Air Quality Index Forecasting using VMD-FAN Hybrid Model

This research project proposes a novel VMD-FAN hybrid model for forecasting Air Quality Index (AQI) time series data. The model addresses the complexity and non-stationary nature of AQI data by combining two powerful techniques: - Variational Mode Decomposition (VMD) for handling noise and decomposing the signal into Intrinsic Mode Functions (IMFs) - Fourier Analysis Networks (FAN) for capturing periodicity patterns in the data The model was tested on air quality data from Taiwan's Annan district, achieving impressive performance metrics: MAE of 0.72, MSE of 1.35, RMSE of 1.16, and MAPE of 1.49%. The model's generalizability was validated through extension analysis on different cities in Taiwan, demonstrating its robustness and practical applicability in environmental monitoring.