Forecasting models of demand in supply chain with high product diversification using gradient boosting machine learning methods

Fahimi, Mohammad Reza; Rajabzadeh Ghatari, Ali; Shoar, Maryam; Khademi, Maryam

doi:10.22075/ijnaa.2024.34189.5103

Forecasting models of demand in supply chain with high product diversification using gradient boosting machine learning methods

Articles in Press

Document Type : Research Paper

Authors

Mohammad Reza Fahimi ¹

Ali Rajabzadeh Ghatari ²

Maryam Shoar ¹

Maryam Khademi ³

¹ Department of Industrial Management, Islamic Azad University, North Tehran Branch, Tehran, Iran

² Department of Industrial Management, Faculty of Management and Economics, Tarbiat Modares University, Tehran, Iran

³ Department of Applied Mathematics, Islamic Azad University, South Tehran Branch, Tehran, Iran

10.22075/ijnaa.2024.34189.5103

Abstract

Forecasting demand in the supply chain is an essential and challenging issue for determining effective strategies and informed decisions. The related challenges also increase by raising the variety and number of products. Providing frameworks and methods with the flexibility, accuracy, and advantages necessary to forecast all product categories is crucial for managers, despite product diversity, differences in applications and features, and varying data volumes. In this regard, two supervised learning models, XGBoost Regressor (XGBR) and Gradient Boosting Regressor (GBR), have been implemented on the Global Superstore dataset on the Kaggle site. This dataset contains 3788 products in three diverse product categories, seventeen subcategories, and 51,290 orders. The limited data volume of the products prevented the possible and helpful forecasting of many products and the obtaining of appropriate results from the methods. In this experimental research, demand forecasting is used in strategic decisions. Therefore, a product aggregation approach has been proposed for this problem, which can be predicted separately according to the similarity in the subcategory products. The data of the dataset was increased using the Data Augmentation technique to investigate the effect of the amount of data on the performance of the models, and the forecasting results of the two models were compared by re-running the models. Based on the forecasting results with increased data with MSE and MAE metrics, the XGBR model achieved the lowest values of 0.12 and 0.10, respectively, and the GBR model achieved values of 0.13 and 0.15. Also, the result of the $D^{2}$ Score Metric was 0.97 in the XGBR model and 0.96 in the GBR model. The values of the error metrics decreased dramatically and by more than 80% as the data increased, and XGBR had a relative advantage in the larger data. The proposed framework and models can be used in industries with similar issues at the strategy level.

Keywords

demand forecasting

high product diversification

supply chain

machine learning

gradient boosting

[1] A. Aamer, L. Eka Yani, and I. Alan Priyatna, Data analytics in the supply chain management: Review of machine learning applications in demand forecasting, Operat. Supply Chain Manage.: Int. J. 14 (2020), no. 1, 1-13.

[2] M. Abolghasemi, B. Abbasi, and Z. HosseiniFard, Machine learning for satisficing operational decision making: A case study in blood supply chain, Int. J. Forecast. In Press (2023).

[3] L.A.C. Andrade and C.B. Cunha, Disaggregated retail forecasting: A gradient boosting approach, Appl. Soft Comput. 141 (2023), 110283.

[4] M.Z. Babai, J.E. Boylan, and B. Rostami‑Tabar, Demand forecasting in supply chains: a review of aggregation and hierarchical approaches, Int. J. Product. Res. 60 (2022), no. 1, 324-348.

[5] B. Berman, J.R. Evans, and P. Chatterjee, Retail Management‑ A Strategic Approach, 13th eds, Pearson Education India, 2018.

[6] J. Brown, Global superstore data of 2016, Kaggle, https://www.kaggle.com/datasets/jamsbrown/global‑superstore‑data‑of‑2016/data, 2023.

[7] T. Chen and C. Guestrin, Xgboost: A scalable tree boosting system, Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discov. Data Min., 2016, pp. 785--794.

[8] C.F. Chien, C.C. Ku, and Y.Y. Lu, Ensemble learning for demand forecast of after‑market spare parts to empower data‑driven value chain and an empirical study, Comput. Ind. Eng. 185 (2023), 109670.

[9] S.‑L. Developers, Metrics and scoring: Quantifying the quality of predictions, User Guide, Disponivelem: https://scikit‑learn.org/stable/modules/modelevaluation.html, 26 (2021).

[10] M. Fisher and A. Raman, Using data and big data in retailing, Product. Operat. Manag. 27 (2018), no. 9, 1665-1669.

[11] J.H. Friedman, Greedy function approximation: a gradient boosting machine, Ann. Statist. 29 (2001), no. 5, 1189-1232.

[12] S. Islam, S.H. Amin, and L.J. Wardley, A supplier selection and order allocation planning framework by integrating deep learning, principal component analysis, and optimization techniques, Expert Syst. Appl. 235 (2024), 121121.

[13] R.V. Joseph, A. Mohanty, S. Tyagi, S. Mishra, S.K. Satapathy, and S.N. Mohanty, A hybrid deep learning framework with CNN and Bi‑directional LSTM for store item demand forecasting, Comput. Electrical Engin. 103 (2022), 108358.

[14] E.R. Kone and M.H. Karwan, Combining a new data classification technique and regression analysis to predict the Cost‑To‑Serve new customers, Comput. Ind. Eng. 61 (2011), no. 1, 184-197.

[15] H.J. Kwon, S. Park, Y.H. Park, S.M. Baik, and D.J. Park, Development of blood demand prediction model using artificial intelligence based on national public big data, Digital Health 10 (2024), 20552076231224245.

[16] K.H. Lee, M. Abdollahian, S. Schreider, and S. Taheri, Supply chain demand forecasting and price optimisation models with substitution effect, Mathematics 11 (2023), no. 11, 2502.

[17] M. Levy and D. Grewal, Retail Management, McGraw‑Hill, 2023.

[18] D.M. Malioutov, K.R. Varshney, A. Emad, and S. Dash, Learning interpretable classification rules with boolean compressed sensing, T. Cerquitelli, D. Quercia and F. Pasquale, Transparent data mining for big and small data, Studies in Big Data, Springer, Cham 32 (2017), 95-121.

[19] A. Natekin and A. Knoll, Gradient boosting machines, a tutorial, Front. Neurorobotics 7 (2013), 21.

[20] F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, and E. Duchesnay, Scikit‑learn: Machine learning in Python, J. Machine Learn. Res. 12 (2011), 2825-2830.

[21] M. Seyedan and F. Mafakheri, Predictive big data analytics for supply chain demand forecasting: methods, applications, and research opportunities, J. Big Data 7 (2020), 53.

[22] Y. Song and X. Hu, Learning‑based demand‑supply‑coupled charging station location problem for electric vehicle, (2023).

[23] E. Sugiharti, R. Arifudin, D.T. Wiyanti, and A.B. Susilo, Convolutional neural Network‑XGBoost for accuracy enhancement of breast cancer detection, J. Phys.: Conf. Ser. IOP Publishing 1918 (2021), no. 4, 042016.

[24] A.A. Syntetos, Z. Babai, J.E. Boylan, S. Kolassa, and K. Nikolopoulos, Supply chain forecasting: Theory, practice, their gap and the future, Eur. J. Oper. Res. 252 (2016), no. 1, 1-26.

[25] Y. Yang, I.G. Morillo, and T.M. Hospedales, Deep neural decision trees, arXiv preprint arXiv:1806.06988 (2018).

[26] Y. Zhang and A. Haghani, A gradient boosting method to improve travel time prediction, Transport. Res. Part C: Emerg. Technol. 58 (2015), 308-324.