I am obviously not a math wiz, so please pardon my crude explanation... I'm just assuming this can be done, and need to know WHAT it is called, so I can research and educate myself on the process and the math.

I know in LINEAR REGRESSION a line is calculated/drawn through the set of points and a SLOPE value describing the line is calculated/returned...

(Please point me in the right direction...)

In place of the SLOPE LINE, my data points resemble a CURVE... I want to calculate/draw a CURVE (both concave and convex looking) through my points and then calculate/return a result (like the linear regression slope) that defines the CURVE...

My goal is to end up with a value (slope, like in LR) that describes the CURVE, so I can compare OTHER CURVES (form other data points) and determine how closely the "new" curves compare to "known" curves.

Is there something like the Linear Regression SLOPE used similarly to define a CURVE?

Thank you!

  • $\begingroup$ Like polynomial regression? $\endgroup$ – Ashwin Iyengar Sep 28 '16 at 19:40
  • 1
    $\begingroup$ If you scatter plot your data in Excel you can try and fit many types of curve to it (right click a data point and Add Trendline). Your description sounds like you want to fit an exponential curve. $\endgroup$ – Paul Sep 28 '16 at 19:42

Check out polynomial regression


and in general, nonlinear regression



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.