v0.4 misc

[lindeloev]

BUGS / VERIFY:

• Fix prior predictive is accurate and is plotted correctly?
• multiple: sim and match: not necessarily applicable? (though I think they are)
• fit$pars$reg should be split to "fixed" and "cp"
• Add test of prior predictive to the unit tests!
• Unit test of a few basic plots? (use seed!)
• Check summary() parameter order.
• Verify names in summary() etc. ("col1:x_1"). What should names be in code? ("col1Xx_1"?)
• Verify the fits of v0.4.0 to v0.3.4! E.g., by re-running the vignettes and comparing the local renders to the online ones.
• Rename assert_() to mcp_assert_()

DOCS:

• Update all vignettes to mention fitted, predict at the first plot.
• Update all vignettes with pp_check()
• The links to external packages fail in the PDF manual, e.g., brms::hypothesis.
• Consider terms: "coefficients", "parameters", "effects"
• Release 0.3.4 on github. The 0.4.0.

A LATER RELEASE?

• Write a vignette on simulating data, including using fit$simulate()
• Perhaps good datasets here: https://www.researchgate.net/publication/260553778_Inference_for_Change-point_and_Post-change_Means_after_a_CUSUM_Test/figures?lo=1
• New article on weighted regression? Weights cannot be zero. Only affects fitting and plotting - not simulate() etc.