Syllabus of IE 553 - Applied Statistical Modeling and Data Analysis

Department: Industrial Engineering

Credits: Bilkent 3,    ECTS 5

Course Coordinator: Savaş Dayanık

Semester: 2019-2020 Fall

Contact Hours: 3 hours of lecture per week,    1 hour of Lab/Studio/Others per week
Textbook and Other Required Material:
  • Required - Textbook: Extending the Linear Model with R, Julian J. Faraway, 2016/Second Edition, CRC Press
Catalog Description:
Generalized linear models (e.g., binomial, Poisson, beta, gamma, multinomial regression for count, proportion, and multi-categorical data), quasi-likelihood methods for dealing with under- and over-dispersion; random effect models for grouped, hierarchical, and panel data; generalized additive models to discover the nonlinear transformations in all of the previous regression problems; applications to real data using a statistical software.
Prerequisite(s): None
Assessment Methods:
  Type Label Count Total Contribution
1 In-class participation Participation 1 5
2 Homework Homework 5 15
3 Midterm:Practical (skills) Midterm 1 40
4 Final:Practical(skills) Final 1 40
Minimum Requirements to Qualify for the Final Exam:
Course Learning Outcomes:
Course Learning Outcome Assessment
Fit a generalized linear, random effect, or generalized additive model, formulate and test various hypotheses about effects, and predict the natural parameters for given explanatory variable values Midterm
Diagnose violations of various model assumptions and take appropriate remedies Midterm
Weekly Syllabus:
  1. Maximum likelihood theory, score functions, Fisher information matrix, asymptotic normality of MLEs, delta rule
  2. Linear statistical models, diagnostics, inference, and prediction
  3. Logistic regression model for ungrouped two-category response data, diagnostics, inference, model selection
  4. Binomial regression model for grouped two-category response data, handling overdispersion with quasi-binomial regression model, beta regression model for the proportion data
  5. Poisson, dispersed Poisson, and negative binomial regression models for count data
  6. Contingency tables and correspondence analysis for cross-classified categorical data on several variables
  7. Multinomial regression models for ordered, unordered, nested, and hierarchical responses; linear discriminant analysis
  8. Generalized linear models (GLMs), hypothesis testing, diagnostics, robust estimation
  9. Gamma, inverse Gaussian, Tweedie GLMs, joining modeling of mean and dispersion, quasi-likelihood GLM
  10. Random effect models, estimation, inference, prediction, diagnostics; block, nested, crossed effects
  11. Repeated measures and longitudinal data, multiple response multilevel models
  12. Bayesian mixed effect models
  13. Generalized linear (binary, count) mixed effect models
  14. Generalized additive models, multivariate adaptive regression splines
Type of Course:   Lecture
Teaching Methods:   Lecture - Exercises - Assignment