Demand forecast with information centralization based on machine learning

• Main Author: Xie, Jiyun
• Other Authors: Chen Songlin
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering; Demand forecasting; Machine learning; Supply chain management
• Online Access: https://hdl.handle.net/10356/173809
• Institution: Nanyang Technological University

Demand forecasting is a critical element in modern supply chain operations and a key subject of study in the field of supply chain academia. It was influenced by various forms of relevant information, encompassing historical demand data, current market conditions, economic trends, and other demand-influencing factors. While much scholarly attention has been directed at mitigating the bullwhip effect through vertical information sharing to enhance demand forecasting accuracy, this paper shifts focus to information centralization. Specifically, it examines how stakeholders at the same supply chain level can improve forecasting accuracy by centralizing demand information horizontally. This paper introduces two novel methods for information centralization to improve demand forecasting and validates these methods using open-source datasets. Artificial neural network models from machine learning are utilized for demand forecasting practices. Additionally, it offers a comparative analysis of these methods, highlighting potential areas for enhancement and further research.