﻿ Curriculum Appllied Statistics f

Curriculum Appllied Statistics for Ph.D. Students

Spring Semester 2017

• Who should attend: PhD students in WI who want to brush up their statistical knowledge.

• Prerequisites: M.S. in Computer Science or comparable field with equivalent amount of mathematical and statistical education.

• Required equipment: Personal Laptop; install R software

• Work approach: For each of the six half days, the statistical concepts are introduced with numerical examples, students will (re-)compute numerical problems and references to further sources will be given (books, Internet). The students work actively during the class hours on selected problems. The focus is not on theoretical discussion but on conceptual discussion and practical applications.

• Goal: Refreshing of statistical knowledge, getting ideas for own PhD work.

• Handout: For each of the six half days, a set of Power Point slides and data will be provided (see below).

• Suggested literature: see various links in syllabus.

• Examination: At the end of each half day, a short assignment must be done in class. To pass the class all six assignments must be done in class.

• Installation of R software: all students should install the R software on their laptop (not need to install Tinn-R).

Content

1. Day Morning: Introduction to R and Sources of Statistics  (3. March 2017, 08:00 to 12:00)

• Overview, installation of R, references to books and Internet links

• Statistical concepts, population, parameters, estimators (point/interval)

• Data manipulation, missing data and imputation, grouping, empirical distributions, graphs and visuals

1. Day Afternoon: Basics and Repetition (3. March 2017, 13:30 to 17:00)

• Basics of probability and stochastic modelling

• Important distributions (Bernoulli, binomial, Poisson, Pascal, geometric, exponential, normal, Weibull etc.)

• Sampling from distributions (inverse sampling technique), random number generation

2. Day Morning: Design of Experiments/Surveys and Testing I (10. March 2017, 08:00 to 12:00)

• Basics of inferential statistics, test of hypothesis, level of significance, types of errors, bootstrapping statistics, permutation tests, etc.

• Meaning and application of most important statistical tests (Z, t, F, Chi2, ANOVA, Friedman, non-parametric, etc.), post-hoc analysis

• Classification of statistical tests (100 Statistical Tests, Gopal K. Kanji) with examples and exercises

2. Day Afternoon: Design of Experiments/Surveys and Testing II (10. March 2017, 13:30 - 17:00)

• Design of experiments (sample size computation, orthogonal designs, (fractional) factorial designs, Latin squares, etc.)

• Conjoint analysis (choice based)

• Design of surveys (representativeness, significance, biases, sampling, (post-) stratification, design of Internet surveys

3. Day Morning: Statistical Modelling I (17. March 2017, 08:00 - 12:00)

• Simple and multivariate linear regression (modeling, practical recommendations), importance of factors, interaction, multicollinearity, step-wise regression, influence of observations, residual analysis, model validation, predicting values

• Logistic, ordinal and multinomial joint-point, orthogonal regression, GLM, Poisson regression, survival analysis, logit/probit, mixed-effect, tree-based models, etc.

3. Day Afternoon: Statistical Modelling II (17. March 2017, 13:30 to 17:00)

• Discriminant analysis, principal components and factor analysis, dimension reduction, structural equation modeling

• Classification and clustering

• Time Series and forecasting (exponential smoothing, moving averages, seasonal effects, regression), autoregressive models

_________________________________

Last update: Oct. 17, 2016