Saved in:
| Main Authors: | , |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2026
|
| Online Access: | https://doi.org/10.5281/zenodo.20116074 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- <p>Resource artifact of undergraduate thesis titled <em>Air Quality Prediction Based on Multivariate Data and Meteorological Variables using Temporal Fusion Transformer and N-BEATS in Gresik Regency</em>.</p> <p>This repository contians five files described as follows:</p> <ol> <li><strong>raw_dataset.zip</strong>: Raw dataset used in research including past and future covariates.</li> <li><strong>processed_dataset.zip</strong>: Results of data preprocessing such as, past covariates without outliers, past covariates without outliers with sqrt transform, past covariates without outliers with log transform, future covariates with sine-cosine encoding, and correlation scores. Each past covarites preprocessed results contains different winsorized scale (1% and 5%), also with or without ozon variables.</li> <li><strong>predicted_dataset.zip</strong>: Prediction results generated by TFT and N-BEATS models. Each model includes three type of prediction results such as, prediction on several point in the data test, backtesting on the entire data test, and future forecasting results.</li> <li><strong>tuned_model.zip</strong>: Results of hyperparameter tuning for TFT and N-BEATS models using Optuna.</li> <li><strong>post_tuning_model_screening.zip</strong>: Results of post-tuning model screening, where several tuned TFT and N-BEATS models were tested on data test before selecting the overall best perfomence on each models.</li> </ol>