Reddit downvotes are random... the system automatically inflates upvotes and downvotes proportional to the velocity of votes so that people don't think they are influencing it in any way... the score is correct though
@rcollyer the edit has the code for it... at the very end is a viewer with next/prev... just remove those bits and take the core grid layout and use Table
for my general, I was talking for 2 1/2 hours straight. Note: the presentation was only supposed to be 45 mins, but I kept getting asked questions ... Really could have used a beer afterwards.
@belisarius technically, it is the candidacy exam. After you pass it, you technically have the equivalent of a masters. At my school, it is referred to as the general exam. We also have a qualifier which weeds you out in the first two years.
But I still have time for some Mathematica. Actually Wolfram is coming to my employer today to present the Mathematica in Finance Platform. I looked at the videos on the weekend, but I can't work out how it's different from normal Mathematica.
They did mention Reuters and Bloomberg data feeds - that would be really nice.
@belisarius yes, more or less. Except, the names for the first two change wildly depending on which university it is (in mine, even different departments have different names for them)
@belisarius yes. With the PhD being a second oral exam, with a similar format to the first: 45 - 60 minutes public lecture, unknown amount of time being grilled in private by one's committee, and finally an unknown amount of time waiting in the hall while your committee debates the merits of letting you pass. Then champagne!
@RM that's the case here as well.
@Verbeia probably not much different. Just the additional feeds.
I doubt they'll NDA us given it's in the wild (and I don't think we can be bound by them anway). So I can pass on my impressions if anyone's interested.
@RM I remember some comment done by someone while reading an einstein bio, remarking that at that time German unis had only one exam ... the last and only one. You studied for six years, and then you go to a committee whose members decide if your last six years were useful ...
I have a question about BinCounts and BinLists; basically I would like to do a histogram of values of x; but I would like each hit to be weighted by the corresponding value of y. My motivation is to rewrite the mathematica code MorphologicalEulerNumber which seems to have a small statistical bias.
Histogram doesn't have any built-in support for weighted data, although it's an interesting idea, and most of the binning algorithms should be amenable to working with it.
That being said, here's a WeightedHistogram function, with some feedback from Andy Ross. It accepts
weighted values (i...
I really need to pay attention to what site I'm on, I nearly blasted this guy for re-posting a now deleted question about Octave. Note: the question is on Computational Science. Otherwise it is identical to the one he posted earlier.
It was remarkably hard to work out what is different: Bloomberg Link, an Excel link, and automatic use of CUDA for derivatives pricing instead of calling CUDADerivative or whatever the function is explicitly.
Apparently they are bringing a lot of time series functionality into the kernel in version 9. At one point they said that it was also in the finance platform, but I haven't been able to find it in the documentation on the web.
@acl Oh, that. The thing in my blog doesn't exactly answer the question in the title, and OP has not been responsive on what he wants or expects. Yeesh.
(I've also been working on an inversion method that only needs evaluation at real arguments. As expected, it works, but not as good as methods that allow complex evaluation...)
@Verbeia So there are two products... I can't understand why they didn't add it as a package. Somehow the way they did seems to have more maintenance... What I understood is that they are integrating the Time Series add-on (wolfram.com/products/applications/timeseries) to M9. I really wished they also integrated an Excel link to standard M9...
@PFonseca Maybe one complete program sells better than Mathematica plus a package. Especially since Mathematica isn't usually associated with the financial market.
But, the way I think about it, Mathematica is so big at this point that it might make sense to split it up in separate packages. Not everybody needs to do image manipulation, for instance.
@J.M. An enormous percentage of the business and industrial world is based exclusively on Excel. This means that they have never heard of Mathematica, and it would be much easier to "sell" the mathematica idea if there was a link into the known world...
@JM Wolfram's strategy is to keep integrating packages into the core product. This may be based on the idea that it differentiates them from their main competitor, MathWorks.
@Heike Hmm, I see what you mean. But, don't you think a case could be made for modularity, as opposed to the current "Japanese toilet" setup of Mathematica?
To use the image manipulation functions again as an example: if you'll be using them a lot, by all means, install them. But if you don't need them, you could get away with a Mathematica installation that doesn't take up that much space...
@PFonseca What does the Excel link do? I think creating a basic integration would not be so difficult (at least on Windows where we have .NETLink). But no one in the sciences (theoretical) uses Excel, so never really needed it beyond importing a few spreadsheets...
Sure, nobody does serious scientific computing in Excel, but if Wolfram wants to cater to the business-types, they really should look into spreadsheet interfaces...
No to modularity !!! I've learned a lot from Mathematica, on themes that I didn't even thought they existed. All because of it having everything inside.
@Szabolcs The excel link lets you call Mathematica functions directly from Excel (and vice versa). I don’t particulary appreciate Excel (I was a heavy user for long time, with lots of programming done in their VBA). But my problem with Mathematica (or MATLAB, or etc), is that most of the companies in my world only see these softwares as huge complicated scientific minded things. If only there was a better link to the real world, I could more easily bring Mathematica into it!
@Heike I agree that the home use is very well priced. But to have a budget to pay the European price (never understood why such a difference to the US price), when I'm the only person that knows how to use it... it was very very hard.
Anyway, the price differences are never greater than a factor of two between regions. The salary differences are much greater. I hate this about software.
Right now in Romania a 400 EUR net salary is pretty good. The VAT is now 24%. That means that the price of the home edition comes very close to that not-so-bad salary.
@Heike I don't necessarily mind there being many different packages, after all I program in c++ and it is all "packages." My issue is having to pay for them.
@rcollyer I agree, having to << them first is fine, so long as I don't have to pay for them. Having mma is a bit like having a huge library of books available, where you can skim through whatever catches your eye.
