Bayesian interpretation of regularization

Bayesian interpretation of regularization is like a way of controlling the complexity of a model. The basic idea is that when you have a complicated model, the model could fit to the data too well. This could be a problem, because if the model was fitted too well to the data it has been trained on, it may have a hard time making accurate predictions on data it has not seen before.

Regularization helps solve this problem by guiding the model to not pay too much attention to the details of the data it is given and instead focus on the general shape. It does this by adding an extra penalty or 'cost' to the model for complexity. This means that the model will be more likely to find simpler solutions which still perform well on the data it is given. The regularization helps to keep the model from over-fitting to the data.