Anestis Gkanogiannis, PhD Candidate of Informatics Department under the supervision of Professor Theodore Kalamboukis, is the winner of the ECML - PKDD Discovery Challenge 2008

Anestis Gkanogiannis, PhD Candidate of Informatics Department under the supervision of Professor Theodore Kalamboukis, is the winner of the ECML - PKDD Discovery Challenge 2008, for detecting spam entries in Social Bookmarking Systems.

The European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML – PKDD), is the top European, and among the top internationally, conference on Machine Learning. Every year, under this conference, a Discovery Challenge is organized in order novel and older Machine Learning and Knowledge Discovery methods to be tested and presented.

This year’s Discovery Challenge deals with the application of such methods in Social Bookmarking Systems of the Internet. Participants had to develop techniques for detecting spam entries in such systems.

Anestis Gkanogiannis participating in the Discovery Challenge took the first place, by solving the problem of the spam detection as a Linear Binary Classification problem. He used a modification of the novel method he presented in the international conference SIGIR 2008 in Singapore on July 2008, titled "An Algorithm for Text Categorization". It is noted that the SIGIR conference (ACM Special Interest Group on Information Retrieval) is the top conference in the field of Information Retrieval.

He presented this method, along with its application to the data of the Discovery Challenge, on the 15th of September 2008 in the ECML - PKDD 2008 Conference in Antwerp, Belgium. More information about the Discovery Challenge can be found here.