Interpretable Forecasting of Energy Demand in the Residential Sector

Abstract

Energy demand forecasting is practiced in several time frames; different explanatory variables are used in each case to serve different decision support mandates. For example, in the short, daily, term building level, forecasting may serve as a performance baseline. On the other end, we have long-term, policy-oriented forecasting exercises. TIMES (an acronym for The Integrated Markal Efom System) allows us to model supply and anticipated technology shifts over a long-term horizon, often extending as far away in time as 2100. Between these two time frames, we also have a mid-term forecasting time frame, that of a few years ahead. Investigations here are aimed at policy support, although in a more mid-term horizon, we address issues such as investment planning and pricing. In this paper, we develop and evaluate statistical and neural network approaches for this mid-term forecasting of final energy and electricity for the residential sector in six EU countries (Germany, the Netherlands, Sweden, Spain, Portugal and Greece). Various possible approaches to model the explanatory variables used are presented, discussed, and assessed as to their suitability. Our end goal extends beyond model accuracy; we also include interpretability and counterfactual concepts and analysis, aiming at the development of a modelling approach that can provide decision support for strategies aimed at influencing energy demand.

Global and Local Approaches for Forecasting of Long-Term Natural Gas Consumption in Poland Based on Hierarchical Short Time Series

“…Bartels and others [42] used gas consumption profiles of households and their economic characteristics, macro-and microeconomic data including regional differences in natural gas consumption and climatic data. Sakas et al [43] proposed a methodology for modeling energy consumption, including natural gas, in the residential sector in the medium term, taking into account economic, weather and demographic data. Hribar et al [44] used various machine learning models to forecast short-term natural gas consumption in the city of Ljubljana.…”

A Future Direction of Machine Learning for Building Energy Management: Interpretable Models

“…This method creates nearby samples with minimal feature changes that alter the model's output. Sakkas et al [208] selected features through statistical analysis and then utilized them for diverse counterfactual explanation (DiCE) model to conduct counterfactual analysis to interpret energy-demand forecasting. Tran et al [209] developed an innovative context-aware evolutionary learning algorithm (CELA) to both increase the capabilities of existing evolutionary learning methods in handling many features and datasets, and to provide an interpretable model based on the automatically extracted contexts.…”

Technology Readiness Levels (TRLs) in the Era of Co-Creation

“…TRUST-AI (www.trustai.eu, accessed on 14 April 2024) is a Horizon-funded research project running in the period 2020-2025 and aiming to develop a framework of so-called explainable AI approaches and their validation in diverse use cases. One such case was in the energy domain [40,41], where our specific responsibility was to customize an explainable energy forecasting approach. Though a variety of AI forecasting models have been developed over the last twenty years, our forecasting approach would differ in that it could provide user explanations of the forecast.…”