Machine Learning and Natural Language Processing Lab

Relational Cond... |--| Asynchrone Rela... |--| Flexible String... |--| Lernen von Inte... |--| Speeding Up Log... |--| Relational Fish... |--| Development of ... |--| Development of ... |--| Collective Clas... |--| Learning Locali... |--| Using Equation ... |--| Classification ... |--| Induktive Daten... |--| Constraint-Base... |--| Magic Bayes-Bal... |--| Logistic Model ... |--| Clustering unte... |--| Two-Level Class... |--| Machine Learnin... |--| Erweiterung des... |--| Bayes'sche-logi... |--|

Master Thesis

Relational Fisher Kernels

Uwe Dick, 2005

One approach to improve the accuracy of classifications based on generative models is to combine them with successful discriminative algorithms. Fisher Kernels were developed to combine generative models with a currently very popular class of learning algorithms, kernel methods. So far, however, Fisher kernels have mainly been considered for examples over flat alphabets. Here we introduce and investigate Relational Fisher Kernels which allow for exploiting the discriminative power of Support Vector Machines in relational probabilistic models. Experiments using this relational kernel show significant improvement over results achieved by merely utilizing the classification performance of the generative model via the plug-in estimate. Moreover, our results indicate for significant improvements when applying the Relational Fisher Kernel on collective settings even when the plug-in estimate does not gain accuracy.