Statistical Computation

(18 hours of lecture 8 hours of exercises)

The course is an introduction how to use the software packages R and SAS for descriptive and exploratory data analysis. This will include graphical presentation of data, basic parametric and nonparametric tests as well as ANOVA models, linear models and logistic regression.

Spring term 2004 at the Department of Mathematics and Statistics, University of Tampere.

Statistical Methods in HCI

(20 hours of lectures and 10 hours of exercises)

The course will deepen the student's knowledge in statistical hypothesis testing and introduce nonparametric and exact testing procedures. Further topics will be the model selection in linear regression models together with model diagnostics, an introduction on how to fit linear mixed models for repeated measurements and an introduction to factor analysis. The course will focus rather on understanding and practical points than on technical details. The software used in the course will be R.

Spring term 2005 at the Department of Computer Science, University of Tampere.

R – An Introduction

(20 hours of lectures and 6 hours of exercises)

A course to familiarize students with the statistical software package R. Students learn how to use R for basic data analysis and for producing high quality graphs.

Autumn term 2007 and Autumn term 2009 at the Tampere School of Public Health, University of Tampere and Autumn term 2011 at the School of Health Sciences, University of Tampere.

Basics of Statistics

(18 hours of lectures and 16 hours of exercises)

The course gives an overview of statistics, its importance and use in different fields of science. Basic statistical concepts and descriptive statistics are introduced, as well as an elementary introduction to estimation and hypothesis testing is given. An introduction to a statistical software package is also included.

Spring term 2008, Spring term 2009, Spring term 2010 and Spring term 2011 at the Department of Mathematics and Statistics, University of Tampere.

Classics and Recent Trends in Statistics

(Joint seminar with Hannu Oja and Jaakko Nevalainen)

The goal of this seminar is to read jointly either classical papers of statistics or recent new interesting publications.

In the seminar each participant will in turn pick a statistical paper that he likes and thinks is important and relevant. These can be for example some classical papers, review papers, tutorials or recent publications.
The chosen paper will be present then by the participant after having it distributed first to the all other participants so that they can read it before the meeting.

Spring term 2009 at the Tampere School of Public Health, University of Tampere.

Classics and Recent Trends in Statistics II

(Joint seminar with Hannu Oja)

The goal of this seminar is to read jointly either classical papers of statistics or recent new interesting publications.

In the seminar each participant will in turn pick a statistical paper that he likes and thinks is important and relevant. These can be for example some classical papers, review papers, tutorials or recent publications.
The chosen paper will be present then by the participant after having it distributed first to the all other participants so that they can read it before the meeting.

Seminar lasting from spring term 2010 to spring term 2011 at the Tampere School of Public Health, University of Tampere.

Classics and Recent Trends in Statistics III

(Joint seminar with Hannu Oja)

The goal of this seminar is to read jointly either classical papers of statistics or recent new interesting publications.

In the seminar each participant will in turn pick a statistical paper that he likes and thinks is important and relevant. These can be for example some classical papers, review papers, tutorials or recent publications.
The chosen paper will be present then by the participant after having it distributed first to the all other participants so that they can read it before the meeting.

Ongoing seminar since autumn term 2011 at the School of Health Sciences, University of Tampere.

Introduction to Dimension Reduction

(16 hours of lectures and 5 hours of exercises)

Modern data sets are increasingly large and therefore more and more problematic to handle. This dimensionality makes the data analysis more complex and also a graphical inspection of the data almost impossible. Statistical dimension reduction attempts to reduce the dimension with as little loss of information as possible. Dimension reduction methods areeither supervised or unsupervised. In unsupervised dimension reduction the whole data matrix is reduced - for that the researcher has to decide on what is really meant by loss of information. Several unsupervised methods like PCA, FA, ICS and ICA will be discussed in this course. Supervised dimension reduction is usually applied in a regression or general modeling context where the number of explaining variables should be reduced without losing information about the dependencies between the explaining variables and the response. For this purpose SIR, SAVE and similar approaches will be discussed in the course. The course will also have a practical part where R is used to learn how to do dimension reduction in practise.

Autumn term 2011 at the School of Health Sciences, University of Tampere.

I gave practicals in the following courses:

Generalized Linear Models (Hannu Oja)

Spring term 2006 at the Tampere School of Public Health, University of Tampere.

Longitudinal Data Analysis (Jianxin Pan)

Spring term 2006 at the Tampere School of Public Health, University of Tampere.

Biometry and Statistical Computing (Hannu Oja)

Autumn term 2007 at the Tampere School of Public Health, University of Tampere.