Announcement: Master Thesis Opportunity on Classification of Rare Classes and Rare Values
From the data science KMD lab: Master Thesis on Classification for Rare Classes and Rare Values
Subject area: Investigating the potential of GraphNNs or multi-layer networks for classification of objects when the classes are rare and some of the predictive features are also rare
Your profile: familiarity with classification in skewed distributions and/or with cost-aware classification and /or cost aware feature selection, proficiency in Pytorch or Tensorflow 2.0
Desiderata: graph-based learning, learning on small data, statistics
How to apply: per mail at myra@ovgu.de; please provide evidence on how your background agrees with the opening