OpinionMining-ML and SentiTagger

OpinionMining-ML is the name of an XML formalism that I have defined together with Luigi Di Caro for tagging users' opinions on products and services.

(Robaldo and Di Caro, 2013) present OpinionMining-ML with respect to a corpus of 1000 comments about restaurants taken from www.2spaghi.it, one of the biggest web2.0 sites about Italian restaurants and pizzerias. However, although the paper presents a corpus about restaurants, OpinionMining-ML is independent from the application domain, i.e. it can be also used to annotate comments that do not concerns restaurants.

In OpinionMining-ML, users' attitude and sentiment are associated with the objects they are about (called facets). Facets are organized in an ontology, which is part of the formalism, and they are associated with the statements denoting appraisals, suggestions, observations, etc. about them.

OpinionMining-ML has been the basis of the SentiTagger system, a project that have been retained for funding from the Working Capital Accelerator 2014, a Telecom Italia initiative to support new startups and innovative research projects.

In the context of the initiative, I participated to ''WCAP 2014 Program'', the two-months intensive seminar series devoted to start-up foundation, pitch preparation, business and marketing analysis, etc., organized in Milan for WCAP 2014 winners. The system and the pitch has been developed during the two-month course. However, at the end I did not found a start-up (not yet :-)), in that I won the project ProLeMAS, and I moved to Luxembourg.

The SentiTagger system has been adapted to the legal domain and it is currently used to carry out Named Entity Recognition on legal documents, in the context of the projects ProLeMAS and MIREL.