MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641065089 A) filed by Srm Trp Engineering College on May 23, 2026, for Interpretable Aqi Forecasting Using Machine Learning And Shap Analysis.
Inventors include Arul Prakash K; Ashvathi R; Hajira Banu S; Hariprasad K; and Maheswari R K.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: Air pollution in urban environments demands accurate, real-time, and interpretable forecasting systems for effective management. This project presents a unified Explainable Artificial Intelligence (XAI) framework for real-time Air Quality Index (AQI) prediction integrated with a policy-driven decision support system. The framework utilizes historical air quality data, live weather information, and user-selected or GPS-based location inputs to generate location-specific AQI predictions. Attention-enhanced XGBoost model helps to effectively capture the complex relationships, which in turn helps to achieve high accuracy in the prediction. It can predict the levels of AQI for the next three days using the trend-based temporal analysis method. It can also explain the results in an transparent way using the SHAP (Shapley Additive Explanations) method to identify the major sources of pollution. A context-aware policy engine develops targeted mitigation strategies based on the sources and environmental factors. Policy prioritization is done using the TOPSIS multi-criteria decision analysis based on factors such as impact, feasibility, cost, and urgency. Furthermore, fuzzy logic is applied to send early warnings before the critical AQI limits are reached. This system can send SMS messages to users with actionable recommendations in real time, thus enabling proactive urban air quality management.
Disclaimer: Curated by HT Syndication.
