Comparative Analysis of Random Forest and ANN Models for Urban Air Quality Index Prediction
Abstract
Predicting AQI accurately is really important these days, especially for places that are growing fast and facing more pollution. In this project, we compared two ML models Random Forest (RF) and Artificial Neural Network (ANN) to see how well they can predict AQI in some big cities of Uttarakhand, like Dehradun, Haldwani, Haridwar, and Rudrapur. We used a secondary dataset for this and built both models using MATLAB. To check how good the predictions were, we used metrics like RMSE, MAE, and R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>. The results showed that the ANN model gave better performance overall. It got an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> value of 0.79289, RMSE of 4.8104, and MAE of 3.7922, which is quite good. On the other hand, the RF model was not as accurate its R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> was only 0.50709, and its RMSE and MAE were 10.124 and 8.0625, which are higher. So, it’s clear and simple that ANN is better at handling the complex and nonlinear relationships between pollution and weather factors. This kind of study is useful for improving air quality monitoring, especially in areas like Uttarakhand where data isn’t always available in real-time. Using AI like this can really help both the public and the government to take quicker actions for health and the environment.