12-18-2018 11:21 AM
Hello,
Is there a way to interpolate an (x;y) cloud of points?
For instance I have get a set of x(t) variable (let say Force) and also y(t) (let say stroke)
If I plot x-y Graph I get a curve like an hysteris.
I would like to get the linear interpolation of all (xi;yi) points.
Workaround is to export xy graph in excel and use Trendline
Thanks for advance
Maxime
12-20-2018 07:47 AM
Hi,
if you are looking to get a curve fitting/regression from the cloud point, Excel is a good option.
In Amesim ,there is an advanced hysteresis submodel in the Signal library which allows to enter a couple of curves for an increasing x and for a decreasing x. This submodel will then 'interpolate' between each consecutive points of your curves but you don't get a regression/trend:
So if you wanted to use this back in Amesim, you could:
- get your clean increasing and decreasing curves from a regression/trend calculation
- put them as a few x,y points into a couple of lookup table as described in that submodel
01-25-2019 11:12 AM
Hi mAx,
Are you talking about "linear interpolation" or "linear data fitting" ?
- Linear interpolation draws piecewise linear segments between each (xi,yi) points.
- Linear data fitting performs a linear regression on sample points to estimate a linear curve (y = a.x+b) that is the best estimate to reproduce the global behavior of the cloud of points. This is what is done by Excel trendline function.
Amesim signal library includes submodels that can perform linear interpolation (e.g. SIGFXA01). In such case the x value must be monotonic.
For linear data fitting, however, the easiest way to find a and b coefficients is indeed to use Excel.
Hope this helps.
01-28-2019 01:45 AM
Good Morning @OliverS
I was talking about linear data fitting