SemanticSentimentAnalysis2014
From Odp
| Line 99: | Line 99: | ||
|deadline=2014/03/06 | |deadline=2014/03/06 | ||
|expired=no | |expired=no | ||
| - | |intime= | + | |intime=6 March, 2014 |
}} | }} | ||
Revision as of 10:30, 21 December 2013
Semantic Sentiment Analysis 2014
The First Workshop on Semantic Sentiment Analysis at ESWC2014, Hersonissos, Crete, 30 May 2014
Contents |
Semantic Web and Sentiment Analysis
As the Web rapidly evolves, people are becoming increasingly enthusiastic about interacting, sharing, and collaborating through social networks, online communities, blogs, wikis, and the like. In recent years, this collective intelligence has spread to many different areas, with particular focus on fields related to everyday life such as commerce, tourism, education, and health, causing the size of the social web to expand exponentially.
The opportunity to capture the sentiment of the general public about social events, political movements, company strategies, marketing campaigns, and product preferences has raised growing interest both within the scientific community, leading to many exciting open challenges, as well as in the business world, due to the remarkable benefits to be had from marketing prediction. However, the distillation of knowledge from such a large amount of unstructured information, is so difficult that hybridizing different methods from complementary disciplines facing similar challenges is a key activity.
Natural Language Processing techniques are widely used in order to process all sorts of texts – and especially those produced within social networks such as Twitter – so as to detect subjective statements and their sentiment [10]. This task is known as sentiment analysis, and overlaps with opinion mining. In currently available sentiment analysis systems such as [1, 2, 3], the user is typically prompted to insert a topic, on which the system returns a list of sentiment scores. However, some of the returned results are of limited use to the user because they are out of the domain the user is interested in. For example, one user could look for opinions about Berlusconi, and get several results related to Berlusconi in politics, business, criminal investigations, or football, or for different events Berlusconi is involved in. When moving from one context to another, or from one event to another, opinions can shift from positive to negative, or neutral. A sentiment analysis system should be able to differentiate its results accordingly, but most existing sentiment analysis systems provide an analysis that can be too coarse-grained, due to missed contextualization. In fact, several works [4, 5, 6, 7, 8, 11) show that considering topics and entities jointly with sentiment features improves the performance of sentiment analysis systems. Existing work (in social media in particular) also tends to focus on identifying sentiment of whole tweets rather than specific topics. Finally, the evolution and dynamics (how and where initiated, spread, flipped from + to -, etc) of sentiment over time are hardly considered. This quick overview already gives room to the role of a founded, semantic sentiment analysis, capable of distinguishing the different tasks and roles involved in features, methods, resources, and tools for SA, and opening to the rich background knowledge that exists, or can be generated and linked within the semantic web.
Mining opinions and sentiments from natural language is a difficult task as opinions and sentiments are often conveyed implicitly through latent semantics, which make approaches based on surface lexical hints ineffective in those cases. To this end, semantic sentiment analysis may go beyond word-level analysis of text, and provide novel approaches to opinion mining and sentiment analysis that allow a more efficient passage from (unstructured) textual information to (structured) machine-processable data, in potentially any domain. Semantic sentiment analysis may benefit from the use of web ontologies (e.g. linked data, multilingual linked lexica, linked vocabularies), or graph-based quasi-ontologies (e.g. ConceptNet, SenticNet, Nell, OIE, etc.) which enable the aggregation of conceptual and affective information associated with natural language opinions.
Combining NLP, lexical and semantic web approaches could provide us with the accuracy and strength of language processing techniques, as well as with the depth and breadth of semantic knowledge bases, through which also sentiments that are expressed in a subtle manner can be detected, as in the case of concepts that do not explicitly convey any emotion, but which are implicitly linked to other concepts that do so. What is challenging is the way techniques can be used and combined to yield highly performant systems. With semantics, we can expand the current state of the art in sentiment analysis to be able to track, correlate, and compare sentiment of specific entities or group of related entities over time and across different contexts.
In brief, Sentiment Analysis techniques mainly count on NLP and lexical resources, and, as seen above, the current state of art targets lightweight analyses of sentences or documents. Besides the companion field of opinion mining, another related field that can benefit from (and contribute to) this cross-disciplinary work is sentic computing (9), which also aims at abridging word-level natural language data and concept-level opinions.
