MUMBAI, India, Aug. 24 -- Intellectual Property India has published a patent application (202641097890 A) filed by Vignan'S Nirula Institute Of Technology And Science For Women on August 13, 2026, for An Intelligent Machine Learning Framework For Water Potability Prediction Using Adaptive Data Processing, Hybrid Classification, And Explainable Artificial Intelligence.
Inventors include P. Sandhya Krishna; K. V. S. S. Rama Krishna; B. Aruna Kumari; V. Pavani; Ala Nandini; Pavuluri Sri Lekha; Patchava Pujitha; and Gumma Aparna.
The application for the patent was published on August 21, 2026, under issue no. 34/2026.
Abstract: The present invention discloses an intelligent machine learning paradigm that can be used to predict potability of water through adaptive data processing, intelligent feature engineering, hybrid machine learning classification, Explainable Artificial Intelligence (XAI), confidence evaluation, and real-time water quality monitoring. The paradigm collects physiochemical attributes of water such as pH, hardness, total dissolved solids, chloramines, sulphate, conductivity, organic carbon, trihalomethanes, and turbidity from lab databases, environmental monitoring systems, water treatment plants, and IoT-based sensors. The obtained data is adaptively processed through detection and management of missing values, duplication, noise, outliers, inconsistencies, and erroneous measurements. Intelligent feature engineering is done within the framework using correlation analysis, entropy, information gain, recursive feature elimination, and random forest feature importance to detect relevant water quality parameters. The improved data is analyzed using various supervised machine learning classification algorithms such as logistic regression, decision tree, random forest, support vector machine, gradient boosting, XGBoost, and artificial neural network. The proposed system evaluates the algorithms on the basis of performance measures including accuracy, precision, recall, F1-Score, AUC-ROC, and Matthews correlation coefficient and automatically chooses an appropriate prediction model. Moreover, the Explainable Artificial Intelligence module creates feature importance details, both local and global, to interpret the prediction result. A confidence analysis module evaluates prediction accuracy and proposes laboratory confirmation if the confidence score is less than the threshold value. The system is also capable of IoT-based real-time monitoring, automatic alert generation, decision-making, and regular updates of the model using new data on water quality. Therefore, the invention represents an adaptive, accurate, scalable, interpretable, and efficient method for classifying water into potable and non-potable.
Disclaimer: Curated by HT Syndication.
