Prediksi Safety Stock Menggunakan Algoritma Support Vector Regression Dengan Optimasi Hyperparameter
Abstract
Inventory management plays an essential role in ensuring smooth supply chain operations and preventing both stockouts and overstocking. One effective approach to determining the appropriate level of safety stock is the use of machine learning–based prediction methods. This study aims to predict safety stock values using the Support Vector Regression (SVR) method with hyperparameter optimization through GridSearch. The dataset used is a public bike rental dataset, which includes variables such as time, weather, season, and holidays. The research stages include data preprocessing, Exploratory Data Analysis (EDA), implementation of the SVR model, and model performance evaluation using MAE, MAPE, and R² metrics. The performance of the Support Vector Regression (SVR) algorithm with a Radial Basis Function (RBF) kernel and optimal parameters demonstrates strong predictive accuracy. Based on the Mean Absolute Percentage Error (MAPE), the model achieved prediction accuracies of 83.8%, 80.1%, and 81.7% on the training, validation, and testing datasets, respectively, indicating its effectiveness in modeling non-linear data. This model is capable of generating more precise safety stock predictions, thereby supporting decision-making in inventory planning and reducing the risk of stock shortages.
Keywords: Safety Stock; Support Vector Regression; Demand Prediction; Hyperparameter Tuning
Abstrak
Manajemen persediaan memiliki peran penting dalam menjaga kelancaran rantai pasok dan menghindari kekurangan maupun kelebihan stok. Salah satu pendekatan yang efektif untuk menentukan jumlah safety stock adalah dengan memanfaatkan metode prediksi berbasis machine learning. Penelitian ini bertujuan untuk memprediksi nilai safety stock menggunakan metode Support Vector Regression (SVR) dengan optimasi hyperparameter tuning melalui teknik GridSearch. Data yang digunakan merupakan dataset publik penyewaan sepeda yang mencakup variabel waktu, cuaca, musim, dan hari libur. Tahapan penelitian meliputi preprocessing data, analisis Exploratory Data Analysis (EDA), penerapan model SVR, serta evaluasi kinerja model menggunakan metrik MAE, MAPE, dan R². Hasil penelitian menunjukkan bahwa model SVR dengan kernel RBF dan parameter optimal menghasilkan tingkat akurasi prediksi yang baik. Berdasarkan nilai MAPE, model mencapai akurasi 83,8% pada data latih, 80,1% pada data validasi, dan 81,7% pada data uji, yang mengindikasikan kemampuan SVR dalam memodelkan data non-linear. Model ini mampu memprediksi kebutuhan safety stock lebih baik, sehingga dapat membantu pengambilan keputusan dalam perencanaan persediaan dan mengurangi risiko kekurangan stok.
Kata kunci: Safety Stock; Support Vector Regression; Prediksi Permintaan; Hyperparameter Tuning
References
E. Fatma and D. S. Pulungan, “Analisis Pengendalian Persediaan Menggunakan Metode Probabilistik dengan Kebijakan Backorder dan Lost sales,” J. Tek. Ind., vol. 19, no. 1, p. 38, 2018, doi: 10.22219/jtiumm.vol19.no1.40-51.
S. Jaipuria and S. S. Mahapatra, “A hybrid forecasting technique to deal with heteroskedastic demand in a supply chain,” Oper. Supply Chain Manag., vol. 14, no. 2, pp. 123–132, 2021, doi: 10.31387/oscm0450291.
D. A. Efrilianda, Mustafid, and R. R. Isnanto, “Inventory control systems with safety stock and reorder point approach,” 2018 Int. Conf. Inf. Commun. Technol. ICOIACT 2018, vol. 2018-Janua, pp. 844–847, 2018, doi: 10.1109/ICOIACT.2018.8350766.
C. J. Lee and S. C. Rim, “A Mathematical Safety Stock Model for DDMRP Inventory Replenishment,” Mathematical Problems in Engineering., vol. 2019, pp. 1-10, 2019, doi: 10.1155/2019/6496309.
T. Gong, J. Xu, W. Bian, and Z. Li, “Optimization of two-echelon supply chain metering device safety stock placement with uncertain demand,” 14th Int. Conf. Serv. Syst. Serv. Manag. ICSSSM 2017 - Proc., pp. 1–6, 2017, doi: 10.1109/ICSSSM.2017.7996162.
