Problems with calculating Tree/branch data

Hi all, I am having an issue which I cannot resolve.

Have posted related questions in the past few days with some insight, but nothing that quite works.

I am attaching an image of my problem.

Basically I have a Tree with x number of branches each have y number of values.

I want to take one branch, and do calculations between that branch and all the other branches.

This is not a problem, if I use the [Branch Item] component. Yet this needs to be scalable and I cannot use [Branch Item] for hundreds of branches.

It seems like there should be a solution to this, but I cannot figure it out.

The example problem shown in the image is trying to do a correlation coefficient calculation, although the exact calculation does not matter.

Any help at all is greatly appreciated.

Thanks!

Attachments:

pearson_problem.jpg, 828 KB

Replies to This Discussion

Permalink Reply by Adam Laskowitz on August 29, 2012 at 3:54pm

A current work-around is to use the [Branch Item] and attach the output of the equation to the [Data Recorder], then use a slider to quickly jump through all the branches in the tree.

This works, but it is not ideal.

Permalink Reply by Jissi Choi on August 29, 2012 at 10:11pm

Hi..

I don't know what is the problem exactly..

I think the attached image shows proper definition meeting your needs..

Do you mean it doesn't work for hundreds of branches..??

How about uploading your bad case..??

or Use rather Params Viewer + List Item than text Panel..

Permalink Reply by Adam Laskowitz on August 29, 2012 at 11:13pm

Thank you for the reply.

You are right, there is nothing wrong with that definition.

While I do mean it doesn't work for hundreds of branches, it doesn't even work for 2 branches.

Well, it works I suppose, but just not doing what I thought/want it to do.

Here is an image of "the problem"

What I want is for there to be twice the number of outputs because I am inputting 2 branches here rather than one. So 12 values rather than 6. I actually am not sure what numbers it is computing in this case.

Attachments:

pearson_problem_2.jpg, 412 KB

Permalink Reply by David Stasiuk on August 30, 2012 at 1:24am

Brian's right...I think a script component is definitely the way to go here. I've written one that I think will do the trick for you. Pardon the inelegance of my VB...I'm sure there are much easier ways to do it, but this at least works.

I've gone ahead and restructured the data in your output, because layering in multiple "lookup" paths requires it, so that the resulting cross-referenced datasets indicate both the original "lookup" path as well as all of the raw paths it's being compared to. Does this accomplish what you need?

Attachments:

tree_cross_ref.gh, 3 KB

Permalink Reply by Jissi Choi on August 30, 2012 at 3:05am

As David's VB script..

It is needed to match numbers of lists..

For example..

when you calculate 3 lists(A) with 5 lists(B)..

the last list of A is matched to the remain lists of B..

So the coefficients of {0} and {2} might be "1"..

If you want n*k Pearson Coefficient..

you need n*k duplicated lists.. (rows/columns)

And David's Script does it..

Permalink Reply by Adam Laskowitz on October 25, 2012 at 4:31pm

I am starting to follow how this is *supposed* to be working.
Why are you feeding 1,3,5 into the "lookup" variable? I am unsure of the significance of these numbers here.

Permalink Reply by Adam Laskowitz on October 26, 2012 at 4:25pm

I see now. This is completely working for me. Thanks so much!

I am assuming the "1,3,5" branches were just used as examples.

Thanks!

RSS