Water Quality Evaluation using Agentic AI
International Journal of Innovations in Science and Technology · 12.05.2026 · 941
DOI: 10.33411/ijist/1895 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Innovations in Science and Technology, online erschienen am 12.05.2026. Verfasst von Rabia Tehseen und 5 weiteren Autor:innen. Die Arbeit nennt 28 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Water is the basic necessity of mankind and evaluation of water quality is highly significant for both environmental sustainability and public health. Traditional approaches applied for water quality analysis lack adaptability and interpretability and are unable to provide real-time assessments and informed decisions. In this paper, a unique framework has been presented to perform water quality analysis based on Agentic AI. The proposed approach combines Rule-Based WQI Calculations and Machine Learning methods, with multi-agent systems. The proposed system is composed of specialized agents, which include data agents, planning agents, analysis agents, knowledge agents, and coordination agents all working together to create a more intelligent and reliable system. The proposed framework was evaluated using a dataset containing 5,200 water quality samples collected from multiple monitoring locations, consisting of physicochemical parameters such as dissolved oxygen, turbidity, pH, temperature, and total dissolved solids (TDS). A stratified 80:20 train-test split along with 10-fold cross-validation was employed to ensure robustness and generalization of the model. Statistical analysis demonstrated that the proposed framework significantly outperformed traditional WQI and stand-alone machine learning models with an average improvement of 13.4% in classification accuracy and 11.2% in F1-score (p < 0.05).The experimental findings indicate high levels of accuracy (91%), and precision (0.89), and recall (0.88) compared with other current WQI-based or traditional stand-alone machine learning models (ROC AUC = 0.92), thus demonstrating the superiority of the proposed model over existing approaches. High classification accuracy and reliability has been achieved by utilizing domain knowledge from the knowledge agent, along with adaptive learning from the analysis agent. In addition, according to feature importance analysis, dissolved oxygen and turbidity have been identified as the two most important characteristics for the assessment of water quality. Finally, proposed framework has the provision of explainability through SHAP/LIME methods that not only makes it user-friendly but also makes it suitable for real-world environments while giving significant importance to transparency.
Zitieren
APA
Tehseen, R., Mustaqeem, A., Qamar, N., Zaheer, A., Omer, U., & Javaid, S. (2026). Water Quality Evaluation using Agentic AI. International Journal of Innovations in Science and Technology, 941. https://doi.org/10.33411/ijist/1895
Vancouver
Tehseen R, Mustaqeem A, Qamar N, Zaheer A, Omer U, Javaid S. Water Quality Evaluation using Agentic AI. International Journal of Innovations in Science and Technology. 2026:941. doi:10.33411/ijist/1895
BibTeX
@article{Tehseen2026,
author = {Tehseen, Rabia and Mustaqeem, Anam and Qamar, Nosheen and Zaheer, Ayesha and Omer, Uzma and Javaid, Sarah},
title = {Water Quality Evaluation using Agentic AI},
journal = {International Journal of Innovations in Science and Technology},
year = {2026},
pages = {941},
doi = {10.33411/ijist/1895}
}
EUVIMED