Integrating Deep Learning and Inverse Distance Weighting for Spatiotemporal Prediction of Air Pollution Level

Motivation: With the rapid development of urbanization, many developing countries are facing serious air pollution. Accurate prediction of air quality is crucial for outdoor planning and government policies of the residents of these countries. Precise air quality forecasts are essential for governments to address emergencies during severe pollution events affecting public health and daily activities. This research emphasizes anticipating air quality data from observation sites, necessitating proficient modeling of spatial-temporal trends. Relevent study: Harmful effects of air pollution have attracted the attention of ordinary citizens, researchers, policy makers and smart city users. In order to minimize its negative impacts on human health and the environment, it is important to detect air pollution levels in a timely and high spatial resolution. Proposed: entails quantifying PM2.5 particle concentrations, meteorological variables, and gaseous contaminants. Following preprocessing, a hybrid approach integrating the Crayfish Optimisation Algorithm (COA) and Rider Optimisation (RO) with Inverse Distance Weighting (IDW) is employed for spatial interpolation. Key outcomes: The regional projections are subsequently utilized in a Bidirectional Long Short-Term Memory (BILSTM) model for temporal analysis and forecasting. This methodology facilitates precise air quality forecasts in Delhi, supporting efficient management.