Brier score and penalized casebase

[Jesse-Islam]

This is about where I am for my section. The brier score from casebase seems to be particularly bad... it seems penalization on cox does not experience the same effect. its possible that removing the penalization from the time variable will improve the plot.

This missing paragraph attempts to explain the lollipop brier and cumulative incidence plots... But the brier score is hard to explain.

[turgeonmaxime]

@Jesse-Islam here are my comments, as we just discussed:

• [x] There's still a comment in red in the "Existing Packages" section. Please fix this.

• [x] Move the discussion about data imputation to the package documentation.

• [x] Consider removing the population-time plot, or reducing the size of the dots.

• [x] Make sure the colours in your graphs match, between graphs (i.e. same methods, same colour) and with the legend.

• [x] Include your code within the text, like the other case studies.

• [x] Think about whether we should discuss the coefficient estimates, and how.