JMySVMLearner - bias calculation

ilaria_gori

Please could you help me to understand how the bias (offset) is calculated in the case of classification of 2D vectors with JMySVMLearner?
The vector w is clearly calculated as linear combination of the training vectors (coefficients equal to the Lagrange multipliers times the target), but bias =?
thanks!!

ilaria

land

Hi,
I suggest, you should take a look in the literature about the JMySVM. Probably it is mentioned somewhere. You could start searching here:

T. Joachims, Making large-scale support vector machine learning
practical, Advances in Kernel Methods: Support Vector Machines,
MIT Press, Cambridge, MA, 1998.

Greetings,
  Sebastian

ilaria_gori

thank you, I found it!

ilaria