Skip to main content
24 events
when toggle format what by license comment
Jul 3, 2012 at 23:21 comment added tau1777 I do believe I have the problem of my initial matrix being close to singular. How can I get the second to last singular value instead of the very last. I've tried SingularValueList[mat /. \[Kappa] -> x,-2]]], but I don't think it is working.
Jul 3, 2012 at 2:50 comment added tau1777 @Jens, I did not mean to get to far off topic, since this is a Mathematica related forum. I guess my main Mathematica related question that remains is with the SingularValueList command? As I was saying in "6th-Edit" I'm still a little stuck about how to interpret the results plot. Other than that I guess you are saying that Det[mat] should be zero, because we are using a singular value, but it is not b/c Mathematica is calculating the Det differently than it calculated the singular values. Thanks again for all your help.
Jul 3, 2012 at 2:36 comment added Jens This doesn't sound like the remaining problems are related to Mathematica, but to the setup of your matrix problem. I can't think of anything to say about that, except that it doesn't sound numerically stable. Many matrix decomposition methods attempt to renormalize the rows and/or columns before doing the decomposition, whereas that may not be possible for the determinant calculation. That could explain the difference in the results.
Jul 2, 2012 at 22:07 history edited tau1777 CC BY-SA 3.0
added 1287 characters in body
Jul 1, 2012 at 6:26 history edited Mr.Wizard CC BY-SA 3.0
added 12 characters in body
Jun 30, 2012 at 21:00 comment added tau1777 @Jens, WOW. I officially think you are a Mathematica wizard. Sorry to have bothered you about this again, it seems like I really need a better understanding of what the SingularValueList command does and what SVD does in general. I will have to present these results to my advisor so knowing exactly what my code does will be very important. But more importantly I would really like to know what my code is doing. If you have any suggestions about where to learn about this material, I would be happy to know. Thanks again.
Jun 30, 2012 at 20:56 history edited tau1777 CC BY-SA 3.0
added 191 characters in body
Jun 30, 2012 at 20:46 comment added Jens Again a stab in the dark: in your Edit 4, the jumps indicate that some of the small singular values are thrown away occasionally and then picked up again. You could try replacing SingularValueList[m] by SingularValueList[m, Tolerance -> 0] which will ensure that all singular values are kept.
Jun 30, 2012 at 19:51 history edited tau1777 CC BY-SA 3.0
added 2648 characters in body
Jun 28, 2012 at 17:59 history edited tau1777 CC BY-SA 3.0
added 243 characters in body
Jun 28, 2012 at 16:54 comment added rcollyer Your example matrix does not appear as if it can be rewritten that way. But, it may be separable into multiple updates. That said, it isn't trivial, and I don't know if it can be done automatically, or a case-by-case analysis is required. I think the latter is more likely. If you can make use of it, I'd be interested in hearing about it.
Jun 28, 2012 at 16:51 comment added tau1777 @TomD, Yes there should be finite roots, as I'm really try to reproduce research. Also I did get finite roots for a small version of my problem. Making the problem much larger is where the trouble starts.
Jun 28, 2012 at 16:50 comment added tau1777 @rcollyer Thanks for the suggestion. I'm not sure if I can rewrite my problem like this, since there are terms like $\kappa$a(1,1) etc, but this is certainly something I will look into.
Jun 28, 2012 at 16:45 comment added rcollyer Finding the determinant of a symbolic matrix is very, very slow, and it gets bad fast. On my machine, I cap out at $11\times 11$ matrices (which take 4 secs) before Mathematica begins to chew up a lot of memory. Since you have a numerical matrix, $M$ plus a symbolic matrix of a single variable, $\kappa A$, it may be worthwhile to see if you can recast $A$ as a rank-1 or rank-l ($l\le n$) update. Then you can use the matrix determinant lemma. As this is non-trivial, I opted to just comment instead of providing an answer.
Jun 28, 2012 at 15:41 answer added Daniel Lichtblau timeline score: 6
Jun 28, 2012 at 7:32 comment added user1066 Are you sure kappa has finite roots? Plot[Det[t2], {\[Kappa], -10, 10}, AxesOrigin -> {0, 0}] (or Solve[Det[t2] == 0, \[Kappa]])
Jun 28, 2012 at 7:01 comment added tau1777 @nightowl, I just checked with $\kappa$ instead of "x" it did not make a difference. I didn't think it would since I defined the function "f" it should have been able to take any variable as it input. Thanks for help, at least that part is troubleshot.
Jun 28, 2012 at 6:43 comment added night owl Should it be FindRoot[f[\[Kappa]],{\[Kappa],.5}]? There was an x. Could this have generated the error messages?
Jun 28, 2012 at 5:17 history edited tau1777 CC BY-SA 3.0
added 1825 characters in body
Jun 28, 2012 at 3:22 history edited Verbeia CC BY-SA 3.0
added 6 characters in body; edited tags
Jun 28, 2012 at 3:20 history edited tau1777 CC BY-SA 3.0
added 318 characters in body
Jun 28, 2012 at 3:14 history edited tau1777 CC BY-SA 3.0
added 318 characters in body
Jun 28, 2012 at 1:56 answer added Jens timeline score: 10
Jun 28, 2012 at 1:41 history asked tau1777 CC BY-SA 3.0