Robust building energy retrofit evaluation under uncertainty: An interpretable machine learning approach
Bibliographic record
Abstract
Improving the energy efficiency and thermal comfort of existing residential buildings is essential for sustainable urban development. However, uncertainties in occupant behaviors and building constructions pose challenges to optimizing retrofit strategies. This study presents an integrated approach combining physics-based energy simulation, interpretable machine learning, and multi-objective optimization to quantify these uncertainties and identify optimal retrofit strategies. Latin Hypercube Sampling was used to generate representative variability in occupant and construction parameters, while Extreme Gradient Boosting (XGBoost) served as a surrogate model to reduce the computational burden of detailed simulations. Shapley Additive exPlanations (SHAP) and standardized regression coefficients were applied to enhance model interpretability and identify key features influencing energy and comfort performance. The approach was applied to representative Canadian single-detached houses across four climate zones and three HVAC systems: natural gas furnaces, air source heat pumps, and ground source heat pumps. XGBoost achieved R 2 values above 0.85 for energy consumption and 0.95 for discomfort hours in most scenarios. Heating setpoint temperature, airtightness, and equipment power density emerged as dominant factors, with their relative importance varying by HVAC types and climate conditions. Passive retrofits were more impactful in colder zones, while behavioral adjustments were more effective in milder climates. The surrogate-based model reduced computation time by 89.7%. Pareto optimal solutions demonstrated up to 48% energy savings and 32% discomfort hour reductions. The proposed method offers actionable insights for policymakers and practitioners to implement targeted, efficient, and adaptive retrofit strategies under uncertainty.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".