You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The method produces a species growth rate for each species in each year: these are imputed for species-year combinations with no data. Both "real" and imputed growth rates are included in the spgrowth object, which we use for bar charts in UKBI.
However, for UKBI it does not make sense to include these imputed values for categorising species. I recall that @03rcooke used a workaround for either the Scottish or English indicator set. Could this be easily incorporated into the package?
To avoid drastically changing the behaviour of the method, I would suggest that we add a variable to indicate which growth rates are estimated from data vs imputed, rather than hard coding the removal of imputed values* or insisting that users define what they want a priori.
*it's sometimes of interest to examine the imputed values, e.g. to test whether there may be biases introduced by species with incomplete time-series.
The text was updated successfully, but these errors were encountered:
Hmm, I don't actually recall this. I don't think I realised that there would be the possibility to separate "real" and imputed growth rates. I thought the spgrowth object just included modelled growth rates, so I treated them all the same. The workaround I built was to accommodate the bma method into the species and indicator assessments, which I have now incorporated into the BRCindicators package.
The method produces a species growth rate for each species in each year: these are imputed for species-year combinations with no data. Both "real" and imputed growth rates are included in the
spgrowth
object, which we use for bar charts in UKBI.However, for UKBI it does not make sense to include these imputed values for categorising species. I recall that @03rcooke used a workaround for either the Scottish or English indicator set. Could this be easily incorporated into the package?
To avoid drastically changing the behaviour of the method, I would suggest that we add a variable to indicate which growth rates are estimated from data vs imputed, rather than hard coding the removal of imputed values* or insisting that users define what they want a priori.
*it's sometimes of interest to examine the imputed values, e.g. to test whether there may be biases introduced by species with incomplete time-series.
The text was updated successfully, but these errors were encountered: