Title: Semi-supervised Learning with Kernels (2009)
Scientific Advisor: Zoltán KÁSA
Ph.D Student: Zalán-Péter BODÓ
Short description: In the thesis we present novel techniques for semi-supervised classification by means of kernel methods. The three main parts of the thesis describe the proposed kernels: Wikipedia-based kernels for text categorization, hierarchical cluster kernels and reweighting kernels.
Bibliograpy in BibTeX [thesis.bib]
Abstract in English [abstract_en.pdf]
Abstract in Romanian [abstract_ro.pdf]
Presentation slides [thesis_pres.pdf]