Home / Current Issue / Paper 1715079
Predicting Indicator Bacteria and Classifying Pathogen Risk in Improved Groundwater Supplies Using ANN, SVR, and SVC: Evidence from Makurdi Metropolis, Nigeria
Subject area: Science,Engineering and Technology · Area of research: Environmental Engineering
Abstract
Improved groundwater sources are widely relied upon for domestic water supply in Nigerian urban centers, yet their microbiological safety remains uncertain. This study assessed the microbial quality of groundwater in Makurdi metropolis and evaluated the performance of Artificial Neural Network (ANN) and Support Vector models for predicting indicator bacteria levels and sanitary safety status. A total of 384 water samples were collected from shallow boreholes and concrete-lined wells across four locations and analyzed for Most Probable Number (MPN), total coliform, Escherichia coli, and Salmonella using standard methods. Overall, 95.6% of samples were classified as Safe, while 4.4% were Unsafe due to detection of faecal indicators. Indicator bacteria levels varied widely, with MPN ranging from 2 to 95 MPN/100 mL and total coliform from 6 to 90 CFU/100 mL, demonstrating substantial spatial and temporal heterogeneity. ANN and Support Vector Regression (SVR) models were developed to predict continuous indicator bacteria outcomes from routinely available variables. ANN outperformed SVR for both MPN (R² = 0.583 vs 0.480) and total coliform (R² = 0.567 vs 0.433). For sanitary risk classification, ANN demonstrated better ability to detect Unsafe samples than Support Vector Classification under class-imbalanced conditions. Synaptic weight analysis revealed that location and month were the most influential predictors, indicating that microbial contamination was driven primarily by spatial and temporal factors. The findings show that improved sources are not inherently safe and that machine learning models can provide valuable decision-support tools for proactive groundwater quality management in resource-limited settings.
Keywords
Groundwater quality, Indicator bacteria, Artificial neural network, Support vector regression, Risk classification, Makurdi
References
[1] Adelodun, B., Choi, K. S., Jo, Y. M., & Kim, S. H. (2020). Assessment of groundwater contamination using multivariate statistical analysis in urbanized regions of developing countries. Environmental Monitoring and Assessment, 192(3), 1–15.
[2] Ahmed, W., Hamilton, K., Toze, S., & Cook, S. (2019). Seasonal variation of faecal indicator bacteria and pathogens in groundwater: Implications for water safety. Science of the Total Environment, 659, 123–131.
[3] Bain, R., Johnston, R., Mitis, F., Chatterley, C., & Slaymaker, T. (2018). Establishing sustainable development goal baselines for household drinking water, sanitation and hygiene services. Water, 10(12), 1711.
[4] Bartram, J., Corrales, L., Davison, A., Deere, D., Drury, D., Gordon, B., … Stevens, M. (2018). Water safety plan manual: Step-by-step risk management for drinking-water suppliers. World Health Organization.
[5] He, H., & Ma, Y. (2018). Imbalanced learning: Foundations, algorithms, and applications. Wiley-IEEE Press.
[6] Heddam, S., & Kisi, O. (2018). Extreme learning machines for water quality prediction: A case study. Environmental Monitoring and Assessment, 190(8), 1–18.
[7] Howard, G., Bartram, J., Pedley, S., Schmoll, O., Chorus, I., & Berger, P. (2019). Groundwater and public health. In O. Schmoll et al. (Eds.), Protecting groundwater for health (2nd ed., pp. 3–23). IWA Publishing.
[8] Odonkor, S. T., & Ampofo, J. K. (2019). Escherichia coli as an indicator of bacteriological quality of water: An overview. Microbiology Research, 10(1), 1–11.
[9] UNICEF & WHO. (2019). Progress on household drinking water, sanitation and hygiene 2000–2017. World Health Organization.
[10] World Health Organization. (2017). Guidelines for drinking-water quality (4th ed.). WHO Press.
[11] Zhu, S., Heddam, S., & Zhang, J. (2019). Artificial intelligence methods for predicting groundwater quality: A review. Journal of Hydrology, 575, 123–136.
How to cite this paper
@article{1715079,
author = {Sini, V. J, Ezeh, E. O, Ashiekaa, F. F, Sule-Otu, M. O.},
title = {Predicting Indicator Bacteria and Classifying Pathogen Risk in Improved Groundwater Supplies Using ANN, SVR, and SVC: Evidence from Makurdi Metropolis, Nigeria},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {9},
pages = {3484-3497},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1715079.pdf},
abstract = {Improved groundwater sources are widely relied upon for domestic water supply in Nigerian urban centers, yet their microbiological safety remains uncertain. This study assessed the microbial quality of groundwater in Makurdi metropolis and evaluated the performance of Artificial Neural Network (ANN) and Support Vector models for predicting indicator bacteria levels and sanitary safety status. A total of 384 water samples were collected from shallow boreholes and concrete-lined wells across four locations and analyzed for Most Probable Number (MPN), total coliform, Escherichia coli, and Salmonella using standard methods. Overall, 95.6% of samples were classified as Safe, while 4.4% were Unsafe due to detection of faecal indicators. Indicator bacteria levels varied widely, with MPN ranging from 2 to 95 MPN/100 mL and total coliform from 6 to 90 CFU/100 mL, demonstrating substantial spatial and temporal heterogeneity. ANN and Support Vector Regression (SVR) models were developed to predict continuous indicator bacteria outcomes from routinely available variables. ANN outperformed SVR for both MPN (R² = 0.583 vs 0.480) and total coliform (R² = 0.567 vs 0.433). For sanitary risk classification, ANN demonstrated better ability to detect Unsafe samples than Support Vector Classification under class-imbalanced conditions. Synaptic weight analysis revealed that location and month were the most influential predictors, indicating that microbial contamination was driven primarily by spatial and temporal factors. The findings show that improved sources are not inherently safe and that machine learning models can provide valuable decision-support tools for proactive groundwater quality management in resource-limited settings.},
keywords = {Groundwater quality, Indicator bacteria, Artificial neural network, Support vector regression, Risk classification, Makurdi},
month = {March},
doi = {https://doi.org/10.64388/IREV9I9-1715079}
}