Recommender System for Coupon Discount of E-commerce Applications

Recommender systems have recently been integrated into many fields to introduce numerous appealing products and help with consumer decisions in item selections. Applying recommender systems to coupon discount applications could improve consumers accessibility to coupons of their interest, leading to increased product sales related to coupons. The research aims to create a recommender system model for coupon suggestions based on customer data interactions. The proposed recommender model is a hybrid model that combines matrix factorization and association rule mining. While matrix factorization utilises customer-item interactions with customer transactions history and information as the implicit feedback, the association rule uses the Apriori algorithm to find frequent itemsets from customer discount transactions to create the weight of each item. As a result, the proposed hybrid recommender model provides overall performance higher than the baseline models and other benchmark models. In addition, the proposed model has a precision rank increase of nearly 20% in the same dataset compared to that of the baseline models.