INVENTORY DATA MINING SYSTEM USING FEDERATED MACHINE LEARNING STRATEGY

Abstract

Inventory management is a particularly important process for keeping track of levels, orders, and transactions in the inventory retailing industry. A significant quantity of data about the stocked items in inventory is generated and gathered daily in the current retailing market. It is often a job to manage stocked goods effectively due to the growing volume of transaction data and their related relationships, but it must be conducted, especially when these inventory databases are scattered in various places. To collect all the inventory data from different distributed databases, it is time consuming, insecure, and inaccurate to implement intelligent management systems; it is essential to investigate the underlying dependencies of the body inventory items. However, current inventory management systems only have a limited ability for intelligent management because they depend on the statistical analysis of historical inventory data. There has not been much progress made in putting in place intelligent inventory management systems to use combined data-driven analysis to uncover hidden relationships. In this paper, we use distributed computing resources to process an enormous amount of inventory data and integrate federated learning, one of the most recent developments in data mining techniques. Federated Learning offers thorough aid for conducting various inventory management duties, including showing anomalous items and examining the aging of the inventory. Keywords: Data Mining, Distributed Database, Inventory, Federated Learning, Machine Learning DOI： https://doi.org/10.35741/issn.0258-2724.58.3.51