Computational challenges for sentiment analysis in life sciences

Data and connectivity between users form the core of social networks. Every status, post, friendship, tweet, re-tweet, tag or image generates a massive amount of structured and unstructured data. Deriving meaning from this data and, in particular, extracting behavior and emotions of individual users, as well as of user communities, is the goal of sentiment analysis and affective computing and represents a significant challenge. Social networks also represent a potentially infinite source of applications for both research and commercial purposes and are adaptable to many different areas, including life science. Nevertheless, collecting, sharing, storing and analyzing social networks data pose several challenges to computer scientists, such as the management of highly unstructured data, big data, and the need for real-time computation. In this paper we give a brief overview of some concrete examples of applying sentiment analysis to social networks for healthcare purposes, we present the current type of tools existing for sentiment analysis, and summarize the challenges involved in this process focusing on the role of high performance computing.