Leveraging LSTM for precision inventory management by future demand forecasting

Abstract

In this work, extraction of high-utility data from massive datasets is one of the most well-known areas of research in data mining. The goal of high-utility itemset mining is to identify the inventory's most lucrative items that users tend to favor. An LSTM-based approach is suggested to determine what consumers buy most frequently. Based on these purchases, high-utility items that are expected to be in demand in the future are then identified from client buying patterns. The development of the design, which will be utilized to manage inventories going forward and identify each consumer's item set, is the main focus rather than just the price or amount of the item purchased. Additionally, related consumer groups can be put together and consumers of similar commodities can be located by expanding the use case of the model. Lastly, an empirical comparison of the algorithms examines the accuracy of the approach and the number of valuable item sets and consumers that have been discovered via the use of the algorithms. The LSTM-based approach is the most effective, as seen by its 98\% accuracy rate. It is especially useful for predicting future consumer purchases, identifying the most lucrative items, and properly managing inventories.