Explainable AutoML models for predicting the strength of high-performance concrete using Optuna, SHAP and ensemble learning
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Accurately predicting key engineering properties, such as compressive and tensile strength, remains a significant challenge in high-performance concrete (HPC) due to its complex and heterogeneous composition. Early selection of optimal components and the development of reliable machine learning (ML) models can significantly reduce the time and cost associated with extensive experimentation. This study introduces four explainable Automated Machine Learning (AutoML) models that integrate Optuna for hyperparameter optimization, SHapley Additive exPlanations (SHAP) for interpretability, and ensemble learning algorithms such as Random Forest (RF), Extreme Gradient Boosting (XGB), Light Gradient Boosting Machine (LGB), and Categorical Gradient Boosting (CB). The resulting interpretable AutoML models O-RF, O-XGB, O-LGB, and O-CB are applied to predict the compressive and tensile strengths of HPC. Compared to a baseline model from the literature, O-LGB achieved significant improvements in predictive performance. For compressive strength, it reduced the Mean Absolute Error (MAE) by 87.69% and the Root Mean Squared Error (RMSE) by 71.93%. For tensile strength, it achieved a 99.41% improvement in MAE and a 96.67% reduction in RMSE, along with increases in R 2 . Furthermore, SHAP analysis identified critical factors influencing strength, such as cement content, water, and age for compressive strength, and curing age, water-binder ratio, and water-cement ratio for tensile strength. This approach provides civil engineers with a robust and interpretable tool for optimizing HPC properties, reducing experimentation costs, and supporting enhanced decision-making in structural design, risk assessment, and other applications.
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.
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 it