Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Total error calculation formula in parameter identification #293

Open
mmalsmadi opened this Issue Mar 13, 2019 · 5 comments

Comments

Projects
None yet
4 participants
@mmalsmadi
Copy link

mmalsmadi commented Mar 13, 2019

Hello,

I need to know how is the "total error" is calculated with the parameter identification inside PK-Sim
what is the formula used for its calculation?

many thanks

@PavelBal

This comment has been minimized.

Copy link
Member

PavelBal commented Mar 13, 2019

Please see Section "Mapping Simulation Outputs to Observed Data - Scaling". The total error is the sum of single residual errors (@msevestre correct me if I'm wrong).

@StephanSchaller

This comment has been minimized.

Copy link
Member

StephanSchaller commented Mar 13, 2019

It is the sum of squares.

@msevestre I still have to update the documentation ;-)

@StephanSchaller

This comment has been minimized.

Copy link
Member

StephanSchaller commented Mar 13, 2019

See discussion here

@mmalsmadi

This comment has been minimized.

Copy link
Author

mmalsmadi commented Mar 13, 2019

So, to confirm that I got this right
Total error = log [sum(((Observed[i]-Predicted[i])/Predicted[i])^2)]

right?

@msevestre

This comment has been minimized.

Copy link
Member

msevestre commented Mar 17, 2019

@mmalsmadi This is incorrect.

Total error = Square of ( Sum of all residual ^2 for all outputs)

Whether log is used or not depend on the output and the scaling selected. Also Weight is also part of the equation (global weight for the output and single weight for each data point). Those weights are one by default so if you did not change them, you can forget about them

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
You can’t perform that action at this time.