• State Verified Answer
  • Replies 5 replies
  • Subscribers 12 subscribers
  • Views 1252 views
  • Users 0 members are here
Related

Agilent Knowledge Portal

Find maintenance, best practice, and troubleshooting articles.

Agilent Knowledge Portal

Calibration Curve Fit Assistant

CKinter

Hello, 

I was wanting to utilize the Curve Fit Assistant for optimizing my calibration curve. I know that I want to use a quadratic fit from my regression analysis, and Curve Fit Assist agrees. It also recommended adding a 1/SD^2 weight to my curve. Can someone help me to understand how this weight is being calculated? I like that the 1/SD^2 is adjusting my y-intercept without having to force any origins. 

Parents
  • 0

    Hi ,

    The Curve Fit Assistant is looking for the best statistical fit of the data. The program calculates 1/SD^2 as "one over the variance (standard deviation squared) of the responses of the replicate calibration points at the same level as that point" according to the Quant Help.  This particular weighting will only work if there are replicates for each calibration level.

    The Curve Fit Assistant is presenting mathematical possibilities, but it is up to you to determine a particular curve and weighting makes analytical sense and is allowed by your lab policies and procedures.

    Britt

Reply
  • 0

    Hi ,

    The Curve Fit Assistant is looking for the best statistical fit of the data. The program calculates 1/SD^2 as "one over the variance (standard deviation squared) of the responses of the replicate calibration points at the same level as that point" according to the Quant Help.  This particular weighting will only work if there are replicates for each calibration level.

    The Curve Fit Assistant is presenting mathematical possibilities, but it is up to you to determine a particular curve and weighting makes analytical sense and is allowed by your lab policies and procedures.

    Britt

Children
Was this helpful?
Privacy Statement
Terms of Use
Contact Us
Site Help