The following senior level STAT courses, to be offered in 2010/2011, have prerequisites which are possibly within the reach of advanced students in other disciplines (there is generally considerable flexibility here). They may be of interest to your senior undergraduate or graduate students. Please pass this notice along to your colleagues as well as to interested students.

GENERAL INTEREST SENIOR STATISTICS COURSES 2010/11

Syllabi for many of our graduate courses are available at http://www.stat.ualberta.ca/stats_centre/STATsyllabi.pdf. In addition, many of the courses have dedicated websites, detailed below.

FALL TERM

STAT 501/502/503/504/505: Directed Study I/II/III/IV/V

These courses allow graduate students, generally (but not always) in departments other than Mathematical and Statistical Sciences, to obtain graduate credit while attending STAT 361 (Sampling Techniques; Fall term), STAT 368 (Design and Analysis of Experiments; Winter term), 378 (Regression Analysis; Fall term), 479 (Time Series; Fall term) or 441 (Data Mining; Winter term). There may be additional requirements; see the calendar descriptions for more details.

STAT 335: Statistical Quality Control and Industrial Statistics - MWF 1200 - 1250

Instructor: Prof. N. Prasad (prasad@stat.ualberta.ca)

Control charts for variables and attributes. Process capability analysis. Acceptance sampling: single and multiple attribute and variable acceptance plans. Prerequisite: STAT 235 or 265.

STAT 361/504: Sampling Techniques - MWF 1100 - 1150

Instructor: Prof. N. Prasad (prasad@stat.ualberta.ca)

Basic sampling techniques. Stratification and clustering. Ratio and regression methods of estimation. Multiphase and multistage sampling schemes.

STAT 378/502: Applied Regression Analysis - MWF 1000 - 1050

Instructor: Prof. P. Li (pengfei@stat.ualberta.ca)

This course is intended as a thorough introduction to the use of linear and non-linear models in statistical analysis. We will be focusing on situations where we have one response variable and one or more explanatory variables. The emphasis will be on the use of the models in question for helping in the analysis of data. We will mostly present sample programs and solutions using R. Tentative topics to be covered are as follows:†

1. Simple linear regression analysis with one explanatory variable: inference on regression parameters, residual analysis, and prediction intervals.

2. Linear regression with more than one explanatory variables: extension of simple linear regression inference procedures; variable selection techniques such as backward elimination, forward selection, stepwise selection, AIC criterion, BIC criterion, cross-validation and generalized cross-validation, and penalized methods such as LASSO (if time permits).

3. Logistic regression and its application.

4. Multi-collinearity and its effects, and possible fixes such as ridge regression.

5. A brief introduction to non-linear regression for continuous data.

Prerequisites: STAT 265 and a course in Linear Algebra; MATH 225 recommended.

STAT 471: Probability I - MWF 1200 - 1250

Instructor: Prof. T. Choulli (tchoulli@math.ualberta.ca)

See the calendar description.

STAT 479/503: Time Series Analysis - TR 930 - 1050

Instructor: Prof. E. Gombay (egombay@ualberta.ca)

This is an introduction to the methods of time series, covering material both from the time domain and frequency domain perspectives. It has in the past been popular with Engineering, Chemistry and Applied Math/Math Finance students, as well as Statistics majors. As STAT 503 it can be taken for graduate credit. A prerequisite is "Consent of Instructor"; adequate preparation would be a STAT course which covered some regression, and a basic facility with mathematics.

STAT 512: Techniques of Mathematics for Statistics - MWF 1300 - 1350

Instructor: Prof. C. Frei (frei@cmap.polytechnique.fr)

We cover a range of mathematical topics useful in Statistics - matrix manipulations (orthogonality, Gram-Schmidt decomposition, diagonalization), calculus and analysis (limits, continuity, differentiation, sequences and series), multidimensional calculus and optimization (extrema, Lagrange multipliers), all with an eye to applications in Statistics and Probability (linear and nonlinear least squares estimation, Laws of Large Numbers, Central Limit Theorem, generating functions, maximum likelihood estimation, etc.). This course has in past years been popular with graduate students from departments of Engineering, Computing Science and others.

The similar course STAT 312 (not for graduate students) is a corequisite or prerequisite for many of our senior courses.

STAT 532: Survival Analysis - TR 1230 - 1350

Instructor: Prof. P. Li (pengfei@stat.ualberta.ca)

This course will focus on the specialized issues related to the analysis of survival or time-to-event data. The course begins by closely examining the features unique to survival data which distinguish these data from other more familiar types. Topics include an introduction to censoring and truncation, parametric survival models, non-parametric survival analysis methods, the proportional hazards model, multi-state models and competing risks and frailty models for clustered time-to-event data. All methods will be illustrated by in-class examples and homework sets. The level of the course is suitable for graduate students working in statistics, public health, medical school, science or engineering.

STAT 562: Discrete Data Analysis - MW 1000 - 1120

Instructor: Prof. KC Carriere (kc.carriere@ualberta.ca)

The course is an introduction to the statistical methods applicable when observed variables do not have numeric, but discrete character - instead of being represented by numbers, they arise from observing qualitative traits like gender, color of eyes, rank on a comparative scale, positive or negative reaction to the experimental stimulus, or alphabetic expressions arising in genetic analysis. These very few examples indicate that data with such variables are common in social, political, biological and medical sciences, as well as in psychology, insurance and economics. The presentation is aimed at understanding of the methods and their application to real data sets; the level of the course is suitable for graduate students working in statistics, mathematics, science, engineering, and, with certain minimal mathematical and computational background, also in humanities. Some knowledge of traditional statistical methods for numerical variables (in particular, linear models) is an asset - although not a formal prerequisite; the presentation is largely self-contained. More details available at www.stat.ualberta.ca/~mizera/562/ .

