Bayesian Machine Learning (Part 8)

Now from the above figure we can say the error is very small as the gaussian distribution is conveniently satisfying the job. Also, in machine learning problems, we really do not want an exact value of the posterior as far as we can get relative values for decision making.