UBC improvements suggestions

[audrey-b]

data plot increase the font (too small when lots of data) DONE
plot single arm studies (currently net.plot gives error)
bias adjustment models
leverage : which arms they come from PROBABLY EASY
sort forest plot DONE
get indirect estimate
color edges by comparison (high vs low vs unclear risk of bias)
weight by number of patients
help file: weight is by number of studies in network plot DONE
help file: se is sd/sqrt(n) not n DONE
meta-regression with categorical (low, high, unclear risk of bias)

[audrey-b]

Jon:

1. Change documentation on BugsNet re: data.plot for nodes/edges (ie: change size/colour related to other categorical/continuous variables)
2. Identification of study and arm when interpreting posterior mean deviance (consistency/inconsistency) plots
3. Identification of indirect effect estimates for pairwise comparisons when available (for comparison vs direct estimates)
4. Meta-regression with continuous or categorical variables, and for more than a single variable at a time (and potentially for interactions?)

Ben:
-It would be great if BUGSnet could produce similar results to CiNeMa but using the Bayesian results.
-Give estimates for both direct, indirect, and Bayesian network estimates.