SERVING THE QUANTITATIVE FINANCE COMMUNITY

Search found 6292 matches

by Paul
October 5th, 2001, 7:52 am
Forum: Technical Forum
Topic: Query on deriving Principal Components on Excel.
Replies: 32
Views: 191854

Query on deriving Principal Components on Excel.

Covariance matrices will be symmetric so it should work. Try it using the same spreadsheet that went wrong before.P
by Paul
October 4th, 2001, 10:46 pm
Forum: Technical Forum
Topic: Query on deriving Principal Components on Excel.
Replies: 32
Views: 191854

Query on deriving Principal Components on Excel.

Er, doesn't the power method only work for symmetric matrices?P
by Paul
October 4th, 2001, 3:48 pm
Forum: Technical Forum
Topic: Query on deriving Principal Components on Excel.
Replies: 32
Views: 191854

Query on deriving Principal Components on Excel.

You are going to find this soooooooooo useful...I knocked up a program in Excel to do PCA in a very non elegant but transparent way. Just as I got the second eigenvalue (2) Excel crashed...before I had even looked at the eigenvalue! When I'm feeling strong I will have another go!P
by Paul
October 3rd, 2001, 7:27 pm
Forum: Technical Forum
Topic: Query on deriving Principal Components on Excel.
Replies: 32
Views: 191854

Query on deriving Principal Components on Excel.

Can you upload the dataset so others can have a go?P
by Paul
October 3rd, 2001, 9:37 am
Forum: Student Forum
Topic: career change for an engineer ??
Replies: 23
Views: 190454

career change for an engineer ??

<t>Are you asking about yourself? If so, why aren't you sure about your age?!You are just about young enough, don't wait until you are 30-32 though. If you're a hot programmer then you should be ok. What sort of engineer are you? Don't forget that there are also jobs in financial software companies....
by Paul
October 3rd, 2001, 9:24 am
Forum: Technical Forum
Topic: Nomenclature for Volatility
Replies: 13
Views: 189900

Nomenclature for Volatility

<t>dS/S=mu*dt+sigma(t)*W(t), where sigma(t) can be an arbitrary stochastic process (possible correlated with W(t)) >>sigma can also be deterministic or uncertain.If I understood you right, you would call sigma(t) actual volatility. >>I would.Next, when I assume that I can observe sigma(t) I can calu...
by Paul
October 2nd, 2001, 12:42 pm
Forum: Programming and Software Forum
Topic: Convertible bonds, smiles, and more...
Replies: 20
Views: 192640

Convertible bonds, smiles, and more...

If the software is bigger than 1Mb just email it to paul@wilmott.com and I'll put it up. Of course, stochastic hazard rate is very fashionable. (But then so is orange and brown furniture from the 1970s.)P
by Paul
October 2nd, 2001, 8:21 am
Forum: Technical Forum
Topic: Nomenclature for Volatility
Replies: 13
Views: 189900

Nomenclature for Volatility

<t>To me1. Realized volatility is what happened in the past, depending upon the path taken by the asset. You'd have to specify the period over which you were measuring the realized vol, like three months. I think of this one as being a measure of volatility of a stock during the lifetime of some par...
by Paul
October 1st, 2001, 8:25 pm
Forum: Programming and Software Forum
Topic: Convertible bonds, smiles, and more...
Replies: 20
Views: 192640

Convertible bonds, smiles, and more...

I don't hate KMV, I just haven't been able to find out what they do! I enjoy modeling, so I like to know what's inside the box.P
by Paul
September 29th, 2001, 11:04 am
Forum: Off Topic
Topic: Bill Gates - The Oracle from Seattle
Replies: 43
Views: 192154

Bill Gates - The Oracle from Seattle

You are right, WILMOTT is the antichrist! And Mom told me that the milkman was my father!P
by Paul
September 28th, 2001, 3:55 pm
Forum: Off Topic
Topic: Bill Gates - The Oracle from Seattle
Replies: 43
Views: 192154

Bill Gates - The Oracle from Seattle

You never know whether you are coming or going with causality!P
by Paul
September 28th, 2001, 6:25 am
Forum: Off Topic
Topic: Bill Gates - The Oracle from Seattle
Replies: 43
Views: 192154

Bill Gates - The Oracle from Seattle

wASN'T THE WORLD SUPPOSED TO HAVE ENDED IN aUGUST 1999? wE SHLD COUNT OURSELVES LUCKY TO HAVE GOT THIS FAR. (Ooops, caps lock on, you can see I've checked out David's theory!)P
by Paul
September 24th, 2001, 4:33 am
Forum: Technical Forum
Topic: Option on PE ratio
Replies: 42
Views: 191510

Option on PE ratio

<t>reza, you've made one giant leap for mankind...backwards! It's easy enough to model E. (Whether it's normal, lognormal, Poisson or whatever is a detail.) But to get from that E to S uses a parameter k that is at best immeasurable, and at worst totally arbitrary!You can presumably hedge the S bit ...
by Paul
September 21st, 2001, 8:51 pm
Forum: Technical Forum
Topic: Option on PE ratio
Replies: 42
Views: 191510

Option on PE ratio

<t>reza, those are the choices as I see them, yes. And I'd go along with all that Aaron says. I also agree with Antoine that we must sort out the big picture approach before we worry about the details. The model that seems most fun to me wld be half way between my B and C (with reza's ideas in there...
by Paul
September 21st, 2001, 6:49 am
Forum: Technical Forum
Topic: Option on PE ratio
Replies: 42
Views: 191510

Option on PE ratio

<t>1. I like the E that changes discretely2. Negative E must be quite common, what is the convention for PE ratios then?3. Modeling E is easy. (Diffusion, jumps, whatever.) Enough data for that. (But don't forget that we can't easily hedge the E risk.) And the model should be more stable than a stoc...
GZIP: On