show NA for n_eff/ESS if k > k_threshold

[avehtari]

Fixes #191

Affects only what is displayed in pareto_k_table. For example, old

Computed from 4000 by 262 log-likelihood matrix.

         Estimate   SE
elpd_loo   -859.1 37.7
p_loo        11.4  2.8
looic      1718.2 75.4
------
MCSE of elpd_loo is NA.
MCSE and ESS estimates assume MCMC draws (r_eff in [0.5, 1.2]).

Pareto k diagnostic values:
                         Count Pct.    Min. ESS
(-Inf, 0.7]   (good)     261   99.6%   117     
   (0.7, 1]   (bad)        1    0.4%   71      
   (1, Inf)   (very bad)   0    0.0%   <NA>    
See help('pareto-k-diagnostic') for details.

New

Computed from 4000 by 262 log-likelihood matrix.

         Estimate   SE
elpd_loo   -859.1 37.7
p_loo        11.4  2.8
looic      1718.2 75.4
------
MCSE of elpd_loo is NA.
MCSE and ESS estimates assume MCMC draws (r_eff in [0.5, 1.2]).

Pareto k diagnostic values:
                         Count Pct.    Min. ESS
(-Inf, 0.7]   (good)     261   99.6%   117     
   (0.7, 1]   (bad)        1    0.4%   <NA>    
   (1, Inf)   (very bad)   0    0.0%   <NA>    
See help('pareto-k-diagnostic') for details.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (68ee019) 92.44% compared to head (fc450fa) 92.44%.

❗ Current head fc450fa differs from pull request most recent head ef4d236. Consider uploading reports for the commit ef4d236 to get more accurate results

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #248   +/-   ##
=======================================
  Coverage   92.44%   92.44%           
=======================================
  Files          31       31           
  Lines        2831     2832    +1     
=======================================
+ Hits         2617     2618    +1     
  Misses        214      214           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.