hello

please see my question! mathematica.stackexchange.com/questions/34647/…

@phdstudent hi, please do not advertise questions in chat.

unless you want to contact someone specifically in chat.

Hi everyone , one question: C_{d0,d0}

C - covariance matrix , d0 - data

how can I understand that?

@Spider Please, I've told you before... this ain't the room for that. You're already asking folks in the math chatroom, so it's best to stick to that room. Thanks :)

$rm-rf ohhh, sorry yeah, it is the second time, I didnt realize!

Can I please get some mathematica help? I don't have a question specific enough to ask directly on the forum but it relates to what I assume are Mathematica fundamentals and InterpolatingFunction[]'s

All I want to do is to be able to do calculations involving some InterpolatingFunctions that I have or even just extract their values at specific time values but I am having trouble doing even this.

I may be late to respond to any responders, sporadically checking chat for replies. If anyone pops by, please let me know

@muzzlator Ok, what's your question? If you have a minimal example, that helps

Hi rm -rf. That's my problem, I can't really figure out how to ask the question in a minimal way

But let me try boil down the question by showing the output I'm getting and perhaps saying the output I want

Firstly, I am getting a lot of these errors: General::ivar: "1/100 is not a valid variable. "

Actually, sorry, I am doing a terrible job at explaining this.

Let me try again

OK, I have a variable bsi. When I call Print[bsi], it prints something like this:

{0. +3.77069 InterpolatingFunction[{{0.4999,0.7501}},<>][t]+70.3797 InterpolatingFunction[{{0.4999,0.7501}},<>][t]+69.5832 InterpolatingFunction[{{0.4999,0.7501}},<>][t],0. +3.77069 InterpolatingFunction[{{0.4999,0.7501}},<>][t]+70.3797 InterpolatingFunction[{{0.4999,0.7501}},<>][t]+69.5832 InterpolatingFunction[{{0.4999,0.7501}},<>][t],

That's only a small part of the output but it's a piecewise defined vector function where each component is defined by a linear sum of interpolating functions

@muzzlator General::ivar usually means you are supplying a numerical value when it expects a symbol. A simple example: D[Sin[x], x] and D[Sin[x], 1]

so for example (t,t^2,t^3) when t is in [0, 0.5] and (1, 1-t, t^4) when t is in [0.5, 1]

Right. OK so the first step should be to ? NumericQ everything

That would be a start, although I'm not sure it is exactly necessary... If you can make bsi smaller (e.g. fewer terms) and show your usage, it would be more helpful

I can't make bsi any smaller because it's generated automatically. I did take a screenshot however: i.imgur.com/4yMPLXL.png

It's just a linear combination of (a linear combination of piecewise defined parametric functions in R^3)

Seems fine. What's the next step? How do you use that?

OK let me try say from the top-down approach what my problem is and I will go deeper when necessary

This plot here works perfectly as intended:

To explain each of the details:

bsi as I explained before is a curve in R^3

Actually, ignore the last 7 lines. I'll try explain again. (Thanks for being so patient by the way)

@muzzlator The joys of debugging :D

OK here's a simple question first, suppose I want to find bsi[i] for {i, 0, 1, 1/10}

I don't even know how to properly compute this without getting a long expression in return

@muzzlator bsi/@{0,1,1/10}

@ssch more like bsi /@ Range[0, 1, 1/10]

@rm-rf yea, that's probably what was intended :)

See, that gets me 'very long output'

Really it should be 11 vectors in R^3

Let me try // N

If Dimensions are wrong have a look at just bsi[0], see if that looks right

Funny, if I do bsi[t] /. t->0, I get {8.67362*10^-19, -3.46945*10^-18, 3.46945*10^-18}[0]

If I do bsi[0], I get 'long output'

what does bsi[y] give?

@muzzlator It's the difference between using exact arithmetic and machine numbers

If I do bsi[blah], I still get t's inside the formula for bsi. I then evaluate the whole thing at blah

seems bsi is not really a function then

Right. More some sort of symbol?

Wow, rojo is still here. I was talking to rojo about this question many months ago. I have been stuck on this for ages

Longest debugging session ever

@muzzlator I am still alive, yeah

Ah, I remember your nick because my file is still called forRojo.nb

Hm, so ParametricPlot3D[bsi , {t, 0, 1}] works fine.

ParametricPlot3D[bsi[t], {t,0,1}] doesn't work however

Yep so this is in agreement with what you said about bsi not being a function at all

Rather it is evaluated as an expression with the symbol t already inside it

And bsi /. t -> 0 works exactly as how I expect!

@IstvánZachar It seems nothing except explicitly inserting a position vector there works.

@Rojo your procrastination was hardly noticeable.

Hehe

I think I'm getting somewhere

yn = HtoSO3[MultQ[SO3toH[yy[t]], expQtn[- bsi ]]];

Print[yn /. t -> 0];

This works as I expect, I get the identity matrix {{1,0,0},{0,1,0},{0,0,1}}. A nice concrete value

I am however having problem with numerically computing the derivative of yn with respect to t

I can't do symbolic differentiation because some of my functions are defined in pretty obscure ways

To this end, I am using the ND function in NumericalCalculus package.

Hm, it seems like I didn't load the library correctly or something. I have a feeling it might now be working

Even though now, my ND function is highlighted in red by the syntax highlighter

omg, I finally understand the difference between = and :=

I'm trying out things like this now, for faster processing

IT FINALLY WORKS! (Sorry about monologue, needed the illusion of not being a loner)

@muzzlator happy to be of service

(I ran SetSystemOptions["CompileOptions" -> "MapCompileLength"->30] )

even after the kernel restart I had {"CompileOptions" -> {"MapCompileLength" -> 100}} again

I wonder if Wolfram Community is rigged such that searches for "matlab" don't return any hits, whereas several posts can be found via Google...

I suddenly seem to be permanently unable to open more than 2 kernels, when I'd always gotten 4 before (on a quad core)

Even after quitting/restarting the kernel or closing mathematica entirely and reopening

Ah, there we go, some of the dead kernels never got deallocated. Had to forcibly close them through the task manager.