Sistemas de recomendação com auxílio de processamento de imagens de redes sociais

Abstract

Recommendation Systems are of great relevance for systems that want to deliver more value to the user. They provide personalized results by the use of particular algorithms. Among the various recommendation techniques the Image Recommendation Systems have become extremely useful since the visual aspect of the items can influence the user’s decision. A recommendation model can be designed to benefit from Social Network data and their interconnection aspects to improve their results. The experiments carried out demonstrate an approach capable of using images from social networks in the construction of a recommendation model and, among other results, finding alternatives to mitigate the cold-start problem.