IHP seminar – November

Posted in Seminar/Conference by Julyan Arbel on 1 December 2010


for those who missed this month IHP seminar organized by Ghislaine Gayraud and Karine Tribouley, below is a very quick summary/introduction of the talks. At closing time, it is not uncommon in that area to pass Field medals, as it happened when we left.

Anne Philippe spoke about long memory time series. For a weak second order stationary time series X_t, define its autocovariance \Gamma(h)=Cov(X_0,X_h). Then X_t has long memory if \sum_h\Gamma(h)=\infty. The long memory parameter is d such that \Gamma(h) decreases as h^{2d-1}. The core of the talk was about tests on long memory parameters for different time series (eg. test of equality), involving fractional Brownian motion…

Then, Judith Rousseau presented results on the Bernstein – von Mises property. It states that the posterior distribution of the parameter is asymptotically Gaussian. A nice consequence of BvM property is that Bayesian credible regions are also frequentist confidence regions, with confidence levels that coincide asymptotically.

Last, Erwan Le Pennec gave a technically involved talk, on a nice subject: the question raised by a group of researchers is to know whether or not Stradivarius violins are varnished differently from other violins. The statisticians’ contribution is to cluster the images of slices of violins, in order to help for their physical analysis. What is found is that the main difference in varnishes is the presence of a red pigment in Stradivarius violins. Which does not really  translates as a factor for exceptional sound and quality.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: