Serving the Quantitative Finance Community

 
User avatar
RVP1234
Topic Author
Posts: 0
Joined: September 4th, 2008, 9:50 am

Importance of Bayesian Statistics

March 11th, 2009, 1:27 pm

Hi guys, I think its fair to say that bayesian statistics in finance is still pretty new. Do any of you see any growth within the next 5 years?
Last edited by RVP1234 on March 10th, 2009, 11:00 pm, edited 1 time in total.
 
User avatar
musterseed
Posts: 0
Joined: December 29th, 2006, 2:31 pm

Importance of Bayesian Statistics

March 11th, 2009, 4:11 pm

QuoteOriginally posted by: RVP1234Hi guys, I think its fair to say that bayesian statistics in finance is still pretty new. Do any of you see any growth within the next 5 years?Bayesian theory is used extensively in machine learning and pattern recognition and so has an use in algorithmic trading. Unlike derivatives pricing, algorithmic trading is actually growing. Therefore, it is not implausible that Bayesian theory will have more widespread use in finance in the future.
 
User avatar
StatGuy
Posts: 0
Joined: November 20th, 2007, 9:03 am

Importance of Bayesian Statistics

March 11th, 2009, 5:14 pm

The recent developments in machine learning involve the use of Nonparametric Bayesian statistics where the form of the prior is not parametric hence closer to reality. Algo trading will use these types of techniques more in future. SG
Last edited by StatGuy on March 10th, 2009, 11:00 pm, edited 1 time in total.
 
User avatar
rusti999
Posts: 0
Joined: October 4th, 2007, 8:19 pm

Importance of Bayesian Statistics

March 13th, 2009, 12:34 pm

QuoteOriginally posted by: StatGuyThe recent developments in machine learning involve the use of Nonparametric Bayesian statistics where the form of the prior is not parametric hence closer to reality. Algo trading will use these types of techniques more in future. SGNot anytime soon, IMHO, especially for stuff that needs decisions in real time. This kind of modeling is nice in principle, but doing inference is slow. Being nonparametric means that you have a much bigger space to explore for inference.
 
User avatar
StatGuy
Posts: 0
Joined: November 20th, 2007, 9:03 am

Importance of Bayesian Statistics

March 13th, 2009, 1:21 pm

QuoteThis kind of modeling is nice in principle, but doing inference is slow. Being nonparametric means that you have a much bigger space to explore for inference. True speed is a problem, but some recent developments in improving computational efficiency shows promise. With nonparametric modelling we might assume, for example, that the data are sampled from a completely unknown distribution, F, and the goal is to make inferences about functions, or even the pdf, of F. We could think of F as belonging to the class of all continuous distributions on the real line for example. So while the space is large in nonparametric modelling, once we combine data information through the posterior we reduced this. It's more complex than this in real life but I will spare you the details SG
 
User avatar
rusti999
Posts: 0
Joined: October 4th, 2007, 8:19 pm

Importance of Bayesian Statistics

March 14th, 2009, 1:15 pm

QuoteSo while the space is large in nonparametric modelling, once we combine data information through the posterior we reduced this. It's more complex than this in real life but I will spare you the details I see your point if you just use the precomputed posterior for new data points. I was under the impression that one might want to update your posterior with each new data point, but maybe it's not necessary to do this (at least not in real-time).