SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
victor123
Topic Author
Posts: 71
Joined: August 26th, 2006, 8:42 pm

Quant Research - Current trends

November 15th, 2012, 9:22 pm

Machine Learning - This keeps coming up in trading research jobs. Is this still not a very academic field ? Or, are the banks, HF's investing time and money to research something that could be profitable? Thought this would be both expensive and long drawn ? Or, do they actually make very limited use of this in real life ? Also, is this mostly used in very short-term trading strategies ?
Last edited by victor123 on November 14th, 2012, 11:00 pm, edited 1 time in total.
 
User avatar
farmer
Posts: 13477
Joined: December 16th, 2002, 7:09 am

Quant Research - Current trends

November 15th, 2012, 11:12 pm

QuoteOriginally posted by: DevonFangsA huge topic would be to find an acceptable way of quantifying model risk.Something that could spit out an estimate of model risk, and a standard deviation or error in the estimate, would be ideal.
 
User avatar
exneratunrisk
Posts: 3559
Joined: April 20th, 2004, 12:25 pm

Quant Research - Current trends

November 16th, 2012, 8:08 am

I echo frenchX and add - knowingly or not, regulators force kind of trading "standards", where pricing needs to include analysis which is done usually "later" with the scope of (counter party) portfolios, scenarios, ...... C/DVA, VaR, ... adapted pricing? This will cause enormous performance requirements, robustness might become questionable and model validation close to impossible? However, if this became true we might be forced to stay in the risk-free world (at least for DVA no other approaches are published)?(In general, if you make a system complex AND tightly coupled you might have unintended consequences, but not enough time to react to them)One thing has been mentioned. Machine learning might become more important to do better "calibration", decision support, .. IMO, machine learning methods for quant finance need to exratct knowledge that is computational AND white box (not so easy).Having the combinatoric complexity in mind logic constraint programming might enjoy a renaissance?.... but this is close to crystal ball consultation ... and all IMO
Last edited by exneratunrisk on November 18th, 2012, 11:00 pm, edited 1 time in total.
 
User avatar
DevonFangs
Posts: 3004
Joined: November 9th, 2009, 1:49 pm

Quant Research - Current trends

November 16th, 2012, 9:48 am

QuoteOriginally posted by: farmerQuoteOriginally posted by: DevonFangsA huge topic would be to find an acceptable way of quantifying model risk.Something that could spit out an estimate of model risk, and a standard deviation or error in the estimate, would be ideal.This is fun because you've no idea. It is actually being discussed, even though it's really that ill-defined.
ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...


GZIP: On