Finding residuals of curves

Hi there,

I am trying to compare the results of data obtained from two different mathematical models. The blue curve represents the meaasured values and the red one represents the predicted values. I would like to find the residuals of these two curves and plot a box-plot for data analysis. Is there any data fitting method that can be done just from the curves generated?

[Rik]

You can calculate the residuals by subtracting the predicted value from the true value. If you attach an example for your data we can help you find the residuals.

As a side-note: the residuals will be quite big, because your function is quite steep. That means your output might not reflect well how near this fit is.

[Linn]

Hi Rik,

These are the function values of two curves. I tried subtracting the predicted value from the measured value to find the residual. As you said it is quite a large residual. I am not sure why box plot would work.

[Rik]

Do those values share x coordinates?

And I didn't mention anything about the feasibility of a box plot. You can use one if you want, but it will not show the high level of agreement between model and data. A small shift in x will result in a large residual (which is measured in y).

[Adam Danz]

It looks like the x values do not align. The blue line ends around x=0.68.

@Linn, you'll need to be a little more descriptive about the data you're working with. I saw that you attached files that appear to be vectors of numbers but I'm interested in knowing their sizes. IDoes the n'th number in one file correspond to the n'th number in the other file? Maybe you could fill in some of these missing pieces to help us get an idea of what you're working with.