Regularization error to prevent overfitting. Our goal is

Regularization addresses overfitting problem by penalizing complex
learning algorithm – it selects one complex hypothesis class and adds regularizer function  to empirical error to prevent overfitting.

 

 

Our goal is to choose best learning algorithm among many possible
hypothesis space – the regularization parameter can be chosen best   in case of ridge regration or best  in .

If we have large dataset – we can split the dataset into three
different parts:

Training dataset – where we model our functions using different
learning algorithms.

Development or Validation dataset – on this we tune the parameters
on held out dataset (development or validation set.)

            Test
dataset – to test best model we have developed on the training dataset.