this is for holding javascript data
Graham McVicker edited Estimating overdispersion parameters.tex
over 9 years ago
Commit id: 9d0ff037d8c012de651334fd6c513c04f34d99b8
deletions | additions
diff --git a/Estimating overdispersion parameters.tex b/Estimating overdispersion parameters.tex
index 37bd808..1e3669f 100644
--- a/Estimating overdispersion parameters.tex
+++ b/Estimating overdispersion parameters.tex
...
\subsubsection{Beta-Negative-Binomial}
To find the maximum likelihood estimate of $\Phi_i$ we need to sum the likelihood across all regions. This presents a problem, as $\eta_j$ must also be estimated for each region. We therefore interatively estimate $\eta_j$
by first finding a maximum likelihood estimate for $\eta_j$ for each region using the
equation equation:
\[
\textrm{L}\left(\eta_j \left| D \right. \right) = \prod_i \left[ \Pr_{\mathrm{BNB}} \left( X = x_{ij} \left| \lambda = T^*_{ij}, \Phi_i, \eta_j \right. \right) \right]
\]
and then
finding a maximum likelihood estimate for $\Phi_i$
for each individual using the
equation equation:
\[
\textrm{L}\left(\Phi_i \left| D \right. \right) = \prod_j \left[ \Pr_{\mathrm{BNB}} \left( X = x_{ij} \left| \lambda = T^*_{ij}, \Phi_i, \eta_j \right. \right) \right]
\]
We repeat this iterative procedure until the improvement in likelihood becomes negligable.
\subsubsection{Beta-Binomial}
To find the maximum likelihood estimate of $\Upsilon_i$ we sum the allele specific likelihood across all regions. We again assume there is no genetic effect, so $p$ = 0.5.