this is for holding javascript data
Xavier Holt edited subsection_Prior_for_Weights_In__.tex
over 8 years ago
Commit id: a8497c5ccfd8b5b86e77d8f54ce20b418d5222f9
deletions | additions
diff --git a/subsection_Prior_for_Weights_In__.tex b/subsection_Prior_for_Weights_In__.tex
index d247a9f..b45eacb 100644
--- a/subsection_Prior_for_Weights_In__.tex
+++ b/subsection_Prior_for_Weights_In__.tex
...
It is then readily apparent that, given our diagonal $\Sigma$ we have $\Sigma^{-1} = \text{diag}(\sigma_1^{-1}, \dots, \sigma_d^{-1})$ and similarly $|\Sigma^{-1}| = |\text{diag}(\sigma_1^{-1}, \dots, \sigma_d^{-1})| = \prod(\sigma_i^{-1})$.
Using this information, it can be shown that finding the MAP estimator of $\mathbf{w}$ is equivalent to solving the maximum likelihood problem with a ridge regularisation parameter \cite{Hoerl_1970}. The success and ubiquity of this approach then speaks to the reasonableness of assuming Guassian priors.
\begin{align}
L(\mathbf{w} | X, Y) \propto P(Y | X, W) P(\mathbf{w})
\end{align}