08-11-2017 11:45 AM
Simply put, I can't find it... Where is the documentation on the algorithm options for the polynomial fit VI? Why is it so hard to find? And if it doesn't exist, why not?
08-11-2017 11:53 AM
http://zone.ni.com/reference/en-XX/help/371361H-01/gmath/general_polynomial_fit/
08-11-2017 11:54 AM
Completely unhelpful. Read the document you linked to. It has zero information on the algorithms.
08-11-2017 11:58 AM - edited 08-11-2017 12:05 PM
Edit: someone replied while I was typing my response.
If you go to the linked page and scroll down, you'll see some algorithms listed under the section marked "General Polynomial Fit Details". Under the "algorithm selector" section where it lists about 6 choices, you can Google any of those entries plus the word "algorithm" to find plenty of details.
I've noticed that the two default options (Least Square, SVD) give me the exact same results as Excel does when I use its polynomial fits, if that helps...
All of the fitting VIs are listed here and you can click on their names to get the full documentation for each one separately. At the bottom of the list there's a link to the advanced fitting VIs as well, in case those are what you need.
These are also found live in LabVIEW. While you have context help enabled (Control-H, or toggle the question mark button on the toolbar), hover your mouse over the VI you need documentation on, and at the bottom of the window there's a blue "detailed help" link you can click, or the blue question mark at the bottom left.
08-11-2017 12:04 PM
@ssegal wrote:
Completely unhelpful. Read the document you linked to. It has zero information on the algorithms.
Ask a general question and get a general answer. The names of the algorithms are listed on the page and there is a wonderful tool called Google where I'm sure you can find more information about them.
08-11-2017 12:04 PM
Clearly I'm not getting my point across.
The documentation does not contain ANY information on the algorithms.
I don't require links to the documentation page.
I know where the page is.
The page doesn't have the info I'm talking about.
Here's what I mean: I know that "SVD" is singular value decomposition. I've never heard of the "Givens" method, nor the "Givens2". What are they, and what's the difference? What's the "householder" method? I could spend a day trying to dig this up on my own via internet and literature searches, but shouldn't at least basic info be provided in the LabVIEW documentation?
Please, please, please do not post yet another link to the documentation page.
08-11-2017 12:10 PM
Maybe go back a step and explain why it's so vitally important that you know the exact algorithm used, step by step?
It's obvious you think you need it for something but it could well be you don't actually need what you're asking for. And right now it sounds like you're demanding that other forums users or NI staff do your work for you.
The general theory is that LabVIEW is made for engineering types, who usually take advanced math classes, and describing the name should be simple enough because if you really truly need to know the algorithm then Googling the names of said algorithms is exactly what you ought to do.
08-11-2017 12:11 PM - edited 08-11-2017 12:30 PM
All these are well known algorithms in linear algebra and have nothing directly to do with polynomial fitting.
Have you tried googling these terms?
They all do basically the same thing and just differ in sensitivity to particular types of bad inputs.
08-11-2017 03:57 PM
The purpose of the LabVIEW Help is (believe it or not) to help you use the particular function or tool in LabVIEW. If you are using a "general polynomial fit", NI (or "LabVIEW") assumes you know something about the meaning of the term "general polynomial fit", but want to know precisely how to "call" (or, in the case of LabVIEW,) "wire" the function. Which inputs are required? Which are optional? What else might you need to know to make an intelligent choice? A complete description of the algorithms does not make sense in this context. Your textbook (or a web search) will allow you to obtain the level of detail you desire.
Bob Schor
08-11-2017 04:31 PM
Came in to ask a question of the community, got a raft of sarcasm.
I'm taking as my comparison point the detailed documentation provided by other software suites that do curve fitting. They assume you know what you're doing, but they still explain what's happening under the hood in enough detail for a user to decipher the algorithm without having to dig up a copy of Numerical Recipes. Which is what I had to do in this case. I'm specifically comparing to Origin and MATLAB, where I do most of my data analysis.
In my experience, it is dangerous to use a canned routine without understanding what it's doing. Your mileage may vary.