Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{204769,
        author = {Settibathula Sridevi and Mr.K.Suresh},
        title = {DEEP LEARNING BASED REAL-TIME AIR QUALITY PREDICTION AND POLLUTION MONITORING SYSTEM},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {1},
        pages = {7036-7041},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=204769},
        abstract = {The Air quality deterioration has emerged as one of the most pressing environmental and public health challenges of the twenty-first century. Rapid urbanization, accelerated industrialization, and the exponential growth of vehicular traffic have collectively contributed to alarming levels of air pollution across cities and towns in India and worldwide. The World Health Organization estimates that approximately 7 million premature deaths occur annually due to household and ambient air pollution, with South Asian countries particularly India bearing disproportionately large share of this burden. The Air Quality Index (AQI) serves as a standardized metric that translates complex atmospheric chemistry in to an intuitive numerical scale, communicating the health implications of air pollution to the general public, policymakers, healthcare providers, and environmental regulators. This project presents a comprehensive machine learning-based system for Air Quality Analysis and Prediction. The system ingests historical pollution data collected from monitoring stations across multiple Indian states under the Central Pollution Control Board's (CPCB) National Air Quality Monitoring Programme (NAMP). The raw sensor readings undergo rigorous preprocessing including missing value imputation, irrelevant feature removal, and data type standardization before being transformed into standardized sub-indices using the established CPCB breakpoint formulae. The final AQI is computed as the maximum of these four sub-indices, consistent with the Indian National AQI standard. Multiple supervised learning algorithms are systematically evaluated for two distinct prediction tasks. For multi class classification of AQI health categories (Good, Moderate, Poor, Unhealthy, Very Unhealthy, Hazardous), four algorithms are compared: Logistic Regression, Decision Tree Classifier, Random Forest Classifier, and K-Nearest Neighbours Classifier. Classification performance is measured using overall accuracy and Cohen's Kappa score. Experimental results demonstrate that Linear Regression achieves an exception alR-squared value exceeding 0.99 on the held-out tests et. Among. classifiers, Random Forest achieves the highest test accuracy of approximately 95% with a Cohen's Kappa of 0.93, substantially outperforming Logistic Regression (73%) and KNN (86%). The entire solution is packaged as an interactive web application using Stream lit, enabling real-time user inputs for SO2, NO2, and SPM values and returning instant AQI predictions with corresponding health category labels and advisory information.},
        keywords = {Air Quality Index, Machine Learning, Deep Learning, Random Forest, AQI Prediction, Environmental Monitoring, Streamlit},
        month = {June},
        }