Evaluation of Explanation Interfaces in Recommender Systems

• Main Authors: Sergio Cleger-Tamayo, Marcela Sávia Picanço-Pessoa, Marcella Sarah Figueiras- de Farias
• Format: Article
• Language: English
• Published: Universidad Distrital Francisco José de Caldas 2017-05-01
• Series: Revista Científica
• Subjects: Filtrado Colaborativo; Métricas de Evaluación
• Online Access: http://revistas.udistrital.edu.co/ojs/index.php/revcie/article/view/11813

Explaining interfaces become a useful tool in systems that have a lot of content to evaluate by users. The different interfaces represent a help for the undecided users or those who consider systems as boxed black smart. These systems present recommendations to users based on different learning models. In this paper, we present the different objectives of the explanation interfaces and some of the criteria that you can evaluate, as well as a proposal of metrics to obtain results in the experiments. Finally, we showed the main results of a study with real users and their interaction with e-commerce systems. Among the main results, highlight the positive impact in relation to the time of interaction with the applications and acceptance of the recommendations received.