Ekaterina Shabunina recently graduated under my supervision as a M.Sc. student of the Como Campus of Politecnico di Milano with a thesis titled “Approach based on CRF to Sentiment Classification of Twitter Streams related to Companies”. Thanks to the innovation of her work, she won the Grand Prize 2013 for the GSE Academic Award for Excellence.
The work is based on the assumption that information produced and shared on social networks is getting more and more interesting as a source for inferring trends and happenings in the real world. She applied sentiment classification of Twitter streams related to companies and calculated statistical correlation analysis with the companies’ securities prices variation. Tweets are labeled with a tailored classification model, which by itself exhibits solid performance indicators, and then are correlated to stock market values. The approach applies the Conditional Random Fields probabilistic model to company-related Twitter data streams and shows that there is high correlation between the classified results and the stock market values, even when adopting a very simple feature model. In particular, it presents a near-perfect adherence of accumulated number of net positive tweets versus the stock’s closing price with an ideal level of significance of the regression and a 97.56% explanatory capacity of the achieved fitted equation in the best case.
GSE (Guide Share Europe), a non-profit association of companies, organizations and individuals who are involved in Information and Communication Technology (ICT) solutions based on IBM architectures, established the GSE Academic Award for Excellence for students.
The project will be presented on the GSE Management Summit in Barcelona on October 14th, 2013. Here is a short interview with Ekaterina.
//www.youtube-nocookie.com/v/g-Y-XE286H0?hl=en_US&version=3

Further information about the awards is available on GSE website.
To keep updated on my activities you can subscribe to the RSS feed of my blog or follow my twitter account (@MarcoBrambi).