Therefore, this workshop focuses on the introduction, presentation, and discussion of novel approaches to semantic sentiment analysis. Special focus will be given to semantic methods, models, and tools that exploit common-sense knowledge bases to perform multi- or open-domain sentiment analysis. The intended audience of the workshop includes researchers from academia and industry as well as professionals and industrial practitioners to discuss and exchange positions on new hybrid techniques, which use semantics for sentiment analysis. The expected number of participants ranges between 20 and 40.
References
- [1] Sentiment 140. http://www.sentiment140.com/.
- [2] Opinion Crawl. http://opinioncrawl.com/.
- [3] Social Mention. http://www.socialmention.com/.
- [4] Chenghua Lin, Yulan He, Richard Everson, and Stefan Ruger. Weakly supervised joint sentiment-topic detection from text. IEEE Transactions on Knowledge and Data Engineering, 24(6):1134–1145, 2012.
- [5] Keke Cai, Scott Spangler, Ying Chen, and Li Zhang. Leveraging sentiment analysis for topic detection. Web Intelli. and Agent Sys., 8(3):291–302, August 2010.
- [6] Ivan Titov and Ryan McDonald. Modeling online reviews with multi- grain topic models. In Proceedings of the 17th international conference on World Wide Web, WWW ’08, pages 111–120, New York, NY, USA, 2008. ACM.
- [7] Richard Johansson and Alessandro Moschitti. Relational features in fine-grained opinion analysis. Computational Linguistics, 39(3):473–509, 2013.
- [8] Aldo Gangemi, Valentina Presutti, Diego Reforgiato Recupero. Frame-based detection of opinion holders and topics: a model and a tool. IEEE Computational Intelligence, 9(1), 2014.
- [9] E. Cambria and A. Hussain. Sentic computing: Techniques, tools, and applications. Dordrecht, Netherlands: Springer, ISBN: 978-94-007- 5069-2, 2012.
- [10] Bing Liu (2012). Sentiment Analysis and Opinion Mining. Synthesis Lectures on Human Language Technologies. Morgan & Claypool Publishers.
- [11] Saif, Hassan; He, Yulan and Alani, Harith (2012). Semantic sentiment analysis of Twitter. In: The 11th International Semantic Web Conference (ISWC 2012), 11-15 November 2012, Boston, MA, USA.
Workshop format
This half-day workshop will consist in an invited lecture by a prominent figure in a related field (e.g. sentic computing, SA-related linked data, etc.), followed by presentations of either full or position papers, demos, and general discussion. A session will be dedicated to a summary of the Concept-based Sentiment Analysis Challenge that is separately organized in the same ESWC2014 Conference.
Related Workshops and Conferences This is a new workshop, the first time in a Semantic Web-related event (whereas workshops on pure NLP-based sentiment analysis are now regular in ACL and IJCAI conferences).
Important dates
- Submission deadline: March 6, 2014 - Notifications: April 1, 2014 - Camera ready version: April 15, 2014
Workshop organizers/chairs
Workshop chairs:
Aldo Gangemi (contact person), U. Paris Nord France/ISTC-CNR Rome Italy, aldo.gangemi@lipn.univ-paris13.fr, http://istc.cnr.it/people/aldo-gangemi
Harith Alani, KMI-OU Milton Keynes UK, h.alani@open.ac.uk, http://people.kmi.open.ac.uk/harith/
Erik Cambria, NUS Singapore, cambria@nus.edu.sg, http://sentic.net
Diego Reforgiato, ISTC-CNR Catania Italy, diego.reforgiato@istc.cnr.it, http://www.istc.cnr.it/people/diego-reforgiato-recupero
Program Committee (to be completed):
- Paolo Rosso, U.P. Valencia
- Harith Alani, KMI-OU Milton Keynes
- Paul Buitelaar, DERI Galway
- Bebo White, U. Stanford
- Fabio Ciravegna, U. Sheffield
- Eneko Agirre, U. Basque Country
- Davide Buscaldi, U. Paris Nord
- Carlo Strapparava, FBK Trento
- J. Fernando Sánchez-Rada, U.P. Madrid
- Björn Schuller, Imperial College London
- Catherine Havasi, MIT Boston
- V. S. Subrahmanian, U. of Maryland
- William H. Hsu, U. Kansas State
|
no |
