SC²S Colloquium - December 02, 2016

Date:	December 2, 2016
Room:	02.07.023
Time:	3:00 pm, s.t.

Volker Jacht: TBA

TBA

Johann Maier: Implementation and Evaluation of Online-Classification Methods based on Sparse Grids

Classification is a frequently occurring task in the field of data mining and knowledge discovery. Various approaches to tackle this class of problems have been developed in the past and successfully applied. However, standard implementations often struggle with large and high-dimensional datasets. Therefore, we focus on sparse grid based variants of popular classification models, such as algorithms based on stochastic gradient descent, support vector machines and density estimation. Additionally, different refinement approaches for sparse grids are introduced and applied to optimize the derived classifiers. Finally, we compare the classification models with respect to qualitative and computational performance on the basis of results obtained by applying the models to different popular datasets.