1

I've looked on how to find hyper parameters of posterior distribution for normal distribution likelihood with unknown mean and precision.

Here is a derivation described https://www.cs.ubc.ca/~murphyk/Papers/bayesGauss.pdf

Im trying to understand how it is done.

I've been looking on some likelihood equation derivation (61 equation in the paper above). I was following on, but I couldn't figure out how one transformation is done.

Can you help me with one part, please? How comes this:

equation form 1

becomes this

equation form 2

the full equation if you like equation

Bernard
  • 175,478
Dmitriy Suniaikin
  • 21

1 Answers1

1

again. I've spent almost half a day to solve this puzzle and transformed this equation to the same form as in the book.

It turns out that the trick of adding -2ab + 2ab to equation sum did it, I was able to collect vars to the form of (a-b)^2 and solved it to the final form with the mean x.

Dmitriy Suniaikin
  • 21