SVM Optimization: Inverse Dependence on Training Set Size
Shai Shalev-Shwartz, Nathan Srebro
25th International Conference on Machine Learning (ICML), July 2008.
Last modified: Wed Aug 13 12:43:03 Central Daylight Time 2008
- The last term of equation (8) is copied incorrectly from equation
(4), and there is an error in the expression for the optimal lambda
following this equation. These errors are corrected in the current online version. The expression in
equation (9) and all subsequent derivations and conclusions are
correct as they appear in the conference proceedings.