Hybrid data driven/thermal simulation model for comfort assessment

TL;DR

This paper introduces a hybrid approach combining real and simulated data to improve thermal comfort prediction using machine learning, achieving high accuracy with a random forest model.

Contribution

It presents a novel hybrid data generation method integrating physical simulations with real data for enhanced thermal comfort prediction.

Findings

Random forest achieved an F1 score of 0.999

Hybrid data improves model accuracy and reduces data collection costs

Benchmarking shows competitive performance of different ML methods

Abstract

Machine learning models improve the speed and quality of physical models. However, they require a large amount of data, which is often difficult and costly to acquire. Predicting thermal comfort, for example, requires a controlled environment, with participants presenting various characteristics (age, gender, ...). This paper proposes a method for hybridizing real data with simulated data for thermal comfort prediction. The simulations are performed using Modelica Language. A benchmarking study is realized to compare different machine learning methods. Obtained results look promising with an F1 score of 0.999 obtained using the random forest model.