Published Paper Details:

Lot of people around the world are affected by one or more mental disorders that involve in their thinking and behavior. A timely finding of these problems is challenging but important, since it could open the possibility to offer abid to people before the illness gets worst. Another Approach is to realize this is to monitor how people communicate themselves, that is, what and how they write, or even a step further, what emotions they accomplish in their social media communications. In this study, we estimate two computational representations that aim to design the presence and changes of the emotions communicated by social media users. In our approach we use two recent public data sets for two important mental disorders: Depression and Anorexia. The acquired results suggest that the presence and variability of emotions, apprehend by the proposed representations, allow to highlight main information about social media users suffering from depression or anorexia. Furthermore, the combination of both representations can enhance the performance, equaling the best reported approach for depression.