Wednesday, March 27, 2013

Release 2.0.33

today, a new version of BibSonomy has been released. Amongst many internal improvements we are happy to announce the following changes:
  • sorting of publication posts by their BibTeX field numbers via url parameter sortPage=number
  • new webservice for querying related tags
  • JCBScraper
  • Improved rest-client-api including more robust charset handling and document uploads
Happy tagging!
Jens

Monday, March 4, 2013

20DC13 - The ECML PKDD Discovery Challenge 2013 on Recommending Given Names

This years ECML PKDD Discovery Challenge is organized by several members of the BibSonomy Team. After the successful challenges on Tag Recommendation in 2008 and 2009 this years challenge focuses on a new domain, that can profit from recommender systems:

All over the world, future parents are facing the task of finding a suitable given name for their children. This choice is influenced by different factors, such as the social context, language, cultural background and especially personal taste. Most books, websites or mobile apps aid parents only by offering alphabetically ordered list of names.


CAN YOU DO BETTER?


In the challenge, participants will explore different recommender strategies and their ability to predict names, that users will find interesting. Hereby, the recommendation of names provides a serious and versatile task, that can be approached from several different perspectives:
  • The influence of personal taste, language and geographic user location, for example, presents an opportunity for context-aware recommendations. 
  • The influence of the social context of a user brings in the social network analysis community as another source to improve the quality of the recommender with its methods. 
  • The etymological roots as well as connotations of names that have arisen over time (e.g., through popular cultural, political or religious bearers of a given name) suggest that methods based on inherent semantics and “content” may also yield successful recommendation approaches. 
  • Methods from the field of sentiment analysis may help to distinguish names which are popular because users like those names from those which are popular because users are interested in their context.
For the challenge, participants can make use of a new data set from the name-search engine nameling where thousands of users are looking up names to find similar or related names to name their children.

Results from an online and an offline experiment will be presented during a workshop, that is held in conjunction with this year's
European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases - the premier European machine learning and data mining conference.

More info on the challenge, on the registration, the data set and the workshop can be found on the official Discovery Challenge Website.

The registration is open!

Your 20DC13 Team
Stephan, Andreas, Robert, Folke & Juergen

Popular Posts