A Framework for Analyzing Big Social Data and Modelling Emotions in Social Media
Isidoros Perikos, Ioannis Hatzilygeroudis · 2018
Social media constitute a rich source of information. The analysis of the big volume of user-generated content in social media can provide meaningful information for gleaning people's emotions and deeper understanding public attitude and mood. In this work, we present a generic framework for the analysis of big social data and the modeling of public emotions and mood. The framework consists of two main parts. Initially an ensemble classifier schema, which combines a generic domain independent, knowledge-based tool with machine learning methods, is used to recognize emotional content in user generated social data. Then, a graph based method is utilized to model the emotional level of a topic based on the emotions recognized and then it creates the topic's emotional graph visualizing public emotions and mood on the topic. The results from the case study are quite encouraging.