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
{{ message }}
Repository navigation
Additive Gaussian uncertainty centered on observed values
#2
Version 1.0.0 – Additive Gaussian uncertainty centered on observed values
This version estimates the unknown true time-weighted average exposure $‘TWA_{true}‘$ by adding Gaussian noise $‘E‘$ to the observed value $‘TWA_{obs}‘$:
$twa_{true,i} = twa_{obs,i}+ε_i,$ where $ε_i$ is a draw from $E \sim N(0,σ_i) = N(0, {CV_{t,i}} . twa_{obs,i} )$
The model assumes the observed value is “true”, and that uncertainty is symmetric and additive. This is mathematically problematic:
The uncertainty should be modeled around the unknown true value, not around the observed realization.
It can yield negative exposures, which are unphysical.
The model is conceptually incorrect, but may produce reasonable UTL estimates for low CVt.
This version is preserved:
For didactic purposes
To compare with the improved version
Because its results may still be acceptable in specific cases
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Version 1.0.0 – Additive Gaussian uncertainty centered on observed values
This version estimates the unknown true time-weighted average exposure$‘TWA_{true}‘$ by adding Gaussian noise $‘E‘$ to the observed value $‘TWA_{obs}‘$ :
The model assumes the observed value is “true”, and that uncertainty is symmetric and additive. This is mathematically problematic:
This version is preserved:
This discussion was created from the release Additive Gaussian uncertainty centered on observed values.
All reactions