(Eliot editing this list)
There is a broad framing for this class if issue, namely, analyses that emerge from multiple spades calls. Originally, this was framed within the experiment and experiment2 functions, but these were not reliable enough and they didn't try to cover all the classes of "meta" spades modeling.
Classes of problems:
[ ] Sensitivity Analyses (e.g., varying parameters and evaluating outputs as a result)
[ ] Replication (for stochastic models)
[ ] Validation (and importantly, inter-model comparisons of validation)
[ ] Consensus modeling (e.g., weighted averaging of multiple inputs, based on e.g., validation scores)
[ ] Ensemble modeling (e.g., range of modeling outcomes)
[ ] Multi-model comparisons (e.g., running many, visualizing each separately, but in organized structures)
[ ] Forecast horizon
Technical issues:
[ ] discuss/decide how best to incorporate specific sensitivity analyses from AB_NRV work;
[ ] consolidate all multi-simulation and parallel-related adjustments from WBI and NRV projects;
[ ] ensure compatibility for workstations with very high core counts;
[ ] implement support for simulation scheduling
[ ] use of simulation results databases
@eliotmcintire @CeresBarros I'd like to meet to discuss these changes/additions
(Eliot editing this list) There is a broad framing for this class if issue, namely, analyses that emerge from multiple
spadescalls. Originally, this was framed within theexperimentandexperiment2functions, but these were not reliable enough and they didn't try to cover all the classes of "meta"spadesmodeling.Classes of problems:
Technical issues:
@eliotmcintire @CeresBarros I'd like to meet to discuss these changes/additions