Delhi's Air Pollution: Can AI Models Predict It Better?

Have you ever wondered how air pollution affects your city? In Delhi, air quality is a big deal. Scientists have been trying to predict it better using something called artificial intelligence, or AI. They've been looking at several pollutants like Carbon Monoxide, Nitrogen Oxides, Ozone, Sulphur Dioxide, and tiny particles called PM2. 5 and PM10. They collected data from air quality stations around Delhi from late 2020 to early 2023.

Five different AI models were tested: LSTM, Bi-LSTM, GRU, MLP, and XGBoost. Each model had its own way of predicting how polluted the air would be. The Bi-LSTM model was the best at predicting most pollutants, including CO, NO, NO2, SO2, PM10, and NH3. It scored really high on something called R², which is like a grade for how well the model predicts things. But not all pollutants are the same. For Ozone and PM2. 5, other models like GRU and LSTM did a better job. This shows that different AI models can be better at predicting different things. These findings are important because they help us understand how AI can improve air quality forecasts. This can lead to better decisions about the environment.

Actions