Daily Arxiv

This page organizes papers related to artificial intelligence published around the world.
This page is summarized using Google Gemini and is operated on a non-profit basis.
The copyright of the paper belongs to the author and the relevant institution. When sharing, simply cite the source.

Hierarchical Evaluation Function: A Multi-Metric Approach for Optimizing Demand Forecasting Models

Created by
  • Haebom

Author

Adolfo Gonz alez, Victor Parada

Outline

Demand forecasting in competitive and uncertain business environments requires models that integrate multiple evaluation perspectives, not just hyperparameter optimization using a single metric. This paper proposes the Hierarchical Evaluation Function (HEF), a multi-metric framework for hyperparameter optimization that integrates explainability (R2), root mean square error (RMSE), and mean accuracy (MAE). Using the Walmart, M3, M4, and M5 datasets, we optimize the model using Grid Search, PSO, and Optuna, and demonstrate that HEF outperforms single-metric reference functions. In particular, HEF excels in heterogeneous monthly time series (M3) and granular daily demand scenarios (M5). HEF improves stability, generalization, and robustness at low computational cost, contributing to improved model selection, supporting accurate demand forecasting, and supporting decision-making in dynamic and competitive business environments.

Takeaways, Limitations

•
Takeaways:
◦
Proposing a multi-metric-based HEF framework: Attempting to diversify model evaluation by integrating R2, RMSE, and MAE.
◦
Applying various datasets and optimization techniques: We validate the generality of HEF using the Walmart, M3, M4, and M5 datasets and Grid Search, PSO, and Optuna.
◦
Improved performance and stability: Outperforms single-metric-based models, particularly effective in heterogeneous monthly time series and granular daily demand forecasting.
◦
Low computational cost: HEF provides effective results at low computational cost, increasing practicality.
•
Limitations:
◦
Lack of description of specific hyperparameter optimization methodology: Insufficient detailed description of the hyperparameter optimization process using HEF.
◦
Lack of comparison with other multi-metric evaluation methods: There is no analysis comparing the performance of HEF with other multi-metric evaluation methods.
◦
Limited application cases in real business environments: Further research is needed on the application cases and effectiveness of HEF in various business environments.
👍
Made with Slashpage