Abstract
The problems involved in the implementation of machine learning by minimizing sample-complexity were studied. These problems were solved in the field of computational geometry based on random walks. A connection between random walks and machine learning notations such as support vector machines was also investigated.
Original language | English |
---|---|
Pages (from-to) | 25-51 |
Number of pages | 27 |
Journal | Theoretical Computer Science |
Volume | 284 |
Issue number | 1 |
DOIs | |
State | Published - 6 Jul 2002 |
Event | Computing Learining Theory - Nordkirchen, Germany Duration: 29 Mar 1999 → 31 Mar 1999 |
Keywords
- Active learning
- Experimental design
- Labeled and unlabelled data
- Selective sampling
- Volume approximation