Sometimes you find interesting things like this.
(and sometimes you just waste days or weeks, but it's still fun)
Two things.
First, a minor point: if you rewrite your compiled function as
tweakC = Compile[{{pixel, _Real, 1}},
Module[{m},
m = Mean[pixel];
Which[
m <= 0.3, pixel*1.5,
m >= 0.85, pixel*0.8,
True, pixel]
]
]
then the ImageApply bit is 20 times faster (due to n...
although some operations tend to be a bit slow in mma :)
@Heike Oscillatory around the $\pi/4$ line (hence the Fresnel integrals), and exponentially growing on the imaginary axis (hence the imaginary error function Erfi[]).
Just to further show off the nice oscillatory behavior of the error function, here's a neat way to see a clothoid: ParametricPlot[Through[{Re, Im}[Erf[(1 + I) t]]], {t, -5, 5}]
I recently discovered that Check does not stop execution of the supplied expression when a Message is thrown. I wrote a version that does so (complete with same semantics), any objections to me posting a question/answer pair?
@Heike yes, but the music is specifically so no one outside of the bathroom can hear what is happening inside. The rest is along the lines of: "Oh my God, what did I just press!"
@JM my mother-in-law has two combination dvd recorder/vcrs which have a full suite of buttons available in a hidden panel. Almost as much as a toilet. :)
@Heike exactly. No one can hear you scream in a Japanese toilet!
for example, a BEC in a deep optical lattice, modulated by a periodic potential with a period incommensurate with the lattice, becomes immediately localized. it's pretty hard to study
eg the excitation spectrum is fractal (or would be if you had an infinite system)
this was actually implemented cleanly in an experiment:
(This is the third in a series of posts about A New Kind of Science. Previous posts have covered the original reaction to the book and what’s happened since it was published.) Today ten years have passed since A New Kind of Science (“the NKS book”) was published. But in many ways the development that started with [...]
@JM wavelets have been on my list of things to understand for quite a bit of time for both practical and recreational reasons. any suggestions for a short introduction?
Unfortunately, the bibliography I compiled isn't with me at the moment... so that and the compilation of Daubechies's lectures are the only ones I remember off the top of my head.
Anyway: if you're comfy with Fourier analysis, wavelets shouldn't be too troublesome...
@JM nothing concrete. I was just wondering how you did it (I wrote some code to do this too, to check some analytical results I had, but in the end I was less convinced of the accuracy of the numerical code than of the analytical results it was meant to confirm...)
Okay, sure. :) But, "I was less convinced of the accuracy of the numerical code", well... as I mentioned in that blog post, numerical inversion's a tricky business.
@JM no, and even if I had, I would not have put in the effort. I am reasonably sure of my analytical result and can check it again by doing the calculation in different ways. I am personally much more likely to screw up a numerical calculation than an analytical one, so I avoid them if possible
In Jens answer to the HermitianQ question, he makes the statement that splitting a matrix into a Hermitian/Skew-Hermitian pair like I did is non-unique. But, I'm having trouble coming up with an example. Do either of you know of one?
Anyway: it wasn't mentioned in that thread, but there's also the option of constructing a positive (semi)definite Hermitian matrix by multiplying some matrix with its conjugate transpose...
@Heike you're right. That's two equations and 4 unknowns. However, the additional equation: b[i, j] - b[j, i]* = 2 m[i, j] would fix the values of $B$.
@Heike and that's what I'm getting at. There are numerous matrices which can be used to produce A and M, yet A[B] and M[B] are unique for a given B. Then, is that enough to fix A and M such that B = A + M, uniquely?
I will show a method that is conjectural, though i believe it is correct. It differs from the more common approach of using (quadratic) birational transformations to force singularities to be double points. More a detailed approach to that, see Madelina's response. Also I cover the exact case, al...
(still reading)
I'm pretty sure I had a comment on this and now it's gone. Why?
@PFonseca Yes, the presenter did say they would integrate Time Series into M9. He mentioned Excel integration but didn't show it, and it wasn't in the documentation for the Finance Platform.
Philosophically they do believe in rolling things into the kernel. He explicitly mentioned R and Matlab as products that have lots of addons that don't work together.
ok math wizzes, I'm trying to find a way to show that -Sum[Log[Tanh[a [Pi]/2 (2 n + 1)]], {n, 0, Infinity}] ==Sum[((2 n + 1) Sinh[a [Pi] (2 n + 1)])^-1, {n, 0, Infinity}]; (a > 0 ). I can plot the difference, and it is negligible, but I can't find a good math reference for sums of hyperbolic functions
@Verbeia the problem with the Excel add-on is that, as far as I know, it doesn't belong to WR, and so, it's not something they can simply add for free, when they feel it's mature enough for global distribution...
Hi, I'm looking for the question where someone asked how to type [[ ]] faster, but my search skills seem to be failing me, I can't find the question anymore
Is it possible to program the Front End to automatically format double square brackets without having to type Esc [[ Esc each time? It's awful to have to type Esc 4 times for each Part expression, and even more annoying to visually parse the unformatted double brackets.
See also this entry in Ma...
@RM OK, I figured it out, you just need to not edit the .tr file in Mathematica, it's auto-replacing the double bracket character with two brackets upon saving.
@PFonseca we tried to get them to talk a bit more about it, but they had pitched the presentation at financial risk management and portfolio modelling, which is not what we do.
Histogram doesn't have any built-in support for weighted data, although it's an interesting idea, and most of the binning algorithms should be amenable to working with it.
That being said, here's a WeightedHistogram function, with some feedback from Andy Ross. It accepts
weighted values (i...
If you search the docs for "autocorrelation", ListConvolve turns up. There's also ListCorrelate. Does this work for you?
