Pendekatan Augmentasi Data Time Series Untuk Peramalan Kebutuhan Obat Antiretroviral Berbasis Long Short-Term Memory

Munazar; Baihaqi Siregar; Amalia

doi:10.30865/json.v7i4.9831

Authors

Munazar Universitas Sumatera Utara
Baihaqi Siregar Universitas Sumatera Utara
Amalia Universitas Sumatera Utara

DOI:

https://doi.org/10.30865/json.v7i4.9831

Keywords:

antiretroviral, augmentasi data, forecasting, LSTM, MAE non-zero, time series

Abstract

Perencanaan kebutuhan obat antiretroviral yang akurat diperlukan untuk menjaga kesinambungan terapi HIV, mencegah stock-out, dan meningkatkan efisiensi distribusi obat. Namun, peramalan masih terhambat oleh keterbatasan data historis, dominasi nilai nol, dan fluktuasi permintaan yang menyulitkan model mempelajari pola time series secara stabil. Penelitian ini mengevaluasi enam metode augmentasi data time series - Gaussian noise jittering, time warping, Fast Fourier Transform-based augmentation, MixUp time series, TimeGAN, dan Seasonal-Trend Decomposition using Loess — untuk meningkatkan akurasi peramalan berbasis Long Short-Term Memory. Dua skema diuji: penambahan data sintetis (25%, 50%, 75%, 100%) dan penggantian data asli (10%, 20%, 40%, 60%, 80%), dievaluasi menggunakan Mean Absolute Error non-zero. Hasil menunjukkan bahwa augmentasi moderat (25%–50%) meningkatkan akurasi, skema penambahan lebih stabil daripada penggantian, dan model terbaik menggunakan Seasonal-Trend Decomposition using Loess mencapai MAE non-zero sekitar 66.

References

[1] S. Ekwaro-Osire et al., “Numerical data augmentation techniques for machine learning: Methods and challenges,” Journal of Data Science and Artificial Intelligence, 2025.

[2] M. Annaki et al., “Evaluating consistency in numerical data augmentation for time series forecasting,” IEEE Access, 2024.

[3] Y. Tan et al., “Impact of synthetic data proportion on forecasting performance,” Expert Systems with Applications, 2025.

[4] C. Shorten and T. M. Khoshgoftaar, “A survey on image data augmentation for deep learning,” Journal of Big Data, vol. 6, no. 60, 2019.

[5] B. K. Iwana and S. Uchida, “An empirical survey of data augmentation for time series classification with neural networks,” PLOS ONE, vol. 15, no. 7, p. e0235799, 2020.

[6] T. T. Um et al., “Data augmentation of wearable sensor data for Parkinson’s disease monitoring using convolutional neural networks,” in Proc. 19th ACM Int. Conf. Multimodal Interaction, 2017, pp. 216–220.

[7] Q. Wen et al., “Time series data augmentation for deep learning: A survey,” in Proc. 30th Int. Joint Conf. Artificial Intelligence (IJCAI), 2021.

[8] H. Zhang et al., “mixup: Beyond empirical risk minimization,” in Proc. Int. Conf. Learning Representations (ICLR), 2018.

[9] G. Forestier et al., “Generating synthetic time series to augment sparse datasets,” in Proc. IEEE Int. Conf. Data Mining Workshops (ICDMW), 2017, pp. 865–872.

[10] J. Yoon, D. Jarrett, and M. van der Schaar, “Time-series generative adversarial networks,” in Advances in Neural Information Processing Systems (NeurIPS), 2019.

[11] I. Goodfellow et al., “Generative adversarial nets,” in Advances in Neural Information Processing Systems (NeurIPS), 2014.

[12] D. P. Kingma and M. Welling, “Auto-encoding variational Bayes,” arXiv preprint arXiv:1312.6114, 2013.

[13] H. I. Fawaz et al., “Data augmentation using synthetic data for time series classification with deep residual networks,” IEEE Transactions on Neural Networks and Learning Systems, 2018.

[14] Q. Wen et al., “Time series data augmentation: A comprehensive survey,” ACM Computing Surveys, 2022.

[15] U. Schlegel et al., “A taxonomy of numerical data augmentation techniques,” Data Mining and Knowledge Discovery, 2024.

[16] C. Esteban et al., “Real-valued medical time series generation with recurrent conditional GANs,” arXiv preprint arXiv:1706.02633, 2017.

[17] M. A. Morid, O. R. L. Sheng, and J. Dunbar, “Time series prediction using deep learning methods in healthcare,” ACM Transactions on Management Information Systems, vol. 14, no. 1, pp. 1–31, 2023, doi: 10.1145/3531326.

[18] R. Pall, Y. Gauthier, S. Auer, and W. Mowaswes, “Predicting drug shortages using pharmacy data and machine learning,” Health Care Management Science, vol. 26, no. 3, pp. 395–411, 2023, doi: 10.1007/s10729-022-09627-y.

[19] J. Yoon, D. Jarrett, and M. van der Schaar, “Time-series generative adversarial networks,” in Advances in Neural Information Processing Systems (NeurIPS), vol. 32, 2019, pp. 5508–5518.

[20] A. Mak, A. Aamer, and W. Mowaswes, “Demand forecasting model for time-series pharmaceutical data using shallow and deep neural network model,” Neural Computing and Applications, vol. 35, no. 2, pp. 1233–1247, 2023, doi: 10.1007/s00521-022-07889-9.

[21] J. Hao and F. Liu, “Improving long-term multivariate time series forecasting with a seasonal-trend decomposition-based 2-dimensional temporal convolution dense network,” Scientific Reports, vol. 14, no. 1, p. 1739, 2024, doi: 10.1038/s41598-024-52240-y.

[22] V. Alexander, “Enhancing time series data predictions: A survey of augmentation techniques and model performances,” in Proc. 2024 Australasian Computer Science Week (ACSW 2024), ACM, 2024, pp. 1–19, doi: 10.1145/3641142.3641143.

[23] M. T. Zeleke, B. W. Atsbeha, B. Y. Melaku, Y. T. Mekasha, A. W. Mekonen, and S. D. Nigatu, “Performance of antiretroviral drugs supply chain management and related challenges in Amhara National Regional State, Ethiopia,” Research in Clinical and Social Pharmacy, vol. 18, p. 100570, 2025, doi: 10.1016/j.rcsop.2025.100570.

[24] Y. Liang, J. Keylock, J. L. Peng, and L. Wu, “Effective LSTMs with seasonal-trend decomposition and adaptive learning and niching-based backtracking search algorithm for time series forecasting,” Expert Systems with Applications, vol. 236, p. 121275, 2024, doi: 10.1016/j.eswa.2023.121275.

[25] X. Li, A. H. H. Ngu, and V. Metsis, “TTS-CGAN: A transformer time-series conditional GAN for biosignal data augmentation,” arXiv preprint arXiv:2206.13676, 2022. [Scopus-indexed via IEEE EMBC proceedings]

[26] M. W. Lim, S. J. Kim, Y. Park, and N. Kwon, “A deep learning-based time series model with missing value handling techniques to predict various types of liquid cargo traffic,” Expert Systems with Applications, vol. 184, p. 115532, 2021, doi: 10.1016/j.eswa.2021.115532.