From 2016 to 2021, scientists studied Chile's air quality by using data from the National Air Quality Information System (SINCA). They focused on three cities: Quintero, Puchuncaví, and Coyhaique. The goal? To create models that can predict pollution levels like sulfur dioxide (SO2), tiny particles (PM2. 5), and bigger particles (PM10). They combined two types of models: Autoregressive Integrated Moving Average (ARIMA) and Artificial Neural Networks (ANN). This mix is like using both a calculator and a computer to get better answers. They also considered factors like wind speed and direction to make the predictions more accurate. Four key monitoring stations—Quintero, Ventanas, Coyhaique I, and Coyhaique II—played an important role. These stations helped understand where pollution comes from and how wind affects it. In Coyhaique, something interesting happened: during winter, pollution gets trapped because of the city's location and cold temperatures. This makes pollution levels higher. The models were tested using different methods like the Akaike Information Criterion (AIC) and Ljung-Box tests. The combined ARIMA-ANN models worked really well, with an R² score over 0. 90. This shows that using both types of models is a great way to predict air quality. The results also show that Chile needs different strategies to manage air quality because the environment plays a big role. Neural networks showed how advanced technology can improve forecasts. In Coyhaique, geography and weather make pollution worse, so special care is needed. This study teaches us that combining different approaches and considering local factors can help us better understand and predict air quality.

Air Quality in Chile: A Mix of Math and AI for Better Predictions

Actions