MLDM 2007: A brief overview

Here is the last post about the MLDM 2007 conference in Leipzig. As mentioned in an earlier post, several different topics were covered in this meeting. In my opinion, there were no trendy topics such as SVM, ANN, GA, etc. that flood other methods. Below, you can find a list of interesting papers.

• “Kernel MDL to Determine the Number of Clusters” by Kyrgyzov et al. where they combine Minimum Description Length (MDL) with Kernel K-means to estimate the number of clusters.

• “A Case-Based Approach to Anomaly Intrusion Detection” by Micarelli et al., a work that combines Case Based Reasoning (CBR) and clustering for intrusion detection.
• “Comparing state-of-the-art collaborative filtering systems” by Candillier et al. (selected for best paper award) presents different collaborative filtering methods to point out their advantages/drawbacks and propose some basic options to consider when using a particular technique.

The title of the best paper award was: “Affine Feature Extraction: A Generalization of the Fukunaga-Koontz Transformation“, written by Wenbo Cao and Robert Haralick. To be noted that R. Haralick gave an additional (very strange) presentation about data mining and religion. Of course, as you can imagine, this presentation was subject to a lot of discussion and controversy. Maybe some risky subjects should not be presented…

Comments