Roadmap for CSEP2 and understood by me. This document contains my notes and progress on a collection of CSEP related things. This represents my unofficial sounding board.
Bold text: High Priority
Italic text: Medium Priority
Normal text: Low Priority
Strikethrough: Complete
Milestones
- paper on testing stochastic event sets
- ucerf3 simulations in Landers sequence
- ucerf3 simulations for other large eqs
- implement n-test, m-test, s-test for stochastic event sets
- determine power of statistical tests using simulation based approach
- run csep1 ETAS simulations in csep2 experiment
- paper on testability of earthquake forecasts that involve faults
- needs thought!
- release csep2 testing center distribution
- write docs for csep2 github
- implement methods to run experiments (db, jobs to hpc, post-processing)
- EGU 2019 (Abstracts due: 01/11/19)
- SSA 2019 (Abstracts due: 01/11/19)
- Statistical Seismology 2019 (Abstracts due: Unknown)
- Read Gneiting work on proper scoring.
- Cramer-von Mises test
- Kolmogorov Smirnov test
- power of evaluations
- how long will we have to perform an experiement so that we
- backend logic to store model data/parameters
- store catalogs
- store results
- store generated forecasts
- maximum likelihood estimation for b-values
- m-test and s-test
- documentation on csep (start with tools to run ucerf3-etas)
- build data model for csep2 runs
- revisit prototypes of end-to-end system; need more runs done before april.