<t>QuoteOriginally posted by: NHmmm,Ax = bOK, so we're rotating (by A) an unknown vector, x into a known vector b assuming the problem is in SO. Unless A is hermitian symmetric, I'd say your completely f'd.I don't see anything that can be minimized in a mean squared sense that might help, except per...

Send my a private message with your email and we can discuss this topic in more detail. Maybe, I can provide you with test data ... I am not aware of something like a standard data set. When it comes to testing, read the paper of Bliss, "testing term structure estimation methods"

<t>In this thread, two topics are adressed. As I am no expert in numerical linear algebra I think I am of no great help with topic no. 1 - ill conditioned linear systems. Usually I follow the advice from the very good book of Golub, van Loan: matrix computation or use regularization methods of the p...

<r>QuoteOriginally posted by: JanDashOne trap in using the SVD is that the minus sign for a negative eigenvalue is hidden in the right eigenfunction. Moreover, the Numerical Recipes code can give incorrect answers if two data series are identical (this happens because sometimes the same data series ...

The recent book of P. Glasserman: Monte Carlo Methods should give you an excellent overview of how to compute derivates (Greeks) in simulation

<t>I more or less completely agree with PJ.I think the reason why Cholesky is still that famous is for two reasons:1) it is very easy to understand and very easy to program2) if you have well-conditioned matrices, then it's robust and it's way faster than spectral decomposition.A very good source fo...

<t>Some partial answer can be found in the bookBouchaud, Potters: Theory of Financial Risk and Derivative PricingFurther, the articles of the authors on covariance matrix estimation seem to be informative, at least.I am not sure, whether Uppal (or someone else?) came up with a similar shrinkage idea...

Ant Farm Financial Engineer - mh, its abbrevation AFFE could give you a warm start in the German job market (affe = monkey, and we all have read the book monkey business, haven't we?)

<t>At the moment, I cannot see, what ACO will add. Standard Metaheuristics as genetic algorithms, simulated annealing, threshold accepting - only to mention a few - have been applied with mixed success. Can you give me any reasons why you think ACO will outperform these on tough real world optimizat...

Unfortunately, this is terrible true. Right you are.Nevertheless, once in a while I can pick up a good idea ... or improve my understanding (that's why I'm around Wilmott's)

<r>QuoteOriginally posted by: ErrrbI saw this thread and it made my day. It looks like there are people who have more than enougth spare time to spend on useless shit like writing this type of PhD works.I'm not sure, whether its my bad English (could be) or if your statement has two meanings (depend...

<t>I'm sorry, maybe I did confuse you a little bit, I was not talking about the leverage factor (which in fact is something from 8 to 10 for well-diversified portfolios), but about the "safety factor" used for calculating risk capital based on a simulation of the gap risk.In fact, in the Luxemburg m...

<r>Thank's a lot, donyoshi. Gave me one or two new insights.Unluckily, the situation in asset management is more difficult as in banks, I suppose. You cannot simply take the risks on your trading books, there, it really becomes business risk. And further, you cannot trade in variance swaps (or whate...

<t>QuoteOriginally posted by: donyoshiliquidity is "relative" and as you mention Ichiban, one tends to use CPPI with less liquid underlyings such as HFs and FoFs as compared to delta hedging a call option, since a big part of the liquidity risk is passed through the the client with a cppi (ie bank i...

<r>Sorry if I sounded insulting, I didn't want to. I am only a little bit fed up with people giving advice on a (almost) zero-knowledge basis. Sooner or later, someone will have to pay for it. And if we talk about CPPI, it already happened twice: stock market crash in 1987 and the last two/three yea...