EUVIMEDEUVIMED entdecken
Crossref

Machine Learning Modeling and Interactive Visualisation for Heavy Equipment Spare Parts Demand Forecasting

Emirul Bahar, Happy Siringoringo

DS Journal of Digital Science and Technology · 30.06.2026 · 5 (2) · 15-26

DOI: 10.59232/dst-v5i2p102 ↗

Kurz gefasst

Fachartikel in der Zeitschrift DS Journal of Digital Science and Technology, online erschienen am 30.06.2026. Verfasst von Emirul Bahar und 1 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Spare parts inventories for heavy equipment fleets are difficult to manage because demand is erratic, lead times are long, and the financial impact of stockouts is high. In many Indonesian operators, planning still relies on simple intermittent time series. This study develops an end-to-end framework that combines modern machine learning models with interactive visual analytics to support spare parts demand planning. The framework integrates data from historical withdrawals, maintenance work orders, equipment characteristics, and logistics records into a single pipeline for cleaning, feature construction, forecasting, and decision support. Using an illustrative sample of 36 monthly observations for 60 Stock-Keeping Units (SKUs) with different criticality levels, eight forecasting families are benchmarked: ARIMA/SARIMA, Croston-type models, Random Forest, XGBoost, support vector regression, Long Short-Term Memory (LSTM) networks, a CNN-LSTM hybrid, and an ensemble stacking model. Forecast accuracy is evaluated with MAPE, RMSE, and MAE, while the managerial effect is quantified through safety stock, inventory value, and service-level indicators. Additional diagnostic analyses convert forecast errors into risk classes so that confusion matrices and ROC curves can be used to assess classification performance. The results indicate that deep learning and especially ensemble stacking outperform classical methods on intermittent series, leading to safety stock reductions of more than 40 percent at the same target service level. The visual dashboard makes these gains transparent for planners and offers a practical blueprint for heavy equipment firms that wish to industrialise data-driven spare parts management.

Zitieren

APA

Bahar, E., & Siringoringo, H. (2026). Machine Learning Modeling and Interactive Visualisation for Heavy Equipment Spare Parts Demand Forecasting. DS Journal of Digital Science and Technology, 5(2), 15-26. https://doi.org/10.59232/dst-v5i2p102

Vancouver

Bahar E, Siringoringo H. Machine Learning Modeling and Interactive Visualisation for Heavy Equipment Spare Parts Demand Forecasting. DS Journal of Digital Science and Technology. 2026;5(2):15-26. doi:10.59232/dst-v5i2p102

BibTeX
@article{Bahar2026,
  author = {Bahar, Emirul and Siringoringo, Happy},
  title = {Machine Learning Modeling and Interactive Visualisation for Heavy Equipment Spare Parts Demand Forecasting},
  journal = {DS Journal of Digital Science and Technology},
  year = {2026},
  volume = {5},
  number = {2},
  pages = {15--26},
  doi = {10.59232/dst-v5i2p102}
}
Kostenlos starten →