Regression markets and application to energy forecasting

TL;DR

This paper introduces regression markets where agents can share data for linear regression models, incentivizing collaboration and improving energy forecasting accuracy through a novel market-based framework.

Contribution

It proposes a new framework called regression markets for distributed data sharing in regression tasks, integrating interpretability and game theory, with theoretical analysis and practical case studies.

Findings

Market design ensures desirable properties for data sharing and incentives.

Simulation studies demonstrate the effectiveness of the proposed approach.

Real-world case studies validate the practical applicability.

Abstract

Energy forecasting has attracted enormous attention over the last few decades, with novel proposals related to the use of heterogeneous data sources, probabilistic forecasting, online learn-ing, etc. A key aspect that emerged is that learning and forecasting may highly benefit from distributed data, though not only in the geographical sense. That is, various agents collect and own data that may be useful to others. In contrast to recent proposals that look into distributed and privacy-preserving learning (incentive-free), we explore here a framework called regression markets. There, agents aiming to improve their forecasts post a regression task, for which other agents may contribute by sharing their data for their features and get monetarily rewarded for it.The market design is for regression models that are linear in their parameters, and possibly sep-arable, with estimation performed based on either batch or online learning. Both in-sample and out-of-sample aspects are considered, with markets for fitting models in-sample, and then for improving genuine forecasts out-of-sample. Such regression markets rely on recent concepts within interpretability of machine learning approaches and cooperative game theory, with Shapley additive explanations. Besides introducing the market design and proving its desirable properties, application results are shown based on simulation studies (to highlight the salient features of the proposal) and with real-world case studies.