Weews: um sistema recomendador de notícias relacionadas apresentadas em web widget

Abstract

Internet facilities as well as its scope, caused a cultural change in the way of reading news. From printed paper to the portals, thousands of news are posted daily and are available just steps. This massive amount of information overwhelm users, who just want to find something interesting to read. This challenge, of delivering only is helpful to users, encouraged the portals to use some form of recommendation through a filtering agent who select what is useful to the user. However, many portals choose to recommendations made by hand, which are more conducive to failure and require constant intervention to keep the content always updated. To introduce an alternative to current forms of recommendation, this work describes the Weews a news recommender system. The Weews aims recommend news based on the similarities between them and the relevance of the extracted terms of the news texts in a given period. For verifying the effectiveness ofWeews, experiments and empirical analysis of the results generated were performed. Through these analyzes, we can validate the Weews recommendation algorithm, proving it to be an efficient alternative to current forms of recommendation.