Product recommendation model for supermarket industry based on machine learning algorithms

Abstract

Recommendation activities automatically display products or content that might interest customers based on previous user activity. Recommendations help customers directly to identify the relevant items that they might otherwise not know. The product recommendation model determines which products are suggested to a consumer, depending on that consumer's shopping history. The main objective of this research was to develop a product recommendation model by considering the shopping history of consumers. The supermarket data used in the study contain customer details, transaction details, and product details. The product recommendation model was built using three machine learning techniques such as the Long Short-Term Memory algorithm, Convolutional Neural Network algorithm, and Collaborative Filtering algorithm. The obtained accuracies of the proposed model with respect to Collaborative Filtering, Long Short-Term Memory and Convolutional Neural Networks are 78%, 54% and 56% respectively. According to the accuracy values the Collaborative Filtering algorithm is more suitable to build the product recommendation model than the Long Short-Term Memory algorithm or Convolutional Neural network.