Maia Kapur1
1SAFS
April 2nd, 2019 9:00 (PST): FSH 203
Sneaking up on a well specified model
Abstract: The sustainable exploitation of marine resources requires management that is based on sound scientific guidance. One of the most important goals of stock assessment modelling is to accurately characterize the important systems model processes by either specifying the correct relationships and by estimating the parameters governing those processes. A well-designed approach to model structuring is needed to ensure good representation of the systems processes. Improper structuring of a model is the result of misspecification of either observation or systems processes. For this work, I plan to use simulation methods to evaluate improvement in assessment model performance (based on a recent Hake assessment) with sequential elimination of a series of known misspecifications of both observation and systems model processes. This talk will go through our proposed methodology and ask for feedback from the audience on how to best execute performance testing and present our findings.