Hier finden Sie von Know-Center MitarbeiterInnen verfasste wissenschaftliche Publikationen


Rexha Andi, Mauro Dragoni, Marco Federici

ReUS: A Real-Time Unsupervised System For Monitoring Opinion Streams


Background An actual challenge within the sentiment analysis research area is the extraction of polarity values associated with specic aspects (or opinion targets) contained in user generated content. This task, called aspect-based sentiment analysis bring new challenges like the disambiguation of words' role within text and the inference of correct polarity values based on the domain in which a text occurs. The former requires strategies able to understand how each word is used in a specic context in order to annotate it as aspect or not. The latter need to be addressed with unsupervised solutions in order to make a system ecient for real-time tasks and at the same time exible in order to adopt it in any domain without requiring the training of sentiment models. Finally, the deployment of such system into real-world scenarios needs the development of usable solutions for accessing and analyzing data. Methods This paper presents the ReUS platform: a system integrating an unsupervised approach, based on open information extraction strategies, for performing real-time aspect-based sentiment analysis together with facilities supporting decision makers in the analysis and visualization of collected data. Results The ReUS platform has been validated from a quantitative and qualitative perspectives. First, the aspect extraction and polarity inference capabilities have been evaluated on three dataset used in likewise editions of SemEval. Second, a user group has been invited to judge the usability of the platform. Conclusion The developed platform demonstrated to be suitable for being used into real-world scenarios requiring (i) the capability of processing real-time opinionbased documents streams and (ii) the availability of usable facilities for analyzing and visualizing collected data. Examples of possible analysis and visualizations includes the presentation of lists ranking aspects by their importance of by their polarity values computed within the whole data repository. This kind of analysis enables, for instance, the discovery of product issues.
Kontakt Karriere

Hiermit erkläre ich ausdrücklich meine Einwilligung zum Einsatz und zur Speicherung von Cookies. Weiter Informationen finden sich unter Datenschutzerklärung

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.