| Popis: |
In today’s business world, where competition is increasing with the increase in product and service diversity, companies are in search of smart methods to bring the right products to their customers. Product bundle generation, in which products that are likely to be purchased together, are collected and presented is one of these methods. In our study, a product bundle production engine is developed based on the sales data of a pioneering chain in the fast-food industry. In the study, which is a component of the product recommendation system, data patterns are learned by extracting product basket statistics and using a customized Gaussian Mixture Model according to the targets. Suitable product bundles for the targets are produced with the depth-first search algorithm, which uses mixture models as a prioritization tool. The study also produces output by considering weighted targets specific to certain customer groups, general purchasing preferences and sales periods. Although the developed model is independent of the sector, it allows for expansion according to business needs, as it consists of discrete modules. |
