-world datasets, including US stock data, US house price index data, and currency exchange rate data.
Gaussian Processes (GPs) provide a general and analytically tractable way of modeling complex time-varying, nonparametric functions. The Automatic Bayesian Covariance Discovery (ABCD) system constructs natural-language description of time-series data by treating unknown time-series data nonparametrically using GP with a composite covariance kernel function. Unfortunately, learning a composite covariance kernel with a single time-series data set often results in less informative kernel that may not give qualitative, distinctive descriptions of data. We address this challenge by proposing two relational kernel learning methods which can model multiple time-series data sets by finding common, shared causes of changes. We show that the relational kernel learning methods find more accurate models for regression problems on several real-world data sets; US stock data, US house price index data and currency exchange rate data.