Temporal fusion transformer based air quality index forecasting using multivariate time series data

Urbanisation, industrial growth, and rising traffic emissions have made air quality a critical public health and environmental concern. Accurate Air Quality Index (AQI) forecasting is essential for enabling timely warnings and informed decision-making by authorities and the public. This study develops an AQI forecasting model using the Temporal Fusion Transformer (TFT), a deep learning architecture well-suited for multivariate time-series prediction. Since real-world air quality data is often unavailable or incomplete, synthetic datasets were generated to preserve the statistical characteristics and temporal trends representative of environmental air quality databases. The proposed approach captures both short-term fluctuations and long-term