Enhancing Food Price Forecasts in Canada: An Integration ofExpert-Driven Covariates and Advanced ML Approaches
154
by Kristina L. Kupferschmidt and Cody Kupferschmidt
Published onMay 27, 2024
Enhancing Food Price Forecasts in Canada: An Integration ofExpert-Driven Covariates and Advanced ML Approaches
·
Abstract
This study addresses the growing issue of food affordability in Canada, exacerbated by recent inflation and other global factors. Canada's Food Price Report (CPFR) is an annual publication that predicts food inflation over the next calendar year. While in recent years the CFPR has leveraged machine-learning (ML), the 2024 report also included a human-in-the-loop approach. This approach included expert-driven economic and climate variables as additional model inputs, with results suggesting that these variables improved forecast accuracy for several food categories. In the present study, we investigate sensitivity of models used for the CFPR report to specific combinations of these covariates. Our preliminary findings suggest potential synergistic effects of combining various covariates, resulting in more accurate forecasts that continue to perform well in changing global conditions.
Creative Commons Attribution 4.0 International License (CC-BY 4.0)