A Virtual Wallet Product Recommender System Based on Collaborative Filtering

Nowadays, there are several options when it comes to making use of products that facilitate financial services to people through virtual wallets. A recommender system quickly provides customers with what they are looking for and helps discover new products that they like. In this paper, a recommender system is proposed that can be customized according to the variables implemented by Movii, a company in the Colombian FinTech sector, taking as input transaction records that indicate the frequency of use of each product, which can be understood as ratings of these products. To determine the model that will implement the recommender system that will be deployed, different models are evaluated, such as techniques based on collaborative filtering. In our evaluation, we found that the model that recommends the most popular products is the one that offers the best performance in recommending a product to users. Thus, it is possible to generate some estimated recommendations on the services available by the company, involving users who consume the available services.