Promoting culture in Peru through a graph-based recommender system

Culture and its dissemination are an important aspect of the identity of a community or country and its population. Peru has a wide diversity of cultural manifestations, but its population tends to overlook them and governmental institutions do not provide efficient support for their promotion and development, limiting their scope and citizen participation. This paper proposes a solution model for a web application that uses recommendation systems to promote culture in Peru. There are many similar proposals that make use of this technology, the most outstanding of them being due to the efficency of the graph neural networks implemented in them. The solution consists of five components and requires web scraping tools and APis to keep the collected data updated. The model presented is a convolutional graph network (GCN) that, when trained with the Amazon Products Electronics dataset, achieves favorable results in terms of recommendation generation. The Machine Learning model designed achieves the mean absolute error less than 1.6 at the level of the experiments performed and it is integrated with three other components that make up our overall proposal at the design level. There are three items that can be developed in future work in order to complete its implementation and improve the efficiency of the components.