Trading Data for Wind Power Forecasting: A Regression Market with Lasso Regularization

TL;DR

This paper introduces a regression market for wind power forecasting that uses Lasso regularization to enable data sellers and buyers to trade data efficiently, improving forecast accuracy and financial outcomes.

Contribution

It proposes a novel regression market framework utilizing Lasso to individualize data payments and enhance feature selection for wind power forecasting.

Findings

Reduced overall losses for data buyers.

Increased financial benefits for data sellers.

Effective feature selection with Lasso in real-world data.

Abstract

This paper proposes a regression market for wind agents to monetize data traded among themselves for wind power forecasting. Existing literature on data markets often treats data disclosure as a binary choice or modulates the data quality based on the mismatch between the offer and bid prices. As a result, the market disadvantages either the data sellers due to the overestimation of their willingness to disclose data, or the data buyers due to the lack of useful data being provided. Our proposed regression market determines the data payment based on the least absolute shrinkage and selection operator (lasso), which not only provides the data buyer with a means for selecting useful features, but also enables each data seller to individualize the threshold for data payment. Using both synthetic data and real-world wind data, the case studies demonstrate a reduction in the overall losses for wind agents who buy data, as well as additional financial benefits to those who sell data.