Classificação automática de facetas de usabilidade e experiência de usuário em postagens relacionadas ao uso.

Abstract

Users‘ feedbacks are very important in Human-Computer Interaction, for we to have knowledge regarding to a system’s Usability and User Experience (UUX). Recently, narratives in the form of text, expressed spontaneously in social systems, app stores or review sites, named Postings Related to Use (PRU), has been showed themselves as precious sources of information about the quality of use. Identifying UUX facets (eg. satisfaction, efficiency) in text is, in essence, a problem of classification into categories, and that has been mainly held in a manual way. Such kind of task is one of the many whose state-of-the-art has been currently overcamed by a class of machine learning techniques called Deep Learning. This research aims to apply these state-of-the-art algorithms, besides more traditional learning techniques, in order to classify UUX goals on PRUs.