Uma investigação em postagens relacionadas ao uso contendo emojis

Abstract

The usability and user experience criteria make the systems more suitable to use however in order to ensure the presence of such criteria is necessary to evaluate them. Mendes (2015) has proposed a model and methodology to evaluate the interaction in systems from the textual language of the user, called MALTU. Recent studies point to the growth in the use of emojis within the digital context, since they are an alternative to express feelings and non-verbal language. This article investigates the impacts that emojis cause in a textual evaluation based on posts made on social networks and evaluation sites. For this, an experiment was conducted that consisted of extracting posts from social systems and app review stores to verify which of these presented the greatest amount of posts with emojis, where it was observed that the Play Store apps best met this condition. With the environment defined 1392 posts from various apps were extracted. The posts were then classified into AcrshortPRUs and Non-AcrshortPRUs, and into facets of usability and user experience. It was investigated the relationship between the emojis in the posts and the facets present in them. Based on the investigation, an algorithm was built to automatically classify the posts according to the emojis presents in them. Where it was obtained for UX - satisfaction and frustration, an accuracy of 0.878%, recall of 0.866% and accuracy of 0.889%, for Usability - efficiency and satisfaction, the precision was 0.784%, recall of 0.654% and accuracy of 0.832%.