cancel
Showing results for
Did you mean:
cancel
Showing results for
Did you mean:

SOLVED
Aquamarine

## Best Smoothing Approach

Hi

I am having issues formulation a smoothing function. Errors like "This value must be a vector of real numbers: pop up.

A working example will help a lot.

Prime 6 File Attached.

Tags (1)
1 ACCEPTED SOLUTION

Accepted Solutions

## Re: Best Smoothing Approach

l having problems plotting.

Tags (1)
12 REPLIES 12

## Re: Best Smoothing Approach

Your approach failed because i is not a vector but a range!

In the pic below I used an undocumented trick to turn a range into a vector by adding a numeric evaluation when defining the variable i. Furthermore I chose the first value to be 1 instead of zero so the magnitude of i and r.dom match.

But of course you are on the save side if you define the necessary vector the "legal" way:

A working example can also be found in the help

You may also be interested in the information about smoothing in the MC15 help. Find it converted to pdf in the attachments.

## Re: Best Smoothing Approach

Is it possible to integrate the supsmooth function?

## Re: Best Smoothing Approach

@DJF wrote:

Is it possible to integrate the supsmooth function?

supsmooth does not return a function but rather a vector of y-values corresponding to the vector of x-values provided.

You either can write your own "integration" function which works on datapoints given by vectors or you use interpolation (either linear or cubic splines) to turn the result of supsmooth into a function.

## Re: Best Smoothing Approach

A function for the supsmooth can be defined using the x and y vectors with interp function.  This new function can be integrated.

## Re: Best Smoothing Approach

I see.  So, then no real advantage to using supsmooth over say a b-spline (unless you happen to like the fit better).  was hoping to save a step.  Thanks all.

## Re: Best Smoothing Approach

l having problems plotting.

Tags (1)
Sapphire III
(in response to muser)

## Re: Best Smoothing Approach

What's the problem...it plots

doesn't it?

Luc

## Re: Best Smoothing Approach

Sorry wasn’t clear. Can’t plot the smoothed  data.

## Re: Best Smoothing Approach

I don't  see any attempt to plot r.smooth ??

Maybe you just have to move down the plot below the definition of r.smooth?

Thanks!

## Re: Best Smoothing Approach

It looks like you got confused by the indexing of vectors. By default Prime starts with zero.

You defined your range i starting from 1 and then plotted Loc[i over i. That way your plot gets shifted to the left for 1 unit - the first value of your data vector "Loc" is ignored and the second one (index 1) is assigned to 1, etc.

Correct plot should look that way:

## Re: Best Smoothing Approach

Thanks for spotting that.

Announcements
Check out the latest