Maulana, Rafie Ishaq (2025) Penerapan Model CEEMDAN-LSTM dengan Optimasi Bayesian Dalam Prediksi Indeks Standar Pencemar Udara di DKI Jakarta. Undergraduate thesis, UPN Veteran Jawa Timur.
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Cover)
21081010073-cover.pdf Download (4MB) |
|
Text (Bab 1)
21081010073-bab1.pdf Download (2MB) |
|
Text (Bab 2)
21081010073-bab2.pdf Restricted to Repository staff only until 4 December 2027. Download (2MB) |
|
Text (Bab 3)
21081010073-bab3.pdf Restricted to Repository staff only until 4 December 2027. Download (2MB) |
|
Text (Bab 4)
21081010073-bab4.pdf Restricted to Repository staff only until 4 December 2027. Download (2MB) |
|
Text (Bab 5)
21081010073-bab5.pdf Download (2MB) |
|
Text (Daftar Pustaka)
21081010073-daftarpustaka.pdf Download (2MB) |
|
Text (Lampiran)
21081010073-lampiran.pdf Restricted to Repository staff only Download (1MB) |
Abstract
Poor air quality can have a negative impact on respiratory health, making the development of a prediction system essential as a risk mitigation effort. However, the Air Pollution Standard Index (ISPU) data tends to be highly fluctuating, making it difficult to predict accurately. To address this issue, the CEEMDAN decomposition method and the LSTM model were employed. Hyperparameter selection was carried out using the Bayesian optimization technique. The research results show that the combination of CEEMDAN and LSTM achieved an RMSE of 13.84, MAE of 10.71, and MAPE of 12.02%. These results indicate an improvement in performance compared to the baseline LSTM model without decomposition and optimization. With an 80:20 dataset split and the use of the Adam optimizer, the model achieved the best evaluation results, indicating an optimal balance between learning capability and generalization. The Adam optimizer proved effective in accelerating convergence and stabilizing the training process on the given dataset. The study also revealed that the use of Bayesian optimization for hyperparameter tuning did not yield a significant performance improvement compared to manual hyperparameter selection. Overall, the CEEMDAN-LSTM model proved effective in improving ISPU prediction accuracy, making it a promising tool for mitigating risks associated with poor air quality.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | Q Science > Q Science (General) T Technology > T Technology (General) |
||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Informatics | ||||||||||||
| Depositing User: | Rafie Ishaq | ||||||||||||
| Date Deposited: | 04 Dec 2025 06:04 | ||||||||||||
| Last Modified: | 04 Dec 2025 06:04 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/47790 |
Actions (login required)
 |
View Item |