MATLAB and Simulink resources for Arduino, LEGO, and Raspberry Pi

Learn moreOpportunities for recent engineering grads.

Apply TodayTo resolve issues starting MATLAB on Mac OS X 10.10 (Yosemite) visit: http://www.mathworks.com/matlabcentral/answers/159016

Asked by Isa Isa on 28 Dec 2012

Hi, I have a figure that contains about 10 curves. I have another figure that has only one curve. I want to math one curve figure with one of the curves in the figure that has 10 curves. Both figures have different scale.It is only the curve I want to match. Please how do I do it? Thanks

Regards, Isa

*No products are associated with this question.*

Answer by Jan Simon on 28 Dec 2012

What does "match" exactly mean? Do you want to compare the mean value, the shape, are the curves shifted horizontally or vertically only, do they have the same mean, but different scales, or are the curves equal expect for some noise?

If only the scales (and means) differ, you can "normalize" them by dividing by the maximum absolute value (and subtracting the mean). Then "matching" could mean to find the minimum difference between the single curve and the set of curves. Now "minimum difference" must be defined also: does "distance" mean the sum of the absolute differences of all points, the maximum difference, the mean or minimal difference, the standard deviation of the differences, etc.?

Show 4 older comments

Isa Isa on 30 Dec 2012

Thanks. I have standard models put in curves, the curves are in both figure and bitmap image. I have uploaded a typical case please check this link http://imageshack.us/photo/my-images/844/curvematch.jpg/ The curves in blue are the standard curves and the curve in red is the one I have numerical values for its x and y and I want to match it with any of the curves in blue. Thanks Isa

Jan Simon on 30 Dec 2012

@Isa: Please insert all important information by editing the original question. This is the location, where readers expect them.

You still did not explain what "match" exactly means, but this is the fundamental point of your question. I have the impression, you want to move and resize the `axes` object with the mouse until you find a visually similar curve. This procedure would be neither reproducible nor exactly enough for a scientific usage. A programmatic comparison and a quantification of the difference is required. But even then there are curves, which have the same distance to two curves of your set. Again I want to stress, that *distance between curves* have to be defined at first according to your exact needs.

It is much better to keep the complete discussion here in the public forum and not to send further information by email.

Isa Isa on 30 Dec 2012

Thanks, by match I mean superimposing a plot of the test data on a set of type curves on the same scale and using the characteristic shapes of the type curves to identify an interpretation model for the test data. By this I will be able to select an interpretation model that fits the test data. In the image I sent, there are two dependent variables on both the test data and type curve. curve matching is generally accomplished through automatic fitting of the test data using linear or non-linear regression techniques. Thanks Isa

## 0 Comments