SERVING THE QUANTITATIVE FINANCE COMMUNITY

Search found 189 matches

by Costeanu
September 9th, 2009, 2:07 pm
Forum: Numerical Methods Forum
Topic: Benchmark methods for the nearest correlation matrix estimation
Replies: 9
Views: 37223

Benchmark methods for the nearest correlation matrix estimation

<t>Hi Bourba,One way to stress correlation matrices is to notice that they form a convex subset of all the n x n matrices. Among all correlation matrices there are two distinguished ones: the identity, and the one where all the entries are one, let's call them R0 and R1. If you want to stress your c...
by Costeanu
September 8th, 2009, 6:01 pm
Forum: Numerical Methods Forum
Topic: Benchmark methods for the nearest correlation matrix estimation
Replies: 9
Views: 37223

Benchmark methods for the nearest correlation matrix estimation

<t>Norms are completely irrelevant. Unfortunately all the papers keep using norms, to make the results look more mathematical. What you care about are pricing errors. Say there is a true correlation matrix of your assets, R1. Due to the finite number of observations, you can estimate R2 which is dif...
by Costeanu
September 4th, 2009, 1:08 pm
Forum: Student Forum
Topic: is derivatives price is a price?
Replies: 40
Views: 38642

is derivatives price is a price?

<t>List,I sympathize with you. I have the same problem with complex numbers, and I'm trying hard to explain people that complex numbers are wrong. If you truly believe that square root of -1 exists then it follows that: And from here you get that negative one is equal to one, and all of mathematics ...
by Costeanu
August 24th, 2009, 1:06 pm
Forum: Student Forum
Topic: High level question: what are the main approaches to derivatives pricing?
Replies: 14
Views: 38518

High level question: what are the main approaches to derivatives pricing?

<t>QuoteSounds interesting, could you please elaborate on this? Counterparty risk means how much you stand to lose if the counterparty defaults. This loss is asymmetric, meaning at the time at the default you lose Max(derivative value, 0). This is similar to an option on the derivative value; that e...
by Costeanu
August 21st, 2009, 6:17 pm
Forum: Student Forum
Topic: High level question: what are the main approaches to derivatives pricing?
Replies: 14
Views: 38518

High level question: what are the main approaches to derivatives pricing?

<t>McGowan, Here's my take. The approaches are: 1a. Closed form solution 1b. Closed form approximation SABR, projection methods in interest rates, etc1c. Semi-closed form solution (i.e. you need a transform like FFT, or a numerical integral, or an ODE solver, or some other solver) - Heston, Jamshidi...
by Costeanu
August 11th, 2009, 7:10 pm
Forum: Programming and Software Forum
Topic: Software that can solve SDEs analytically?
Replies: 8
Views: 37724

Software that can solve SDEs analytically?

<t>QuoteThe drift-corrected Predictor-Corrector is reasonably robust. And Milstein is so-so in my experience.I guess we'll find more details in your upcoming book (which I have on order; btw, how certain is the November date?). Let me give a bit more color on the higher order schemes. In all our dif...
by Costeanu
August 11th, 2009, 6:15 pm
Forum: Programming and Software Forum
Topic: Software that can solve SDEs analytically?
Replies: 8
Views: 37724

Software that can solve SDEs analytically?

<r>Hi Jimk83,First off, nice links by Polter. In particular, if you browse away from the fourth link you get to the following link: <URL url="http://www.math.uni-bayreuth.de/~lgruene/publ/maplesde.pdfSo"><LINK_TEXT text="http://www.math.uni-bayreuth.de/~lgruen ... esde.pdfSo">http://www.math.uni-bay...
by Costeanu
July 31st, 2009, 6:31 pm
Forum: Numerical Methods Forum
Topic: Optimization problem
Replies: 8
Views: 37749

Optimization problem

<r>Hi Barny, High dimensional optimization problems can be very difficult, but for some classes of problems there are good algorithms. Before you start learning everything there is to know about optimization, you could give Excel a try. It has a nice solver (some people hate it though ...), but it's...
by Costeanu
July 31st, 2009, 6:19 pm
Forum: Programming and Software Forum
Topic: boost, UBLAS, coordinatewise multiplication
Replies: 10
Views: 38987

boost, UBLAS, coordinatewise multiplication

<t>There is a way, namely use element_prod. Not sure it helps though. Here's some code you can play with. Try various combinations. To me it looks like there is no benefit of vectorizing code, as opposed to using straight loops. However, it's possible the compiler notices some shortcuts (for example...
GZIP: On