A Simple Recommendation Model Using the Item’s Global Popularity and Frequency-Based User Preference

Next-basket recommendation (NBR) is one of the emerging methods in the recommendations system and it has many real-world applications, like where you want to predict the set of items in the next transaction based on its previous continuous transactions. In the literature, many algorithms are proposed for the next-basket recommendation, but very little attention is given to the frequency information in making the recommendation model. In this paper, we presented a simple recommendation model, and in this model, the user representation is done based on the frequent purchasing information, and finally, when the recommendation is made, it considers the user representation and items global support. We made experiments on publicly available three real-time datasets and compared them with two of the existing recommendation algorithms using recall as a measure, and our algorithm outperforms the existing two models. The proposed algorithm is simple in terms of its complexity and the number of operations involved.