prince-computer-vision
prince-computer-vision copied to clipboard
jwdinius
work-through of Computer Vision: Models, Learning, and Inference by Simon J.D. Prince
Hi jwdinius First of all, thank you very much for this repository. I found a typo in chapter 3 problem 2 where lambda was incorrectly typed as alpha. See the...
$A$ is not necessarily invertible. Use moment generating function instead.
After fixing the solution to Problem 9 of this chapter, recompute the predictive distribution values for the Bayesian approach.
For the predictive density, the factor in the denominator should not include the $$+1$$. The term with the sum $$\tilde \alpha + \tilde \beta$$ should be just that (no $$+1$$).
I think there is an arithmetic error. The final result should be $$\frac{\alpha + \sum x_i - 1}{\alpha + \beta + I}$$.
The expression for $\mu$ is correct, but there is a simpler analytical approach for $\sigma^2$: $\sigma^2 = \frac{\beta}{\alpha + \frac{3}{2}}$