Optimizing Supply Chain Partnerships for Incheon-Based Suppliers - A Deep Neural Network Approach -

Abstract

The purpose of this study is to predict the transaction probability between Incheon-based suppliers and nationwide buyers using a Deep Neural Network (DNN)-based Business Partner Recommendation (BPR) model and to identify key predictors of transaction success. Focusing on the unique transaction patterns of Incheon-based suppliers, this study integrates multidimensional data, including industry sector, product characteristics, geographical distance, and transaction volume, to present a Transaction Probability Score that assesses the likelihood of successful partnerships. The analysis reveals that industry compatibility and product specialization are more significant predictors of successful transactions for Incheon-based suppliers than geographical proximity. These findings highlight the value of deep learning models in enhancing supply chain efficiency and optimizing partnership recommendations within complex trade networks.