How can I subtract vectors of different lengths?

Hello,

I have two vectors of different lengths:

• LargerVec (length=7012)
• SmallerVec (length=7004)

The best way I've come up with to solve this to be able to subtract these vectors is to use SPILLOVER to crop off the last eight values in the larger vector:

spillover=length(LargerVec)-length(SmallerVec);
LargerVec(length(LargerVec)+1-spillover:length(LargerVec))=[];

% Gradients, lower-higher inlets
delta=LargerVec-SmallerVec;

This works, but I don't want to arbitrarily lose those last eight values from LargerVec. I would prefer to extend an average of SmallerVec.

Ultimately, I want the method that will be the most true to the data.

Thank you for your help.

[Stephen23]

"I want the method that will be the most true to the data"

If you know the underlying model that exactly predicts your data then this might be a reasonable request. Otherwise you will just be inventing data.

[Adam]

What do you expect the last 8 values to be then? I don't really understand what you mean by 'extend an average of SmallerVec'.

LargerVec(1:numel(SmallerVec)) - SmallerVec

will do the subtraction of valid values. How you then introduce the other 8 values is up to you I guess and you can just tag them back on the end.

[balsip]

Thanks guys. I'm now understanding that my question is more conceptual than technical, and I think that the "most true" approach would be to throw out the extra data of the longer vector.

[dpb]

That would like be so in the spirit of "most true to the data" as it isn't doing any extrapolating which is more risky than interpolating, even.

[balsip]

Agreed, dpb. Thank you much.

Answer 1

[Thorsten]

You do not need much code, you can simply write

delta = LargerVec(1:numel(SmallerVec))-SmallerVec;

or if you want to delete the values in LargerVec, you can use

LargerVec = LargerVec(1:numel(Smallervec));