Explanative Bewertungsanalyse für die Saarländische Hotellerie

The main goal of the project is the creation of a web portal that collects, analyzes and evaluates guest comments, ratings and reviews of hotels and that makes the information accesible in comprehensive form. The portal is based on a search engine that systematically collects guest reviews from the web, not just from the commercial hotel review and hotel booking services but also from the increasing number of other publication forms in the Web-2.0, such as blogs, forums, social networks, Twitter, etc. Often hotel operators are not aware of the large number of sites that contain reports and comments about their hotels. The main focus of the project is not so much on the ratings but more on discovering the reasons for positive or negative comments. Such a service can be useful for quality monitoring as well as help in planning, e.g. about changing demands and expectations of guests, implicit in their reviews. Because of the large amount of possible sources and publication forms in the web it is very difficult for the hotel management to get a comprehensive overview.

The analysis is based on opinion mining and sentiment detection techniques, a field of text mining that has gained large research and application interest because of the increasing importance of "user generated content" on the web. Opinion ming poses a number of special challenges in contrat to fact oriented information extraction and retrieval. With respect to its interpretation functionalities, BESAHOT goes beyond basic functionalities of opinion mining as classification and extraction of ratings from texts. To enable later extensions of the service to other areas in the tourism domain, such as gastronomy, places of interest, etc., also the automatic acquistion of domain knowledge are investigated.

Funded by:Land (Saarland)
Project Manager:Walter Kasper (Walter.Kasper@dfki.de)
Contact:Walter Kasper (Walter.Kasper@dfki.de)
Duration: 1.1.2010 - 31.12.2011