18 July 2017, 12:00 - 14:00
Data of MILANO EXPO2015 are crawled from the event’s twitter page before and after its opening. After data processing, a Latent Semantic Analysis is performed in order to build a semi-supervised ontology with the aim to perform a model and evaluate the global sentiment about the event. The sentiment polarity for the same users who have written before and after the EXPO2015 opening has been calculated and a gap analysis performed in order to evaluate the effect of each type of semantic argumentation on users’ opinions. The Generalized Cross Entropy (GCE) approach method is shown and applied to evaluate the regression coefficients effect. This GCE approach is applied for the first time to web data with the peculiarity to evaluate semantic class effect on the global sentiment.
Enrico Ciavolino is researcher and professor of Statistics and Psychometrics at Salento University and a member of scientific committee of the Human and Social Sciences PhD course. He has coordinated or participated in more than twenty regional, national and international projects. The activity of methodological research concerns the models of multivariate analysis and structural equation models based on parametric estimators (maximum likelihood), non-parametric (partial least squares - PLS) and semi-parametric (Generalized Maximum Entropy). The methodological research finds applications in the fields of psychology, economics, decision support systems, services evaluation and big data. He is the Executive Editor of the of the WoS & Scopus Journal "Electronic Journal of Applied Statistical Analysis". Moreover, he is the director of the spinoff sara-lab (Statistical Analysis for Research and Applications) - www.sara-lab.it