MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202621073893 A) filed by Symbiosis International Deemed University on June 12, 2026, for Machine Learning-Based Multi-Parameter Water Quality Prediction And Classification System.
Inventors include Dr. Sandeep Kumar; Shivangi Kushwaha; Sharvari Wakde; and Shreya Maske.
The application for the patent was published on July 31, 2026, under issue no. 31/2026.
Abstract: ABSTRACT MACHINE LEARNING-BASED MULTI-PARAMETER WATER QUALITY PREDICTION AND CLASSIFICATION SYSTEM The present invention discusses machine learning-based multi-parameter water quality prediction and classification system (100) and method for automated classification of water quality into Poor (Class 0), Medium (Class 1), and Good (Class 2) categories. The system (100) comprises a data ingestion module (110), a data preprocessing module (200) performing mean substitution, duplicate removal, and label encoding, an exploratory data analysis module (300) computing Pearson correlation coefficients and frequency distributions, a feature engineering module (400) generating four interaction-derived composite features from physicochemical parameter pairs, a model training module (500) training Logistic Regression (510), Random Forest (520), and XGBoost (530) classifiers on a 4300-sample water quality dataset, a model evaluation module (600) computing accuracy, precision, recall, and F1-score metrics, and a prediction output module (700) generating real-time water quality class predictions. The XGBoost classifier achieves the highest classification accuracy of approximately 91.90%, providing scalable, cost-effective alternative to traditional laboratory-based water quality assessment. [
Disclaimer: Curated by HT Syndication.
