Towards a Strategy for Recognition of Collective Emotions on Social Networking Sites

Since the emergence of emotional theories and models, which explain individuals feelings and their emotional processes, diverse research areas have shown interest in studying these ideas in order to obtain relevant information about behavior, habits and preferences of people. However, there are some limitations on emotion recognition that have forced specialists to search ways to achieve it on particular cases. This article treats collective emotions recognition case focusing on social networking sites applying a particular strategy, as follow: Firstly, state of art investigation regard emotions representation models in individual and collectives. In addition, possible solutions are provided by computing areas regarding collective emotions problems. Secondly, a collective emotion strategy was designed where it was retrieved a collection of data from Twitter, in which some cleaning and processing steps were applied, in order to keep the expression as purest. Afterward, the collective emotion tagging step arrived, whither based on consensus theory approach, the majority tagged-feelings were grouped and recognized as collective emotions. Finally, prediction step was executed and resided on modeling collective data, wherein one part was supplied into the Machine Learning during training and the other one was served to test the machine accuracy. Thirdly, An evaluation was set to check the fit of the collective recognition strategy, where results obtained allow to place the proposed work in the right path as consequence of minor differences observed, that indicate higher precision according to the distances measures used during the study development.