Aplicación de técnicas de aprendizaje automático para la generación de recomendaciones en base a los intereses de los usuarios en las redes sociales

Abstract

Abstract: Due to the growth of information in social networks, an information overload has been created, making it difficult for users to find information according to their interests, so the use of recommendation algorithms allows us to find information, products or services that are of interest of users. This thesis consists of the application of machine learning techniques for the generation of recommendations based on the interests of users in social networks, where the items to be recommended are the hashtag of the social network twitter. The recommender system is based on collaborative filtering using the K-NN algorithm to recommend hashtags to Twitter users. To measure the precision of the recommender system, the algorithm is evaluated from two points of view: evaluation of the prediction in which the Mean Square Error is used and the precision of the recommendations in which metrics such as Precision, Recall and F are used.Measure. Finally, a web application is developed where the recommendations of items and graphs of results are presented to users.