Abstract:Black-box models have demonstrated remarkable accuracy in forecasting building energy loads. However, they usually lack interpretability and do not incorporate domain knowledge, making it difficult for users to trust their predictions in practical applications. One important and interesting question remains unanswered: is it possible to use intrinsically interpretable models to achieve accuracy comparable to that of black-box models? With an aim of answering this question, this study proposes an intrinsically interpretable machine learning-based method to forecast building energy loads. It creatively combines two intrinsically interpretable machine learning algorithms: clustering decision trees and adaptive multiple linear regression. Clustering decision trees aim to automatically identify various building operation conditions, allowing for the training of multiple models tailored to each condition. It can reduce the complexity of model training data, leading to higher accuracy. Adaptive multiple linear regression is an improved regression algorithm tailored to building energy load prediction. It can adaptively modify regression coefficients according to building operations, enhancing the non-linear fitting capability of multiple linear regression. The proposed method is evaluated utilizing the operational data from an office building. The results indicate that the proposed method exhibits comparable accuracy to both random forests and extreme gradient boosting. Furthermore, it shows significantly superior accuracy, with an average improvement of 10.2%, compared with some popular black-box algorithms such as artificial neural networks, support vector regression, and classification and regression trees. As for model interpretability, the proposed method reveals that historical cooling loads are the most crucial for predicting building cooling loads under most conditions. Additionally, outdoor air temperature has a significant contribution to building cooling load prediction during the daytime on weekdays in summer and transition seasons. In the future, it will be valuable to explore integrating the laws of physics into the proposed method to further enhance its interpretability.

A Tailored Artificial Intelligence Model for Predicting Heat Strain of Construction Workers

Development of an Early-Warning System for Site Work in Hot and Humid Environments: A Case Study

An Artificial Neural Network-based Model That Can Predict Inpatients’ Personal Thermal Sensation in Rehabilitation Wards

Safety Risk Prediction Model of High-Rise Building Construction Based on Key Physiological Index

Hourly Heat Load Prediction for Residential Buildings Based on Multiple Combination Models: A Comparative Study

Explainable machine learning framework to predict the risk of work-related neck and shoulder musculoskeletal disorders among healthcare professionals

Comparison of algorithms for heat load prediction of buildings

Towards non-intrusive and high accuracy prediction of personal thermal comfort using a few sensitive physiological parameters

The Personalized Thermal Comfort Prediction Using an MH-LSTM Neural Network Method

Forecasting personal heat strain under extremely hot environments: Utilizing feature importance in machine learning

An integrated thermophysiological model for predicting thermal response of medical staff wearing medical protective clothing in summer

Construction Worker Fatigue Prediction Model Based on System Dynamic

Application of machine learning techniques for predicting the consequences of construction accidents in China

Intrinsically Interpretable Machine Learning-Based Building Energy Load Prediction Method with High Accuracy and Strong Interpretability

Tunnel construction worker safety state prediction and management system based on AHP and anomaly detection algorithm model

A Multi-Segmented Human Bioheat Model for Asymmetric High Temperature Environments

PREDICT THERMAL SENSATION OF CHINESE PEOPLE USING A THERMOPHYSIOLOGICAL AND COMFORT MODEL

A Model-Driven Learning Approach for Predicting the Personalized Dynamic Thermal Comfort in Ordinary Office Environment

Development of Personal Comfort Model and Its Use in the Control of Air Conditioner

Physics-guided LSTM model for heat load prediction of buildings

Thermal preference prediction based on occupants’ adaptive behavior in indoor environments- A study of an air-conditioned multi-occupancy office in China