SERVING THE QUANTITATIVE FINANCE COMMUNITY

Search found 9882 matches

by Alan
August 26th, 2003, 4:28 pm
Forum: Technical Forum
Topic: Measures of downside risk - semi standard deviation
Replies: 8
Views: 193811

Measures of downside risk - semi standard deviation

<t>QuoteOriginally posted by: sudhakar682Hi,Thanks for the suggestion. I had done the search prior to posting my question. I am particularly looking for any articles that explain the superiority/inferiority of using downside risk measures over standard measures.thanks,SudhakarI will plug an ancient ...
by Alan
August 22nd, 2003, 3:51 pm
Forum: Student Forum
Topic: Daily Volume Seasonal
Replies: 3
Views: 189467

Daily Volume Seasonal

<t>Ok, here's a model:V(t) = a1 D(1) + a2 D(2) + ... + a6 D(6) + b V(t - 1) + c (R(t-1))^2 + error,where V(t) is the volume in the hour surrounding time t, D(i) is a dummy variable equal to 1 at hour i of the trading session and zero otherwise R(t-1) is the log price return in the previous hour.find...
by Alan
August 22nd, 2003, 3:09 pm
Forum: Technical Forum
Topic: Random Portfolio Weights' Algorithm
Replies: 11
Views: 190574

Random Portfolio Weights' Algorithm

<r>QuoteOriginally posted by: gregoriosTo pburns:You're correct, in the case you are comparing your portfoilio against a benchmark.I need to see the whole cloud of points not just the efficent frontier line.To jazzjulien:Did you somehow dispatch only part of your reply or are you not willing to shar...
by Alan
August 21st, 2003, 2:55 pm
Forum: Numerical Methods Forum
Topic: integrals over multivariate dist.
Replies: 11
Views: 190035

integrals over multivariate dist.

<t>QuoteOriginally posted by: twHi,Does anyone know of a smart way to perform integrals of the formWell, I suspect that you really haven't solved your problem in a smart wayif the answer is a 100-D integral. So, for the best advice, I would post theoriginal problem and perhaps someone will see a muc...
by Alan
August 20th, 2003, 7:46 pm
Forum: Student Forum
Topic: GARCH(1,1) and Markov process
Replies: 10
Views: 189874

GARCH(1,1) and Markov process

<t>QuoteOriginally posted by: dummycarHi, take "Pricing Hang Seng Index options around the Asian financial crisis - A GARCH approach" as an example. Duan used empirical martingale simulation method. However, the random variable, e, in the stock price dynamic is an element of an iid standard normal r...
by Alan
August 20th, 2003, 6:48 pm
Forum: Student Forum
Topic: GARCH(1,1) and Markov process
Replies: 10
Views: 189874

GARCH(1,1) and Markov process

I don't see that one at his home page for download -- what's the URL?
by Alan
August 20th, 2003, 3:46 pm
Forum: Technical Forum
Topic: Random Portfolio Weights' Algorithm
Replies: 11
Views: 190574

Random Portfolio Weights' Algorithm

For N weights, why not just draw N-1 random numbers U_i, uniformly from [0,1].Order them from smallest to largest, include 0 and 1 and take the differences.
by Alan
August 20th, 2003, 1:17 pm
Forum: Student Forum
Topic: Simulating CIR Path
Replies: 10
Views: 190622

Simulating CIR Path

I suspect your simulation will work if you reflect off the origin -- see this thread
by Alan
August 20th, 2003, 12:56 pm
Forum: Student Forum
Topic: GARCH(1,1) and Markov process
Replies: 10
Views: 189874

GARCH(1,1) and Markov process

You're right -- that wouldn't be a self-consistent way to do GARCH optionpricing. It wouldn't yield a smile, for example. For the right way, see work by j.c. duan/toronto, for example.
by Alan
August 20th, 2003, 12:37 am
Forum: Student Forum
Topic: GARCH(1,1) and Markov process
Replies: 10
Views: 189874

GARCH(1,1) and Markov process

Hi,If we are reading the same paper, then in theline just above that result, you will see that he is derivingk = E[e^4]/(E[e^2])^2 - 3, which is the _excess_ kurtosis.
by Alan
August 20th, 2003, 12:22 am
Forum: Student Forum
Topic: quadratic variation and jumps
Replies: 4
Views: 189461

quadratic variation and jumps

<r>Suppose the the continuous part of theprocess was dX(t) = sig(t) dB(t), where dB(t) is a Brownian motionand sig(t) is a stochastic volatilityBy pathological, I was thinking that perhaps there would bea problem with my argument if sig(t) exploded; i.e. went to infinity in finite time.This generall...
by Alan
August 19th, 2003, 7:25 pm
Forum: Student Forum
Topic: quadratic variation and jumps
Replies: 4
Views: 189461

quadratic variation and jumps

<t>I assume you mean you can sample at any frequency. Then, unless thecontinuous part of the process is pathological, I'd say yes, the jumps are alwaysvisible because you can do some experiments.Experiment 1: Measure |X(t) - X(t - eps)| at t fixed and smaller and smaller eps. If it goes to zero ther...
by Alan
August 19th, 2003, 6:09 pm
Forum: Student Forum
Topic: GARCH(1,1) and Markov process
Replies: 10
Views: 189874

GARCH(1,1) and Markov process

<t>Non-Markov models can often be turned into Markov ones by adding state variables. That's what's going on here. If the state only consists of the price history P(t),then you're right it's non-markov because you need the whole history of prices to determine whathappens next. But, once you add the m...
by Alan
August 19th, 2003, 1:39 pm
Forum: Student Forum
Topic: A Grounding in Stochastic Volatility
Replies: 16
Views: 190682

A Grounding in Stochastic Volatility

Johnny, thanks for the plug.I sometimes see a used copy available from amazon.Otherwise, I am working on a vol 2, and when I print that,say by year-end or next spring, then I'll also do a second printing ofvol. 1.regards,alan
by Alan
August 12th, 2003, 9:42 pm
Forum: Technical Forum
Topic: SPX vVol
Replies: 12
Views: 190530

SPX vVol

<t>QuoteOriginally posted by: mmantaAlan,I prepared a small spreadsheet using some volatility estimators to see if I can match your observations with the GARCH type models.Given that I don't have 5-min intraday data, I used daily prices.Are those calculations correct?As I understand your later expla...
GZIP: On