System for Entropy-Based Product Expiration Alerts for Customers with Serious Issues

There is a significant problem with selling things that have expired, particularly among customers who purchase the products from supermarkets or stores. In order to prevent this problem from occurring, it is possible to create a web application that will notify the proprietor of the products that are going to expire. This paper presents three proposed approaches for imbalanced learning in order to handle imbalanced data, which consists of a large set of uploaded products with different expiration dates. The first approach is the Entropy-based Over Sampling approach (EOS), the second approach is the Entropy-based Under Sampling approach (EUS), and the third approach is the Entropy-based Hybrid Sampling approach (EHS), which combines oversampling and undersampling simultaneously as a single approach. When taking into consideration the divisions of information on the product's expiration date, these three methods contribute to the classification of the imbalanced classes, which is known as the Entropy-based Imbalance Degree (EID). Last but not least, we arrange all of the products in accordance with their expiration dates, with the most recent ones being placed at the top. As a result, notifications can be issued on a regular basis to all of the products that have been submitted and will soon expire.