Paper Title
IoT-Based Air Quality Monitoring and AQI Prediction

Abstract
The increasing issue of air quality-related health problems has resulted in the need for a reliable, consistent, and precise means by which we can evaluate and monitor air quality on a regular basis. The main reason for using the Enhanced Gradient Boosting (XGBoost) regression model for estimating and predicting future air quality indices is its ability to provide accurate, reliable results while handling a large number of features. A second important consideration was including multiple environmental parameters such as NO₂, CO₂, CO, temperature, and relative humidity within the dataset inputted into the model. In addition to estimating and predicting future air quality index, the model was integrated with a rule-based anomaly detection engine that utilized WHO-approved thresholds for the determination of unhealthy air quality levels and hazardous gases. This enabled the detection of abnormal environmental conditions and corresponding public health concerns. The performance of the method was evaluated using various standard evaluation metrics, including MAE, RMSE, R2 (coefficient of determination), and MAPE, therefore confirming that the proposed system will provide very accurate and reliable results as well as extensive prediction capabilities. The hardware captures data on various contaminants—including particulate matter (PM2.5), carbon monoxide, carbon dioxide, and nitrogen dioxide—as well as ambient temperature and humidity. Local data analysis is performed using predictive algorithms. Utilizing efficient platforms such as TensorFlow and Keras libraries, the setup can instantly recognize hazardous pollution spikes or gas leaks, issuing warnings through them. Beyond monitoring, it delivers practical guidance, such as suggesting safer times for outdoor activities, aligning with smart city objectives through a design that is both modular and cost-effective for deployment in residential, educational, or public environments. Keywords - Edge AI, Air Quality Monitoring, IoT Sensors, PM2.5, AQI Prediction, Anomaly Detection, ESP32, LSTM, XGBoost, TensorFlow Lite, Matplotlib.