AnsweredAssumed Answered

OpenLab 2.1 - residuals plot (calibration)

Question asked by tomasz.kubowicz on Jul 29, 2019
Latest reply on Aug 1, 2019 by ryoboyle

Like • Show 0 Likes0
Comment • 3

Hello

Is there any built-in chart in OL to generate residuals plot for calibration curve?

Or it needs to be created from scratch?

I've been trying to find it in reports items (charts/calibration etc.) but it looks like there isn't one.

Also in calibration curve tab even if I have few injections per level I can see only one residual value (Residual per level, rel residual, rel residual%) - I guess it is average. Is there any chance to display all of them (for every single injection)?

Regards

Tomasz Kubowicz

1 person has this question

Outcomes

ryoboyle

Jul 31, 2019 1:17 PM

Hello,

I added tags to this post to increase visibility.

There is not built-in chard for displaying residuals for a calibration curve.

For your second question it depends on how you have your curve calculation set in your processing method. The options are 'From average per level' or 'From individual calibration points'.

If you have 'From average per level' selected then you will only see one Residual per level, rel. residual, and rel. residual % value displayed for each level. These values would be the based off of the average of the calibration points for that level.

If you have 'From individual calibration points' selected then you will see Residual per point, rel. residual, and rel. residual % values displayed for each calibration point.

Hopefully that helps.

Actions

tomasz.kubowicz @ ryoboyle on Aug 1, 2019 12:29 AM

Hello

Thanks for answer.

For second question I asked: if I choose "From individual calibration points" I'd have residuals for all injections but my calibration curve will based on one injection per level only (it will not take average from all inj/level). Is that right?

Actions

ryoboyle

@ tomasz.kubowicz on Aug 1, 2019 4:59 AM

Hello,

If you choose "From individual calibration points," the calibration curve will still use all calibration points for every level. It will just use each individual point when performing the linear regression. This is different than "From average per level" which first finds the average of all calibration points on each level and then only uses those average points when performing the linear regression.

This difference is shown graphically on the calibration curve. Below are two calibration curves built using the same data files. The only difference between them is the curve calculation selection in the method. The first calibration curve had "From individual calibration points" selected while the second had "From average per level" selected. Notice that the second curve has a light blue or cyan square point that sits directly on the line. That point represents the average of the 5 calibration points.

Changing the calibration curve calculation in the above example had no impact on the curve fit since it was just a 1 level calibration, but it can have an quite an effect on multi level calibrations.

Hopefully this helps.

Actions

3 Replies

ryoboyle Jul 31, 2019 1:17 PM
Mark Correct
Correct Answer
Hello,

I added tags to this post to increase visibility.

There is not built-in chard for displaying residuals for a calibration curve.

For your second question it depends on how you have your curve calculation set in your processing method. The options are 'From average per level' or 'From individual calibration points'.

If you have 'From average per level' selected then you will only see one Residual per level, rel. residual, and rel. residual % value displayed for each level. These values would be the based off of the average of the calibration points for that level.

If you have 'From individual calibration points' selected then you will see Residual per point, rel. residual, and rel. residual % values displayed for each calibration point.

Hopefully that helps.
Like • Show 0 Likes0
Actions
- tomasz.kubowicz @ ryoboyle on Aug 1, 2019 12:29 AM
  Mark Correct
  Correct Answer
  Hello
  
  Thanks for answer.
  For second question I asked: if I choose "From individual calibration points" I'd have residuals for all injections but my calibration curve will based on one injection per level only (it will not take average from all inj/level). Is that right?
  Like • Show 0 Likes0
  Actions
  - ryoboyle @ tomasz.kubowicz on Aug 1, 2019 4:59 AM
    Mark Correct
    Correct Answer
    Hello,
    
    If you choose "From individual calibration points," the calibration curve will still use all calibration points for every level. It will just use each individual point when performing the linear regression. This is different than "From average per level" which first finds the average of all calibration points on each level and then only uses those average points when performing the linear regression.
    
    This difference is shown graphically on the calibration curve. Below are two calibration curves built using the same data files. The only difference between them is the curve calculation selection in the method. The first calibration curve had "From individual calibration points" selected while the second had "From average per level" selected. Notice that the second curve has a light blue or cyan square point that sits directly on the line. That point represents the average of the 5 calibration points.
    
    Changing the calibration curve calculation in the above example had no impact on the curve fit since it was just a 1 level calibration, but it can have an quite an effect on multi level calibrations.
    
    Hopefully this helps.
    Like • Show 0 Likes0
    Actions

OpenLab 2.1 - residuals plot (calibration)

Outcomes

Related Content

Recommended Content