PSI - Issue 70

Abdul Musavvir et al. / Procedia Structural Integrity 70 (2025) 432–439

435

0.00

0.00

75% 486.000 804.450 997.890 180.000

0.00

80.000

72.000

68.200

Max 104.000 Fine Aggregate (FA); Coarse Aggregate (CA); Super plasticizer (SP); Fly Ash (FA); Ground Granulated Blast Furnace Slag (GGBS); Silica Fumes (SF); Nano Silica (NS). 600.000 992.600 1520.00 228.000 58.420 375.000 282.800 75.000 58.420

Fig. 1. Density of data in different range of compressive strength of HSC.

Fig. 2. Feature Importance Chart of XGBoost Regression Model.

3.1. Training and Testing collected data: To develop the most suitable algorithm for the extracted dataset, the ML model was developed in four different algorithms (Linear Regression, Decision Tree Regression, Random Forest and XGBoost Regression). The data was split into 80% and 20% for training and testing in the each model. Ten Hyper parameters were applied from the dataset that are Cement, Fine Aggregate, Coarse Aggregate, Water, Super plasticizer, Fly Ash, Ground Granulated Blast Furnace Slag, Silica Fumes, Nano Silica and Compressive Strength. The R2 Score, Mean Squared Error, Mean Absolute Error and Root Mean Squared Error were used for evaluating the models based on their output. On the tested ML models, the linear regression had the least performance as the HSC has lot of complex data with non-linearity.