380: Idea to solve convergence problem

[danielinteractive]

part of #380

[danielinteractive]

@yonicd maybe you are interested in reviewing this design PR - I think I solved the puzzle why SAS converges much better in the example

[github-actions]

Code Coverage Summary

Filename                    Stmts    Miss  Cover    Missing
------------------------  -------  ------  -------  -------------------------------
R/between-within.R             59       0  100.00%
R/component.R                  67       0  100.00%
R/cov_struct.R                 97       1  98.97%   407
R/empirical.R                   7       0  100.00%
R/fit.R                       223      11  95.07%   186-189, 407, 450-453, 468, 498
R/interop-emmeans.R            39       0  100.00%
R/interop-parsnip.R            59       1  98.31%   12
R/kenwardroger.R               92       2  97.83%   41, 63
R/mmrm-methods.R              120       0  100.00%
R/residual.R                    8       8  0.00%    10-31
R/satterthwaite.R             116      12  89.66%   238-249
R/testing.R                    55       4  92.73%   29, 31, 70, 72
R/tidiers.R                    72       2  97.22%   46-47
R/tmb-methods.R               287       3  98.95%   277-278, 338
R/tmb.R                       287       0  100.00%
R/utils-formula.R              27       0  100.00%
R/utils-nse.R                  16       0  100.00%
R/utils.R                     124      10  91.94%   276-286
R/zzz.R                        65      19  70.77%   7-17, 50-55, 85, 113, 133
src/chol_cache.h               63       0  100.00%
src/covariance.h              101       1  99.01%   177
src/derivatives.h             126       0  100.00%
src/empirical.cpp              72       0  100.00%
src/exports.cpp                47       0  100.00%
src/jacobian.cpp               47       1  97.87%   54
src/kr_comp.cpp                56       0  100.00%
src/mmrm.cpp                   76       0  100.00%
src/predict.cpp                93       0  100.00%
src/test-chol_cache.cpp        58       5  91.38%   9, 18, 26, 55, 62
src/test-covariance.cpp       123       5  95.93%   9, 29, 40, 61, 72
src/test-derivatives.cpp      108       7  93.52%   44, 53, 62, 85, 94, 106, 124
src/test-utils.cpp            195       7  96.41%   9, 16, 24, 34, 44, 57, 119
src/testthat-helpers.h         15       5  66.67%   36-37, 41, 50, 53
src/utils.h                    84       0  100.00%
TOTAL                        3084     104  96.63%

Diff against main

Filename      Stmts    Miss  Cover
----------  -------  ------  --------
TOTAL             0       0  +100.00%

Results for commit: da369a6

Minimum allowed coverage is 80%

♻️ This comment has been updated with latest results

[github-actions]

Unit Tests Summary

       1 files       43 suites   23s ⏱️
   462 tests    425 ✔️ 37 💤 0 ❌
1 819 runs  1 777 ✔️ 42 💤 0 ❌

Results for commit da369a6.

♻️ This comment has been updated with latest results.

[yonicd]

In general i think that the computation cost of the empirical covmat is small and would relax the constraint that this simulated data showed.

To the question of MIVQUE0, i think that could be a longer term goal in the next release. It would give more flexibility for larger dimension designs (per reference paper) so we may hit against that problem anyways as more people use the package.

[yonicd]

this blogpost may give a direction on a consistent way to resolve NA in the covmat

here is an R package that has similar solutions

[danielinteractive]

this blogpost may give a direction on a consistent way to resolve NA in the covmat

here is an R package that has similar solutions

Thanks Yoni - yeah I think we don't need those more complex solutions because we can start from the residuals data set

[danielinteractive]

@yonicd ah this is cool! thanks