3rd Seminar by CAST

Alkaa kello
Päättyy kello

Linna rakennus, luentosali K108, osoite: Kalevantie 5.


CAST - Centre for Applied Statistics and Data Analytics

Celebrating European Statistics Day Seminar

The third Seminar by the Centre for Applied Statistics and Data Analytics (CAST) will gather researchers, other faculty and students interested in applied statistics and data analytics working at the University of Tampere and other Universities, Research Institutes and Companies.

The main aims of the seminar events by CAST are: (i) to bring awareness of the importance of statistics and data analysis in research; (ii) to create a forum of discussion where researchers present their work and research questions followed by discussion and feedback from the audience; (iii) strengthen the links between schools and research groups that might lead to future collaborations in terms of research articles and funding applications.

The European Statistics Day: At the initiative of the European Statistical Advisory Committee (ESAC) and with the support of the members of the European Statistical System and the European System of Central Banks, the European Statistics Day will be celebrated by the European statistical community on 20 October.

The visiting speakers of the seminar are Professor Hannu Oja (University of Turku),  Emeritus Professor Antti Penttinen  (University of Jyväskylä) and Emeritus University Lecturer Simo Puntanen (University of Tampere).

Please register for the event by Wednesday 19 October, at 8:00 a.m.


9.30—9.35 Opening

·         9.35 – 10.20 Hannu Oja (University of Turku, Finland)

·         Third and fourth cumulants in search of independent components

·         In independent component analysis it is assumed that the observed random variables are linear combinations of latent, mutually independent random variables called the independent components. In this talk projection pursuit is used to extract the non-Gaussian components and to separate the corresponding signal (non-gaussian) and noise (gaussian) subspaces.  Our choice for the projection index is a convex combination of squared third and fourth cumulants and we estimate the non-Gaussian components either (i) one-by-one (deflation-based approach) or (ii) simultaneously (symmetric approach). The properties of the unmixing matrix estimates are considered in detail through the corresponding optimization problems, estimating equations, algorithms and asymptotic properties. The talk based on the joint work with Joni Virta and Klaus Nordhausen.

·         10.20 – 11.05 Simo Puntanen (University of Tampere, Finland)

·         Ingram Olkin (1914-2016): A man who knew inequalities, multivariate analysis and more

·         Ingram Olkin, Stanford Emeritus Professor, Master of multivariate statistical analysis, majorization, linear algebra, and meta-analysis, passed away on 28th of April 2016 at home in Palo Alto, California. In this talk we say a few words about his career and go through some personal memories and joint experiences that we were having with him.

11.05-11.30 Coffee Break

11.30—12.15 Antti Penttinen (University of Jyväskylä, Finland)

·         Statistics for eye movement data

·         A moving target in a heterogeneous environment is considered where the target is allowed to have memory. The motivating example arises from eye movement experiments where a participant is inspecting a picture and her eye movements are recorded using an eye movement camera (during a few seconds). The long memory effect, which we call self-interaction, is associated with learning during the inspection time. A (future) application is to model animal movements through GPS tracking. Both type of sequences are complex in the sense that they may be inhomogeneous both in space and in time.

The general model for eye movement sequences is spatio-temporal point process. We consider sequential point processes which are informative marginals of the general spatio-temporal point process.  Models with explicit constructions for self-interaction in heterogeneous environments are presented, the Monte Carlo maximum likelihood method for parameter estimation as well as summary based model criticism are suggested and applied to eye movement experimental data. Generalization to spatio-temporal point processes and open questions are discussed.


Scientific Organisation: Paulo Canas Rodrigues: Paulo.Rodrigues@uta.fi
Local Organisation: Ansa Lilja: Ansa.Lilja@uta.fi