Full-Feature versus Limited-Input Machine Learning for Residential Energy Estimation: A Comparative Analysis of RECS and ResStock Under Realistic Input Constraints
2026-08-10 • Machine Learning
Machine Learning
AI summaryⓘ
The authors studied how well computer models can predict home energy use using easy-to-get information versus detailed data. They found that complex models work best when they have all the detailed data, but when limited to just a few simple inputs, the predictions are less accurate. For similar types of homes, fewer inputs still gave pretty good predictions. Their work shows that some machine learning models can mimic big energy datasets well, but success depends on the data used and how similar the homes are.
residential energy consumptionmachine learningCatBoostfeature selectionR2 scoreensemble modelsenergy modelingsynthetic databenchmarkingheterogeneous populations
Authors
Aditya Ramnarayan, Fatih Evren, Patti Gunderson, Samuel Rosenberg
Abstract
Residential energy estimates are often needed before detailed envelope characteristics, equipment efficiencies, infiltration, sensor, or billing data are available. This study quantifies the trade-off between predictive accuracy and input accessibility using two nationally representative U.S. residential-energy datasets: the survey-based Residential Energy Consumption Survey (RECS) and the simulation-based ResStock dataset. Full-feature models were first used to establish dataset-specific performance benchmarks. For total-energy estimation, the models were subsequently restricted to ten low-burden variables obtainable from occupants, administrative records, or location-based weather data without an on-site energy audit. Among CatBoost, XGBoost, LightGBM, Random Forest, and Neural Networks, CatBoost consistently achieved the highest predictive performance for the full-feature analysis, reaching R2 = 0.90 for ResStock and R2 = 0.73 for RECS. When the feature set was restricted to ten homeowner-accessible inputs to simulate realistic deployment conditions, model performance converged to R2 = 0.61 for RECS and R2 = 0.62 for ResStock, showing that algorithmic complexity cannot fully compensate for missing physical and behavioral information. However, for a more homogeneous ResStock cohort consisting of single-family detached, natural-gas-heated homes in Climate Zone 6A constructed between 2000 and 2010, a reduced-input model improved accuracy to R2 = 0.85, demonstrating the value of targeted modeling for homogeneous populations. The results indicate that tree-based ensemble models can serve as high-fidelity emulators of national-scale residential energy datasets. However, careful consideration of feature availability, dataset origin (empirical vs. synthetic), and applicable use cases are also important.