Make two datasets match by shifting one along x-axis

Question

I have two datasets, each of which consists of a number of (x,y) pairs. Both have a similar shape when y is plotted vs x, but they are shifted along the x-axis with respect to each other. I would like to find the offset in x required to make the two datasets overlap (see attached images).

Is there an elegant way to do this in JMP? I could write a loop that shifts one dataset by small increments and calculate the xi^2 between the two for each iteration to find which offset gives the best match, but I was wondering if there was a smarter way of doing it.

goal.PNG

start.PNG

gandi2223 · Answer

Ian,

* Not necessarily
* Yes
* I was thinking of starting with the smaller dataset (i.e. fewer x/y pairs) and calculating the squared difference between each of its datapoints and the (interpolated) y value of the 2nd dataset at the same x value. The x offset between the two datasets where this value becomes a minimum should correspond to best overlap. It shouldn't be too hard to script this, but before doing so, I thought I'd ask if JMP has a built-in function to achieve this goal.

Thanks,

Marcus

ian_jmp · Answer

I don't think you have overlooked anything that's built in.

Useful prescriptions probably depend on what you can and can't assume about the data, and what figure of merit to use:

* Would you expect each sample to have the same number of (x, y) pairs?
* Is it possible that one or more x or y value is missing?
* How would you wish to quantify 'degree of overlap'?