Adaptive Sequential Bayesian Change-point Detection

author: Ryan Turner, University of Cambridge
published: Jan. 19, 2010,   recorded: December 2009,   views: 6512
Categories

Slides

Related content

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
Lecture popularity: You need to login to cast your vote.
  Delicious Bibliography

Description

Nonstationarity, or changes in the generative parameters, are often a key aspect of real world time series, which comprise of many distinct parameter regimes. An inability to react to regime changes can have a detrimental impact on predictive performance. Change point detection (CPD) attempts to reduce this impact by recognizing regime change events and adapting the predictive model appropriately. As a result, it can be a useful tool in a diverse set of application domains including robotics, process control, and finance. CPD is especially relevant to finance where risk resulting from parameter changes is often neglected in models. For example, Gaussian copula models used in pricing collateralized debt obligations (CDOs) had two key flaws: assuming that subprime mortgage defaults have a fixed correlation structure, and using a point estimate of these correlation parameters learned from historical data prior to the burst of the real-estate bubble [1, 2]. Bayesian change point analysis avoids both of these problems by assuming a change point model of the parameters and integrating out the uncertainty in the parameters rather than using a point estimate.

See Also:

Download slides icon Download slides: nipsworkshops09_turner_asbcpd_01.pdf (561.3 KB)


Help icon Streaming Video Help

Link this page

Would you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !

Reviews and comments:

Comment1 Tariq Abuhashim, April 9, 2010 at 3:33 a.m.:

There are no slides attached to the video, it's really hard to read the presentation from the video, please attack the slides.

Thanks


Comment2 Ryan Turner, April 13, 2010 at 1:24 p.m.:

Here are the slides:
http://mlg.eng.cam.ac.uk/rdturner/BOC...

I will try to get these added

Write your own review or comment:

make sure you have javascript enabled or clear this field: