When the number of random effect entities is small, e.g. in the cohort-based models, users would like to check the evaluations on each group (data grouped by member id, job id, cohort id etc) and analyze the importance of each group to overall evaluation. This PR aims to output the per group evaluation in separate output files.
junshi15
commented
4 years ago
When the number of random effect entities is small, e.g. in the cohort-based models, users would like to check the evaluations on each group (data grouped by member id, job id, cohort id etc) and analyze the importance of each group to overall evaluation. This PR aims to output the per group evaluation in separate output files.