This study proposes a new collaborative filtering model that integrates Restricted Boltzmann Machines. The proposed two-stage model is applied to household-level supermarket purchase data. Results show that our model fits the data better and outperforms existing collaborative filtering methods in predicting shopping patterns. The proposed model also improves interpretations of market complexity and common causes of coincidence associated with customers’ multi-category purchases.

In 2005, Asia Pacific had the largest percentage of recreational shoppers in the world, while at the same time, Indonesia achieved the highest percentage for recreational shopping in Asia Pacific. (AC Nielsen, 2006). Shopping malls in Indonesia are growing rapidly and have currently increased in size, theme and location, and in the number of different types of shops. Given the rapid growth of malls, mall managers and retailers face the challenge to develop new strategies to survive in this highly competitive market. To gain greater insight into the relationship between shoppers and retailers, our aim is to explore shoppers’ behavior during their visit to the mall, specifically with respect to the type of shops visited. We will identify different shopping patterns and their characteristics based on 171 unobtrusive observations of shoppers using hierarchical clustering analysis. The results indicate that four shopping styles can be identified: fashion shoppers, basic shoppers, social shoppers, and entertainment seekers. Our results enhance an understanding of shopping styles. Mall managers and retailers could use this knowledge to decide the appropriate retail mix in the mall.