Difference between revisions of "CSEP Minutes 01-23-2019"
From SCECpedia
Jump to navigationJump to search (Created page with "CSEP Working Group Home Page<br><br> ''Participants'': D. Jackson, M. Werner, W. Savran, J. Gilchrest, D. Rhoades, and P. Maechling '''Minutes:''' * P...") |
|||
Line 7: | Line 7: | ||
* Need to correct KS critical value computation for discrete distributions. | * Need to correct KS critical value computation for discrete distributions. | ||
* Try binning data before computing ECDF. | * Try binning data before computing ECDF. | ||
− | * | + | * Distinction between two-types of tests: |
** Operational Tests: Used to communicate model results. | ** Operational Tests: Used to communicate model results. | ||
** Developmental Tests: More severe tests used to aid in model development. | ** Developmental Tests: More severe tests used to aid in model development. | ||
+ | *** Cumulative N-test visualization. | ||
+ | *** M-test visualization with error bars around binned data. | ||
* Important to understand the deficiencies of the catalog, ie the data are of limited accuracy. | * Important to understand the deficiencies of the catalog, ie the data are of limited accuracy. |
Latest revision as of 19:56, 24 January 2019
Participants: D. Jackson, M. Werner, W. Savran, J. Gilchrest, D. Rhoades, and P. Maechling
Minutes:
- Presented M-test notes here: https://github.com/SCECcode/csep2/blob/master/notes/m_test_notes.ipynb
- Need to correct KS critical value computation for discrete distributions.
- Try binning data before computing ECDF.
- Distinction between two-types of tests:
- Operational Tests: Used to communicate model results.
- Developmental Tests: More severe tests used to aid in model development.
- Cumulative N-test visualization.
- M-test visualization with error bars around binned data.
- Important to understand the deficiencies of the catalog, ie the data are of limited accuracy.