Title :
Multi-genre summarization: Approach, potentials and challenges
Author :
Elena Lloret;Ester Boldrini
Author_Institution :
University of Alicante, Dept. of Software and Computing Engineering, Natural Language Processing Research Group, Ctra San Vicente del Raspeig s/n, San Vicente del Raspeig, 03690, Alicante, Spain
Abstract :
In our current context, the Social Web is fostering the creation of a wide range of new textual genres (e.g., review sites, blogs, fora, social networks or microblogs). These new textual genres are becoming more and more important for both users and also for the business sector. Their effective analysis and exploitation may offer an added-value for the decision-making processes of customers but also from the company perspective to monitor possible competitors. Given these premises, in this paper, we outline and discuss the importance of a novel type of text summarization, i.e., multi-genre summarization, by means of reviewing the most relevant literature aiming at generating automatic summaries from new textual genres, specifically, reviews and microblogs, since they are growing in popularity. We report the main findings in the literature, and, different from what is being done at the moment regarding the production of automatic summaries, we stress the need for addressing different genres within the same approach, and we describe the challenges involved in the creation of this new type of summaries.
Keywords :
"Companies","Feature extraction","Context","Twitter","Real-time systems","Media","Pragmatics"
Conference_Titel :
eChallenges e-2015 Conference, 2015
Electronic_ISBN :
2166-1677
DOI :
10.1109/eCHALLENGES.2015.7440970