Software actors for continuous social media analysis

Social media analysis is rapidly becoming a widespread tool for various applications. Motivations for interest range from commercial marketing to the monitoring of social trends and political opinions. This work presents an actorbased platform, augmented with machine learning algorithms. To evaluate its possible application for the continuous analysis of social media streams, we also present a case study about sport. We argue that the actor paradigm allows modeling well this kind of scenario, and it can scale to more complex applications, where multiple actors can cooperate to realize structured analysis tools. Results also show the advantage of a system for continuous analysis, which can learn from new data and improve its internal classification model.