STAT 578: Regression Analysis - MWF 1200 - 1250

Instructor: Prof. Ivan Mizera (mizera@stat.ualberta.ca)

Fitting lines and curves accounts for the single most frequently applied statistical technology. The aim of STAT 578 is mastering the craft of linear and nonlinear regression modelling, including hypothesis testing and diagnostics, non-quantitative responses via generalized linear models (logistic regression, for instance), nonparametric curve fitting, and robust procedures. The presentation is oriented to practical skills in applying these methods, as well as on building necessary conceptual background for more sophisticated methods and other statistical disciplines. The level of the course is suitable for graduate students working in statistics, mathematics, science, or engineering, provided they have some background in matrix algebra and computing, corresponding to the advanced character of the course.

STAT 580: Stochastic Processes - TR 1400 ñ 1520

Instructor: Prof. B. Schmuland (schmu@stat.ualberta.ca)

This course is an introduction to stochastic processes without measure theory. We will concentrate on Markov chains in discrete time, but also do some martingales, Brownian motion, and stochastic integration. Prerequisites are probability and linear algebra. Students will need to do easy programming (including use of a random number generator) and should have access to some software that will do matrix computations (inverses and finding eigenvalues and eigenvectors.) For instance, Maple works nicely for both purposes. Further information is available at www.stat.ualberta.ca/~schmu/580.htm or by contacting the instructor.

=====================================================================

WINTER TERM

STAT 501/502/503/504/505: Directed Study I/II/III/IV/V

STAT 368/501: Introduction to the Design and Analysis of Experiments, MWF 1000 - 1050

Instructor: Prof. Ivan Mizera (mizera@mathstat.ualberta.ca)

This course introduces statistical techniques for the design and analysis of experiments. The main elements of design are replication, blocking, and randomization. The course begins with simple designs for comparing two treatments, and then proceeds to more complex designs and methods of analysis. Topics include: completely randomized design, randomized block design, analysis of variance, analysis of covariance, multiple comparisons, factorial designs, nested designs, and random effects.

Prerequisites: STAT 265 and a course in linear algebra. MATH 225 recommended.

STAT 432: Survival Analysis, TR 1230 - 1350

Instructor: Prof. P. Zhang (pengz@ualberta.ca)

We will study various statistical techniques used in the analysis of time-to-event dada as well as recurrent events data:

1. General introduction to survival analysis, basic concepts such as survival function, Hazard function, Censoring, Truncation.

2. Parametric approaches to survival analysis, Models for survival functions and hazard functions and inference.

3. Non-parametric approaches to survival analysis, Kaplan-Meier estimator and its variants, Comparison of survival functions, estimation of quantities such as median survival time.

4. Cox's proportional Hazard model.

It is assumed that students have exposure to statistics in general and to basic statistical distributions in particular. A good background in linear regression analysis (e.g. Stat 378) is very helpful.

STAT 441/505: Applied Statistical Methods for Data Mining, MWF 900-950

Instructor: Prof. Ivan Mizera (mizera@stat.ualberta.ca)

The course objective is to give the student understanding of and practical skills for a variety of advanced statistical methods suitable for practical data analysis. It is particularly suitable for those who, after learning essential, linear-model-based statistical techniques in regression and/or design of experiments, desire to widen their scope with more sophisticated methods not covered in other statistical courses. Topics include principles of statistical† model building and analysis applied in linear and generalized linear† models and illustrated through multivariate methods such as repeated† measures, principal components, and supervised and unsupervised† classification. The course is accompanied by lab sessions, which are devoted to analyzing real data sets with students. More details are available at http://www.stat.ualberta.ca/~mizera/441/.

STAT 568: Design and Analysis of Experiments, MWF 0900 - 0950

Instructor: Prof. P. Li (pengfei@stat.ualberta.ca)

Design and analysis of experiments can be seen as the two twin branches of statistics: the former being concerned with how to collect data appropriately, and the latter with how to analyze the data after they have been collected. In STAT 568, we will study the techniques for designing the appropriate experiments for different purposes and the corresponding data analysis methods. Tentative topics to be covered are as follows:

1. Basic principles of replication, randomization and blocking; introduction to linear model and experiment with single factor.

2. Experiments with more than one factor, blocking, Latin squares, analysis of variance and covariance.

3. Full factorial and fractional factorial experiments at two levels: the concept of confounding; maximum resolution and minimum aberration for choosing optimal fractional designs and blocking schemes.

4. Full factorial and fractional factorial experiments at three levels: two typical analysis methods, orthogonal components system and linear quadratic system.

5. Non-regular orthogonal design and data analysis.

6. Brief introduction to response surface methodology and robust parameter design.

Prerequisite: STAT 368 or STAT 378 or equivalent, OR my consent.

STAT 575: Multivariate Statistical Analysis, TR 1230 - 1350

Instructor: TBA

Most data sets include measurements of correlated variables. To be effective, a statistical must incorporate an appropriate model for the correlation structure. STAT 575 provides an introduction to the theory and application of multivariate statistical methods. The level of the course is suitable for graduate students working in statistics, mathematics, science, or engineering. In recent years, a substantial proportion of students have come from engineering.