Regresar

An adaptive hybrid recommendation system architecture

Abstract:

There is a wide variety of recommendation systems, mainly based on content and collaborative filters. Each one has advantages and disadvantages, so combining them in a hybrid system is a promising approach, and to do that, there are multiple proposals in the literature. Most hybrid approaches are static, so it is of current interest to explore their ability to follow the characteristics, both domain and available data. This work proposes an adaptive hybrid recommendation architecture, which follows the dynamic behavior of the environment through the use of metrics (meta-characteristics), from which the hybrid configuration for the recommendation is determined. In the experiments in a case study, its adaptive capacities are shown, exemplifying its operation, and showing its flexibility to be implemented in various ways and in multiple contexts.