Dynamic Rule-Based Approach for Shelf Placement Optimization Using Apriori Algorithm

Abstract

In the current era of customers, retail industries are transforming themselves into the customer-centric business models, where predetermination of customer needs and serving according to that may increase the reliability of business and enhance the profit. With the advent of new technologies, the retail industries need to be updated and one step ahead from the customers of new generation, whose demands are increasing based on continually changing trends. Conventional machine learning algorithms enable such industries to determine the needs and interests of their customers and make them able to attain the maximum profit from their businesses and look toward the new directions to expand the business. Correct implementation of these algorithms and techniques helps in anticipating the retail needs of the customers. Shelf placement plays a vital role in sale of product and customer engagement. A well-organized and associated placement of products on shelves increases the sale and makes customer comfortable with the shopping. A well-known technique, association rule is implemented in this paper using Apriori algorithm in Python, to identify the most common item sets sold together, which further helps in figuring out the more beneficial shelf placement for better customer engagement. It was found that items having more confidence rate are more likely to be purchased together and should be placed together for profit maximization. Our research produces a maximum confidence of 30% which is the result of our novel work.