EUVIMEDEUVIMED entdecken
Open Access · CC BY-SA 4.0

A NEURAL NETWORK APPROACH FOR WATER QUALITY PREDICTION AND IDENTIFICATION OF SIGNIFICANT WATER QUALITY INDICATORS

Fareed Ahmad, Muhammad Usman Ghani Khan, Muhammad Imran Najeeb, Irfan Irshad

Pakistan Journal of Scientific Research · 01.06.2026 · 6 (01) · 42-49

DOI: 10.57041/0n1ms924 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Pakistan Journal of Scientific Research, online erschienen am 01.06.2026. Verfasst von Fareed Ahmad und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Water is as essential as air for the sustenance of life. Nowadays, the quality of water remains a global concern, particularly in developing countries, where inadequate infrastructure, limited resources, and lack of standard rules and regulations has further exacerbated the issue. Consumption of substandard water can cause severe illnesses and even death. Accurate prediction of water quality is a challenging task due to the complex nature of environmental datasets, which often contain noise, skewed distributions, class imbalance, and missing values. Machine learning techniques can help provide an effective solutions for handling such complex data and aid in accurate classification and prediction of features that influence the quality of water. This study focuses on classification and feature ranking using machine learning classifiers for water quality assessment. ReliefF feature ranking identified total coliforms, alkalinity, bicarbonate, pH, and TDS as the most influential water quality indicators, while fluoride, nitrate, and iron as least significant features. We further performed nested subset analysis by progressively adding ranked features and evaluated classifiers i.e. SVM, OneR, IBk, and MLP using 10-fold cross-validation. Although our proposed Neural Network model i.e., MLP(4,2) achieved the highest accuracy of 84.62% on the complete feature set, both MLP(4,2) and IBk attained a peak accuracy of 91.21% using 19, and 13 top-ranked features, respectively. Overall, the findings exhibit that the feature selection along Neural Networks can improve prediction accuracy and enable accurate water quality assessment, which can especially assist decision-making systems in environmental monitoring and public health management in countries like Pakistan that face severe water-related challenges.

Zitieren

APA

Ahmad, F., Khan, M. U. G., Najeeb, M. I., & Irshad, I. (2026). A NEURAL NETWORK APPROACH FOR WATER QUALITY PREDICTION AND IDENTIFICATION OF SIGNIFICANT WATER QUALITY INDICATORS. Pakistan Journal of Scientific Research, 6(01), 42-49. https://doi.org/10.57041/0n1ms924

Vancouver

Ahmad F, Khan MUG, Najeeb MI, Irshad I. A NEURAL NETWORK APPROACH FOR WATER QUALITY PREDICTION AND IDENTIFICATION OF SIGNIFICANT WATER QUALITY INDICATORS. Pakistan Journal of Scientific Research. 2026;6(01):42-49. doi:10.57041/0n1ms924

BibTeX
@article{Ahmad2026,
  author = {Ahmad, Fareed and Khan, Muhammad Usman Ghani and Najeeb, Muhammad Imran and Irshad, Irfan},
  title = {A NEURAL NETWORK APPROACH FOR WATER QUALITY PREDICTION AND IDENTIFICATION OF SIGNIFICANT WATER QUALITY INDICATORS},
  journal = {Pakistan Journal of Scientific Research},
  year = {2026},
  volume = {6},
  number = {01},
  pages = {42--49},
  doi = {10.57041/0n1ms924}
}
Kostenlos starten →