Planteamiento de un Sistema Recomendador Basado en Contenido y Colaborativo para la Elección de Libros

Abstract

The project presented in this book contains the approach of a recommendation system based on content (recommendation with the attributes of the books that the user has already read) and collaborative (recommendation with the attributes of users with profiles similar to those of the user who requests the recommendation), by means of the approach of an algorithm based on the weights of the attributes; It is suggested that the weight of the attributes for which the recommendation of books is made, are important in order to generate a better result. Start with the approach of equations that distribute the weight of 100% between the collaborative recommendation and the content-based one; after this the weight of each type of recommendation is redistributed among the attributes that each one possesses. After carrying out the prototype of the system, 20 users were searched for the initial tests of the system, to which 5 recommendations were made to each one, the first is done by giving each type of recommendation 50% of weight and also divided in equal parts to the attributes of each recommendation. The second recommendation was made with the weights obtained in a survey carried out initially to 100 users of which 60 answered and with which initial pesos were generated. The other three recommendations were made with the feedback obtained from the previous recommendation. To be able to determine the results, surveys were carried out with the 20 users, with which it was concluded that the weight of each attribute varies with each recommendation and that these are important for the user to generate the recommendation that best suits your taste.