BibSonomy publications for /author/ObermayerMon Feb 06 00:00:00 CET 2012The New Review of Hypermedia and Multimedia3267-294Tripartite community structure in social bookmarking data.172011dblp Fri Dec 16 00:00:00 CET 2011IEEE Transactions on Neural Networks2390-398Soft nearest prototype classification.142003dblp Mon Dec 05 00:00:00 CET 2011CoRRGraph Quantizationabs/1001.09212010dblp Mon Dec 05 00:00:00 CET 2011CoRRAccelerating Competitive Learning Graph Quantizationabs/1001.09272010dblp Mon Dec 05 00:00:00 CET 2011CoRRExtending Bron Kerbosch for Solving the Maximum Weight Clique Problemabs/1101.12662011dblp Mon Dec 05 00:00:00 CET 2011CoRRA Necessary and Sufficient Condition for Graph Matching Being Equivalent to the Maximum Weight Clique Problemabs/0912.45842009dblp Mon Dec 05 00:00:00 CET 2011CoRRElkan's k-Means for Graphsabs/0912.45982009dblp Mon Dec 05 00:00:00 CET 2011CoRRA Unified Framework for Risk-sensitive Markov Decision Processes with Finite State and Action Spacesabs/1110.63172011dblp Mon Dec 05 00:00:00 CET 2011CoRRProbabilistic prototype models for attributed graphsabs/1109.47442011dblp Wed Nov 30 07:55:16 CET 2011New York, NY, USASIGWEB Newsl.September4:1--4:10Hyperincident connected components of tagging networks2009bibsonomy-dataset networks spam tagging Data created by social bookmarking systems can be described as 3-partite 3-uniform hypergraphs connecting documents, users, and tags (tagging networks), such that the toolbox of complex network analysis can be applied to examine their properties. One of the most basic tools, the analysis of connected components, however cannot be applied meaningfully: Tagging networks tend to be almost entirely connected. We therefore propose a generalization of connected components, <i>m</i>-hyperincident connected components. We show that decomposing tagging networks into 2-hyperincident connected components yields a characteristic component distribution with a salient giant component that can be found across various datasets. This pattern changes if the underlying formation process changes, for example, if the hypergraph is constructed from search logs, or if the tagging data is contaminated by spam: It turns out that the second- to 129th largest components of the spam-labeled Bibsonomy dataset are inhabited exclusively by spam users. Based on these findings, we propose and unsupervised method for spam detection.Thu Nov 24 00:00:00 CET 2011Web Intelligenceconf/webi/2007607-613Distance Measures in Query Space: How Strongly to Use Feedback From Past Queries.2007dblp Thu Nov 17 00:00:00 CET 2011Neural Computation61472-1510Support Vector Machines for Dyadic Data.182006dblp Thu Nov 17 00:00:00 CET 2011Neural Computation5959-970The Joint Development of Orientation and Ocular Dominance: Role of Constraints.91997dblp Thu Nov 17 00:00:00 CET 2011Neural Computation71589-1604Soft Learning Vector Quantization.152003dblp Thu Nov 17 00:00:00 CET 2011Neural Computation3425-468Models of orientation and ocular dominance columns in the visual cortex: a critical comparison.71995dblp Thu Nov 17 00:00:00 CET 2011Neural Computation3555-575Singularities in Primate Orientation Maps.91997dblp Thu Nov 17 00:00:00 CET 2011Neural Computation112573-2595An Analysis of Orientation and Ocular Dominance Patterns in the Visual Cortex of Cats and Ferrets.122000dblp Thu Nov 17 00:00:00 CET 2011Neural Computation1139-155A Stochastic Self-Organizing Map for Proximity Data.111999dblp Thu Nov 10 00:00:00 CET 2011NIPSconf/nips/20091383-1391Correlation Coefficients are Insufficient for Analyzing Spike Count Dependencies.2009dblp Fri Sep 09 00:00:00 CEST 2011Parallel Computing3381-404Large-scale simulations of self-organizing neural networks on parallel computers: application to biological modelling.141990dblp