Wolfram Mathematica

5:39 AM

Any hints how to approach a problem that would naturally show itself as integer nonlinear programming (polynomial fitness function) with constraints?

Nasser

5:54 AM

I must be the only person in the world running CDF player 10 and firefox 36+ version on windows 7. Since I can't run CDF's inside Firefox browser version 36 and above after updating to version 10 of CDF player and I see no one else is complaining about it. This is very strange. Either no body cares about CDF's any more, or no body runs windows 7 anymore and uses CDF's. community.wolfram.com/groups/-/m/t/449591?p_p_auth=KKGXGHw5

Heh. Parameters are constrained precisely to permutations of all values in the region. I think I should just turn into trying out MCMC in this case.

6 hours later…

Yyrkoon

11:25 AM

Hi all,
this may be really trivial but I've the following issue:
I'm generating automatically a list of list made of conditions:
L={{a>1,a<1},{b>1,b<1},{c>1,c<1}}
the lengt of this list may vary depending on the input I gave. So now is 3 sublist but can be more
I want to generate all the possible combination (without repetition) of these 3 sublist.
I know that with
Outer[List,L[[1]],L[[2]],L[[3]]]
this gets done, I tried to use the levelspec of Outer
Outher[List,L,2]
but does't work, do you guys have any suggestions?

Kuba

11:57 AM

@Yyrkoon Try
Outer[List, ##] & @@ L

or

Tuples[L] (*need to add Partition if you want the same structure*)

2 hours later…

balboa

1:50 PM

Hello,

sorry I didn't know that enter sends a message, my question is coming in few minutes :)

I want to draw a power spectral density plot of audio signal that I record. I've found out about PowerSpectralDensity[] function, but I'm not sure whether I should take Fourier[] of a signal first, then PowerSpectralDensity[] or this PowerSpectralDensity does fft by itself?

I was thinking of asking this as a question, but it would most probably be closed, since I admit that this could be researched... but I'm not yet familiar with underlying mathematics so it makes it hard...

3 hours later…

Pickett

4:53 PM

@balboa Its input argument is a time series so no, you should not apply the fourier transform by yourself first.

Rolf Mertig

5:46 PM

@Murta Great! Enjoy. Poor me has to work in Mexico City. Though the climate is great!

2 hours later…

7:22 PM

@kirma I went ahead and visualized the cycles on the cube and octahedron.

@MichaelHale Pretty fancy :)

Did you look at possibility - or impossibility of using only limited number of "face tiles"?

Only for 1 tile on tetrahedron. I know that's impossible. The hasty code I wrote won't scale to the cube though.

I'm not surprised.

@kirma The icosahedron isn't too bad. I'm afraid to try the dodecahedron though.

I think dodecahedron had something like 1280 Hamiltonian cycles...

7:31 PM

Yeah. Oh well. Might as well try.

Length@FindHamiltonianCycle[PolyhedronData["Icosahedron", "SkeletonGraph"], All]
-> 1280

I'll give it 20 minutes.

Only needed 4-5. Definitely a mess though.

Not quite as tractable as tetrahedron, I must say. :)

7:44 PM

Hi.

Can somebody give me an advice on fitting method?

I have some experimental points of a harmonic oscillations and I want to find its period.

And it looks like I do something wrong.

@mikeonly I'd start by putting a constraint on omega like you put on A.

8:04 PM

@MichaelHale Ok, so constraints helped. Though I wonder why does Mathematica behave like this?

Consider Nyquist-Shannon sampling theorem...

@mikeonly because it found a great fit, and it doesn't really understand your sampling in the sense you implicitly do it visually...

@kirma Oh, ok. Thanks.

@mikeonly I guess you could try to perform NonlinearModelFit also on TimeSeries in v10... maybe it would be more intelligent. Or maybe not.

@kirma Thanks, I'll try. Are there any good approximation tutorials for Mathematica?

8:20 PM

@mikeonly I fear I'm not particularly well-equipped to answer that question. Yet, Mathematica Stack Exchange and Mathematica documentation go a long way...

@kirma Yes, of course. But as far as approximation methods are considered, Mathematica Documentation is not comprehensive at all on that.

Q: Why do some question pages have the first tag in the document title but others don't?

halirutan

8:33 PM

@mikeonly And have you considered using correlation (in the time domain) or fft (in the frequency domain)?

Meta

8:53 PM

Compare the document title (typically shown on the tab for browser with tabs) for these questions: programming - How can I use NestWhileList instead of While in this case? equation solving - Elimination returning true instead of correct answer Plotting difference between a function and its inve...

support