<t>A friend argues that neural nets are just the greatest tool and that most financial researchers deprive themselves of using this panacea that "can solve any nonlinear problem". That's a bunch of huey, but I need to substantiate my argument better. Hence, I ask. Has there been any important papers...

<t>Thanks for the links, I got some reading to do. In the mean time, can any of you please criticize the following approach to estimating a covar matrix by GARCH(1,1).1. Estimate GARCH(1,1) models for the diagonal elements. That is, sigmahat(i,i)(t) = GARCH(1,1|historical data )2. With these estimat...

<t>I'm looking for some literature on estimating the covariance matrix of a portfolio of equities via GARCH techniques. This is basically a learning exercise.I've done some google searches but I'm getting materials that are too advanced. What's a good source for this kind of application. I'll be hap...

<t>Hi Rogermc, thanks for elaborating further on those items. This is a very clever approach:"you create a benchmark that is short your portfolio (100M) and long the benchmark (150M). You can then take your existing portfolio, and solve for the minimum risk solution setting the optimal portfolio equ...

<t>Thanks Rogermc! I will get Grinold and Kahn per your suggestion. Below are some points I would like to go over, just to reenforce my understanding.1. It seems that MFM could use something like Principal Component Analysis (PCA) to reduce the dimensions of VCV matrix say from a 3000x3000 to i.e., ...

<t>The Mean Variance theory that I learned is about finding weights of an existing portfolio, it says nothing about how to pick stocks. If I understand this correctly, it means that a holding can get a positive or negative weight (go long or go short) but the stocks have already been chosen. CAPM is...

Nevermind, I reframed my problem in a more elementary way; one were P(B) = B. Still if any of you of have some insights about problems such as the above then I'm curious to read about it. Thanks!

I am seeking for some leads/theory/links on how to solve a matrix equation of the form A*P(B) = C. Here A and C are known square matrices and "P(B)" is an unknown polynomial in B of degree n. How do I find the "best" P(B) in some appropriate sense OLS or something else?

I have both Glasserman's and Jackel's book. I recommend Glasserman. Jackel's book is good but spread too thin (in case it matters the math is less demanding than Glasserman).

Thanks so much, I didn't know about them. The price is steep for my budget, which one is the most practical?

<t>Just wondering if there's a book about fitting stochastic processes to data. I imagine it as sort of an econometrics book but with emphasis (or at least a few chapters) on the processess that are important in mathematical finance. For example, one that considers a Levy jump-diffusion process with...

Aaron, thanks for the extensive and clear explanation! You rock, man!

<t>QuoteOriginally posted by: AaronThere are several things you could be talking about here.... If you estimate the parameters of your trading system based on historical information, and backtest it on the same information, your performance measurement is biased high. That is indeed very interesting...

<r>If the question is, that a backtests yields probability of win = 0.5 then why doesn't the next draw have probability 0.5? 1. Well, forget backtesting for a second. The answer is maybe it does or maybe it doesn't. It depends on your model but doesn't necessarilly have to do anything with time. pro...

Not sure which one of the DiPrimas's is better. The one that includes boundary value problems sounds more interesting. Monasteb's suggestion is also excellent.