IKH

Regularisation

Having established that we need to find the right balance between model bias and variance, or simplicity and complexity, we need tools which can reduce or increase the complexity. In this segment, we will learn regularisation methods which are used to keep the model complexity at bay.

Regularisation is the process of deliberately simplifying models to achieve the right balance between keeping the model simple and yet not too naive. It is a part of the learning algorithm with some explicit steps to control the model complexity. Recall that there are a few objective ways of measuring simplicity – choice of simpler functions, lesser number of model parameters, using lower degree polynomials, etc.

Regularisation discourages the model from becoming too complex even if the model explains the (training) observations better. It is used to find the optimal point between extreme complexity and simplicity.

Report an error