G. Han and X. A. Sun, “A research on the storage strategy of university apartment spare parts based on the stochastic inventory model,” 2016 Int. Conf. Logist. Informatics Serv. Sci. LISS 2016, pp. 1–4, 2016, doi: 10.1109/LISS.2016.7854353.
Y. Boulaksil, “Safety stock placement in supply chains with demand forecast updates,” Oper. Res. Perspect., vol. 3, pp. 27–31, 2016, doi: 10.1016/j.orp.2016.07.001.
W. Bing and P. Chen, “Research on Supply Chain safety Inventory forecast based on GA-BP Neural Network,” RASSE 2021 - IEEE Int. Conf. Recent Adv. Syst. Sci. Eng. Proc., pp. 1–5, 2021, doi: 10.1109/RASSE53195.2021.9686846.
T. Inprasit and S. Tanachutiwat, “Reordering point determination using machine learning technique for inventory management,” ICEAST 2018 - 4th Int. Conf. Eng. Appl. Sci. Technol. Explor. Innov. Solut. Smart Soc., 2018, doi: 10.1109/ICEAST.2018.8434473.
Y. Cai, M. F. Zhang, and L. Huang, “Safety stock management based on lead time optimization,” 2011 Int. Conf. Electr. Technol. Civ. Eng. ICETCE 2011 - Proc., pp. 6150–6153, 2011, doi: 10.1109/ICETCE.2011.5774201.
K. Jakkraphobyothin, S. Srifa, and T. Chinda, “Factor Analysis of Inventory Management in Thai Construction Industry,” TIMES-iCON 2018 - 3rd Technol. Innov. Manag. Eng. Sci. Int. Conf., pp. 1–5, 2019, doi: 10.1109/TIMES-iCON.2018.8621817.
B. B. Barrios, A. A. Juan, J. Panadero, K. Altendorfer, A. J. Peirleitner, and A. Estrada-Moreno, “On the use of Simheuristics to Optimize Safety-Stock Levels in Material Requirements Planning with Random Demands,” Proc. - Winter Simul. Conf., vol. 2020-Decem, no. 2017, pp. 1539–1550, 2020, doi: 10.1109/WSC48552.2020.9383988.
P. R. Hakim and H. Prastawa, “Forecasting Demand &Usulan Safety Stock Pasir Silika Dengan Metode Time Series Pada Pt Solusi Bangun Indonesia Tbk. PABRIK …,” Ind. Eng. Online J., vol. 11, no. 4, 2022, [Online]. Available: https://prosiding.seminar-id.com/index.php/sainteks
A. Jain, V. Karthikeyan, B. Sahana, S. Br, K. Sindhu, and S. Balaji, “Demand Forecasting for E-Commerce Platforms,” 2020 IEEE Int. Conf. Innov. Technol. INOCON 2020, pp. 9–12, 2020, doi: 10.1109/INOCON50539.2020.9298395.
S. Singh, “production system using various Machine Learning,” 2019 9th Int. Conf. Cloud Comput. Data Sci. Eng., pp. 422–425, 2019.
B. K. Almentero, J. Li, and C. Besse, “Forecasting pharmacy purchases orders,” Proc. 2021 IEEE 24th Int. Conf. Inf. Fusion, FUSION 2021, pp. 1–8, 2021, doi: 10.23919/fusion49465.2021.9627017.
M. B. de Oliveira, G. Zucchi, M. Lippi, D. F. Cordeiro, N. R. da Silva, and M. Iori, “Lead Time Forecasting with Machine Learning Techniques for a Pharmaceutical Supply Chain,” Int. Conf. Enterp. Inf. Syst. ICEIS - Proc., vol. 1, no. Iceis, pp. 634–641, 2021, doi: 10.5220/0010434406340641.
K. Goswami and A. B. Kandali, “Electricity Demand Prediction using Data Driven Forecasting Scheme: ARIMA and SARIMA for Real-Time Load Data of Assam,” 2020 Int. Conf. Comput. Perform. Eval. ComPE 2020, pp. 570–574, 2020, doi: 10.1109/ComPE49325.2020.9200031.
R. Gustriansyah, D. I. Sensuse, and A. Ramadhan, “Decision support system for inventory management in pharmacy using fuzzy analytic hierarchy process and sequential pattern analysis approach,” CONMEDIA 2015 - Int. Conf. New Media 2015, 2016, doi: 10.1109/CONMEDIA.2015.7449153.
How To Cite This :
Refbacks
- There are currently no refbacks.
