Machine Learning for detecting Microbial and chemical Contaminants in sachet Water

This study addresses the critical challenge of detecting microbial and chemical contaminants in sachet water in Nigeria using machine learning (ML) techniques. Traditional methods for water quality assessment are often time-consuming, costly, and ill-suited for real-time monitoring, particularly in resource-limited settings. We propose a novel approach that leverages supervised ML algorithms, including Gradient Boosting (GBC) and Random Forest (RF), to predict water potability based on an augmented dataset of 20 parameters, encompassing both microbial contaminants (e.g., Escherichia coli, Salmonella) and chemical contaminants (e.g., lead, arsenic). The dataset was enhanced using synthetic data generation techniques to address gaps in the original dataset, which lacked comprehensive coverage of critical contaminants. Our results demonstrate that the Gradient Boosting Classifier (GBC) achieves an accuracy of 99.8% and an F1 score of 99.7% on the augmented dataset, significantly outperforming other models. Feature importance analysis revealed that Escherichia coli, Salmonella, and lead were the most critical predictors of water potability, aligning with public health concerns. This study highlights the potential of ML for enhancing water quality monitoring, offering a scalable and cost-effective solution to mitigate waterborne diseases in regions like Nigeria, Nigeria. Future work will focus on integrating real-time sensor data and validating the model in real-world scenarios to further improve its applicability and impact.