A comparative performance analysis of machine learning models for compressive strength prediction in fly ash-based geopolymers concrete using reference data
Fly ash-based geopolymer (FAGP) is a promising supplementary cementitious material in the concrete industry, which can improve the sustainability and performance of concrete. This study will have made attempt to address the complexities involves in the concrete mix designs process with the aim of ac...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-07-01
|
| Series: | Case Studies in Construction Materials |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2214509525000063 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Fly ash-based geopolymer (FAGP) is a promising supplementary cementitious material in the concrete industry, which can improve the sustainability and performance of concrete. This study will have made attempt to address the complexities involves in the concrete mix designs process with the aim of achieving the desired 28-day compressive strength for FAGP. This study conducts a comparative performance analysis of the machine learning models for compressive strength prediction of FAGP using comprehensive dataset of 563 samples from 55 literature studies. Seven models such as multiple linear regression (MLR), artificial neural networks (ANNs), support vector machines (SVMs), K-nearest neighbor (KNNs), decision trees (DT), and ensemble methods combining DT with boosting and bootstrapping were employed. Model performance was evaluated using sensitivity analysis and statistical indices such as MSE, MAD, RMSE, MAPE, and SI. The findings indicate that ensemble methods and ANNs deliver superior accuracy, with R² values ranging from 0.88 to 0.92 and minimal errors. The ranking of predictive models is: ANNs > Boosting > Bootstrap SVM > DT > KNN > MLR. The prediction and profiler plots are also confirm the validity of the ANNs model performance. In addition, SHAP (SHapley Additive exPlanations) analysis demonstrated that compressive strength is significantly influenced by fly ash content and curing temperature. This study provides valuable insights for optimizing concrete mix designs, enhancing the performance and sustainability of FAGP, and contributes to the broader understanding of cementitious materials, fulfilling both industrial and environmental goals. |
|---|---|
| ISSN: | 2214-5095 |