Air Quality Prediction using ANN and RNN – A Comparative Study

Abstract:- This work aims to predict Absolute Humidity (AH) using Artificial Neural Networks (ANN) and Recurrent Neural Networks (RNN). The dataset is first preprocessed by removing columns that contain unused, null, or negative values. Noise symbols are then stripped from the remaining data, which is subsequently converted to a floating-point format to facilitate processing. Finally, feature scaling is applied to enhance model performance. The dataset includes different sensor values like carbon monoxide (CO), benzene (C6H6), nitrogen oxides (NOx), non-methane hydrocarbons, and temperature, all of which influence AH — a key measure for weather forecasting and environmental monitoring. Both ANN and RNN models are used to capture complex relationships and patterns in the data. The final step involves evaluating both models using performance metrics like Mean Absolute Error (MAE) and Mean Squared Error (MSE) to select the one that performs the best.