Faldo, Raey and Mandala, Satria and Astuti, Rina Pudji and Prihatmanto, Ary Setijadi and Mohd Zahid, Mohd Soperi (2025) APD-BayNet: Jakarta Air Quality Index Prediction Using Bayesian Optimized Tabnet. IEEE Access, 13. 57734 – 57752. ISSN 21693536
Full text not available from this repository.Abstract
The Air Quality Index (AQI) is a critical measure for assessing air pollution levels and their impact on public health and the environment. Jakarta, the capital of Indonesia, has consistently ranked among the world’s most polluted cities. Various machine learning-based studies have attempted to predict AQI levels in Jakarta, demonstrating promising results. However, existing studies suffer from several limitations, including the use of outdated datasets, lack of dataset normalization, suboptimal hyperparameter tuning, absence of k-fold cross-validation in certain experiments, and limited exploration of deep learning models for AQI detection. Consequently, the robustness of existing air pollution detection systems remains inadequate. To address these challenges, this study introduces APD-BayNet, an advanced air quality detection system that integrates the TabNet deep learning architecture with Bayesian Optimization (BO) for enhanced predictive performance. Furthermore, we compare APD-BayNet with two widely used deep learning models: Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN). Our methodology consists of four key stages: data preprocessing, model development, hyperparameter tuning using BO, and performance evaluation through 5-fold cross-validation, applied consistently across all models. Extensive experiments demonstrate that APD-BayNet consistently outperforms LSTM and CNN in AQI prediction. The TabNet model within APD-BayNet achieves high performance on the training set, with precision, recall, F1-score, and accuracy of 96.64, 95.28, 95.79, and 98.17, respectively. On the test set, these metrics remain strong at 94.30, 94.99, 94.36, and 98.17, respectively. These findings highlight the effectiveness of APD-BayNet in providing a robust and scalable solution for air quality monitoring. Future research could explore its adaptability to other geographical regions, enhancing its applicability on a global scale. © 2013 IEEE.
Item Type: | Article |
---|---|
Impact Factor: | Cited by: 0; All Open Access, Gold Open Access |
Uncontrolled Keywords: | Air quality; Long short-term memory; Air quality index prediction; Air quality indices; Convolutional neural network; Deep learning; Hyper-parameter; Index predictions; Jakarta; Learning models; Short term memory; Tabnet; Convolutional neural networks |
Depositing User: | Mr Ahmad Suhairi Mohamed Lazim |
Date Deposited: | 08 Jul 2025 16:29 |
Last Modified: | 08 Jul 2025 16:29 |
URI: | http://scholars.utp.edu.my/id/eprint/38927 |