ABSTRACT

Groundwater wells are the main source of drinking water in Hadhramout and throughout Yemen. Regression equations have been developed after a methodical examination of the relationships between water quality parameters. The quality of 41 water well data for coastal Hadhramout, Yemen, was examined in this study. In this study, correlation analysis and regression equations were used. Hydrogen ion concentration (pH), electrical conductivity (EC), turbidity (Turb), total dissolved solids (TDS), total hardness (TH), total alkalinity (TA), chloride (cl-), sulfate (SO42-), fluoride (F-), iron (Fe), nitrate (NO3-), manganese (Mn), sodium (Na+), and potassium (k+) were the physiochemical parameters that were examined. The results collected indicated that certain factors have strong relationships with one another, whereas the other parameters have weak and moderate relationships. Thus, by predicting parameters through the establishment of appropriate regression equations between them, this study facilitates the allocation of these results to reduce the expenses associated with certain analytical tools.

Keywords: Water quality parameters, Correlation Coefficient, Regression Equations, Hadhramout, Yemen.

ABSTRACT

This study explores the use of the XGBoost regression model to predict the Nano-Silica (NS) content in concrete based on features such as concrete type, compressive strength, and flexural strength. The dataset, which includes these features, was processed to handle missing data and convert categorical variables into numerical values. The data was then split into training and testing sets to evaluate the model's performance. The model achieved an R² score of 0.82, explaining approximately 82% of the variation in NS content. The Root Mean Squared Error (RMSE) of 2.25% indicated that the predictions were accurate with relatively small errors. Visual assessments, including scatter plots and residual plots, showed that the model's predictions closely matched the actual values, with residuals randomly distributed and no clear bias.The model performed well for traditional concrete types, especially Ordinary Portland Cement (OPC), where the relationship between flexural strength and NS content was clear. However, for more specialized mixes, such as High-Strength Concrete (HSC) and Self-Compacting Concrete (SCC), the model showed greater variability. This suggests that additional features, such as specific mix designs or material properties, could enhance the model's accuracy for these concrete types. Overall, the XGBoost model is a valuable tool for predicting NS content in concrete, with opportunities for further refinement to improve predictions for specialized concrete mixes.

Keywords: XGBoost, Nano-Silica, Concrete type, Compressive Strength, Flexural Strength.