ABSTRACT
This work aims to determine if the deep learning models can predict air quality using spectral imaging of the time series data. Spectrograms provide an insight into the frequency content over time, and they can detect variations in air quality that are not immediately known in raw sensor data. We utilize several deep learning architectures and models, such as convolutional neural networks (CNNs) and Long short term memory (LSTM) to these spectrograms. By doing this, the model can generate intricate patterns that relate to specific air quality situations to spectral properties. When using this method rather than more conventional time series forecasting techniques, the prediction accuracy has increased. Additionally, this method provides a new way to utilize spectral information in air quality monitoring. This study also examines the effectiveness of deep learning models in this situation.
