PSI - Issue 70

Santhosh Kumar N V et al. / Procedia Structural Integrity 70 (2025) 440–446

445

[a]

[b]

[c]

[d]

[e] Fig. 1. (a) MLR; (b) KNN; (c) RF; (d) XGB; (e) SVM; Scatter plots of actual vs. predicted values for training and test datasets 5. Conclusion This research demonstrates the effectiveness of machine learning models in predicting lime mortar compressive strength based on key material inputs. MLR performed consistently well with both raw and scaled data (Adj. R² = 0.68), capturing linear relationships. SVM showed a significant improvement with scaled data (Adj. R² = 0.79), indicating its sensitivity to scaling. KNN also benefited from scaling (Adj. R² = 0.60), highlighting its advantage in non-linear models. RF and XGBoost performed consistently with respect to both data formats, being insensitive to scaling because of their tree-based structure. The improvement in the performance of SVM with normalization highlights its promise for mortar strength prediction with appropriate pre-processing. Future studies may investigate the inclusion of other variables such as moisture content and curing conditions, and more sophisticated methods such as deep learning or hybrid models for more intricate relationships. Optimal data pre-processing such as feature engineering and dimension reduction may enhance performance. Real-time construction site prediction systems, transfer learning for generalized applicability, and longitudinal studies to compare long-term durability could increase model robustness and real-world utility. References Ahmad, A., Ahmad, W., Aslam, F., & Joyklad, P. 2022. Compressive strength prediction of fly ash-based geopolymer concrete via advanced machine learning techniques. Case Studies in Construction Materials, 16, e00840. Akbari, M., & Jafari Deligani, V. 2020. Data-driven models for compressive strength prediction of concrete at high temperatures. Frontiers of Structural and Civil Engineering, 14(2), 311 – 321. Akossou, A. Y. J., & Palm, R. 2013. Impact of data structure on the estimators R-square and adjusted R-square in linear regression. International Journal of Mathematics and Computation, 20(3), 84 – 93.