SERVING THE QUANTITATIVE FINANCE COMMUNITY

## Search found 48 matches

November 28th, 2010, 5:53 pm
Replies: 12
Views: 23301

interesting, thanks for the paper Alan.thanks a lot guys, I think this kills it. I will post once I get through all of it.
November 28th, 2010, 5:10 pm
Replies: 12
Views: 23301

<t>Oh, its just that I am not sure if just setting the first derivative to zero wrt the parameters kappa, theta, sigma1, for lets say the MLE estimation for the OU process is enough. How can we be sure that it indeed is the global maxima that we are looking for.Also if the above works, we can basica...
November 28th, 2010, 2:37 am
Replies: 12
Views: 23301

ok another question Alan, I was reading more abt the OU fitting, seems people seem to use the MLE. I did not quite understand that coz the form of the function is definitely not convex.
November 28th, 2010, 2:05 am
Replies: 12
Views: 23301

gotcha!, thanks Alan, will ask once I complete more steps.
November 27th, 2010, 6:28 pm
Replies: 12
Views: 23301

<t>Thanks a lot for the note Alan, so in nutshell basically we can regress both the equations to estimate the parameters. For the alpha equation also we can do a linear regression. Presumably alpha is the expected value or the mean of all the deltaY's i.e E(Yt-Yt-1)Also for sig1, u mean estimation f...
November 26th, 2010, 8:45 pm
Replies: 12
Views: 23301

<t>Hi All,Did not really get any response on the student forum, so posting here.I am trying to fit the following model I have time series for X and Y, the two epsilons are correlated brownians, with some correlation rho.Please help, I am thinking MLE but the form of the equations would make the para...
November 25th, 2010, 7:29 pm
Forum: Student Forum
Topic: parameter fitting question.
Replies: 0
Views: 21461

### parameter fitting question.

<t>Hi All,I am trying to fit the following model I have time series for X and Y, the two epsilons are correlated brownians, with some correlation rho.Please help, I am thinking MLE but the form of the equations would make the parameter estimation non convex. Parameters to be estimatedtheta, kappa, s...
November 21st, 2010, 5:13 pm
Topic: commodity structural models
Replies: 3
Views: 24817

### commodity structural models

November 21st, 2010, 1:53 pm
Topic: commodity structural models
Replies: 3
Views: 24817

### commodity structural models

November 21st, 2010, 11:17 am
Forum: Student Forum
Topic: parameter estimation of commodity structural models.
Replies: 0
Views: 21482

### parameter estimation of commodity structural models.

October 29th, 2010, 12:05 am
Topic: crude oil term structure modeling.
Replies: 9
Views: 29506

### crude oil term structure modeling.

<t>Hi, I am interested in getting to know some commonly used term structure models used in the industry. In general I found two parallel approaches, work by Schwartz et al that focuses on stochastic modeling of the relevant factors, and some papers on PCA modeling. Did not fnd something that tries t...
October 24th, 2010, 10:09 pm
Forum: Student Forum
Topic: optimization problem-need help.
Replies: 11
Views: 22919

### optimization problem-need help.

Hmm gotcha, thanks again for your help Alan. It feels great to be in a position to discuss things with stalwarts like yourselves
October 24th, 2010, 9:54 pm
Forum: Student Forum
Topic: optimization problem-need help.
Replies: 11
Views: 22919

### optimization problem-need help.

Oh I hear you, I guess you did say that it has atleast one zero eigen value. Ok. I guess we're good if all eigen values are strictly positive, >0.Is there a way to generalize the solution characteristics depending on M?
October 24th, 2010, 9:52 pm
Forum: Student Forum
Topic: optimization problem-need help.
Replies: 11
Views: 22919

### optimization problem-need help.

Sorry I guess I did not quite follow that. I just thought that for any psd matrix Mz'Mz >=0for any vector z
October 24th, 2010, 9:44 pm
Forum: Student Forum
Topic: optimization problem-need help.
Replies: 11
Views: 22919

### optimization problem-need help.

Thanks for the note again Alan. Isn't it the case that if M is positive sd, then we are in a good shape, coz then the problem is strictly convex.

GZIP: On