Stat 520, Fall 2015

Stat 520: Forecasting and Time Series, Fall 2015

Syllabus: Syllabus (pdf).
Instructor: Tim Hanson. E-mail: hansont@stat.sc.edu.
Office Hours: Tuesday/Thursday 11:30am to 12:30pm, and by appointment.
Office: 219C LeConte College, (803) 777-3859.
Class Meeting: Tuesday/Thursday 8:30am - 9:45am in Wardlaw College room 116 or online.
Textbook: Time Series Analysis with Applications in R, Second Edition by J. Cryer and K. Chan.
Prof. Joshua Tebbs' complete set of course notes.
Video link for STAT 520.
Kung-sik Chan's textbook page with all R code used in the book.

Lecture schedule, notes, and homework

  • Thur., Aug. 20: Logistics, syllabus, installing R and the TSA package. Notes Chapter 1 pp. 1-3. Chapter 1 R code for Tebbs' notes.
  • Tues., Aug. 25: Notes Chapter 1 pp. 4-19. Looking at time series examples, purpose of STAT 520.
  • Thur., Aug. 27: Notes Chapter 2 pp. 20-27. Joint distributions of random vectors, mean, covariance, correlation, etc.
  • Tues., Sep. 1: Notes Chapter 2 pp. 28-38. Important stochastic processes. Homework 1 due Thursday Sep. 10. R code for Chapter 2.
  • Thur., Sep. 3: Chapters 2 & 3 pp. 38-45. Stationarity, trends.
  • Tues., Sep. 8: Chapter 3 pp. 45-55. Constant & linear trends, residual process. R code for Chapter 3.
  • Thur., Sep. 10: Chapter 3 pp. 55-68. Polynomial and seasonal models. Homework 2 due Thursday Sep. 17.
  • Tues., Sep. 15: Chapter 3 pp. 68-79. Regression output, residual analysis (normality, independence), sample autocorrelation. Derivation of sample mean variance for stationary process.
  • Thur., Sep. 17: Chapter 4 pp. 80-88. Linear process, MA(1), MA(2), and MA(q). R code for Chapter 4.
  • Tues., Sep. 22: Chapter 4 pp. 88-96. AR(1), AR(2). Homework 3 due Thursday Oct. 1.
  • Thur., Sep. 24: Chapter 4 pp. 96-107. AR(2), AR(p), invertibility.
  • Tues., Sep. 29: Chapters 4 & 5 pp. 107-118. ARMA(p,q) models, differencing to achieve stationarity.
  • Thur., Oct. 1: Chapter 5 pp. 118-126. ARIMA(p,d,q) processes. R code for Chapter 5. Note: the polyroot function finds roots of polynomials in R.
  • Tues., Oct. 6: CLASS CANCELLED due to flooding.
  • Thur., Oct. 8: CLASS CANCELLED due to flooding.
  • Tues., Oct. 13: Chapters 5 & 6 pp. 129-138. Transformations, sample ACF. Here is Exam I, due by noon, Wednesday October 14. Solution.
  • Thur., Oct. 15: Chapter 6 pp. 138-153. Sample ACF for use with MA(q) and sample PACF for use with AR(p). R code for Chapter 6.
  • Tues., Oct. 20: Chapter 6 pp. 159-169. Sample EACF for choosing an ARMA(p,q) model, unit root test for differencing.
  • Thur., Oct. 22: Fall Break! No class.
  • Tues., Oct. 27: Chapters 6 & 7 pp. 168-180, 197-202. BIC best subsets, strategy for fitting, MOM and MLE estimation. R code for Chapter 7.
  • Thur., Oct. 29: Chapters 7 & 8 pp. 202-210.
  • Tues., Nov. 3: Examples. Homework 4 due Thursday Nov. 12.
  • Thur., Nov. 5: Chapter 8 pp. 208-215. Diagnostics. More examples. R code for Chapter 8.
  • Tues., Nov. 10: Chapter 8 pp. 215-217.
  • Thur., Nov. 12: Chapters 8 & 9 pp. 217-233. Guidelines for final project. R code for Chapter 9.
  • Tues., Nov. 17: Chapter 9 pp. 233-243. Free online textbook by R. Hyndman and G. Athanasopoulos on automated forecasting using ARIMA and exponential smoothing models.
  • Thur., Nov. 19: Chapter 9 pp. 244-266. Homework 5 due Tuesday Nov. 24; hand in Dec. 3 for full credit.
  • Tues., Nov. 24: Chapter 10 pp. 267-278.
  • Thur., Nov. 26: Thanksgiving! No class.
  • Tues., Dec. 1: Chapter 10 pp. 278-290. R code for Chapter 10.
  • Thur., Dec. 3: Seasonal differencing, including trends. Black-box forecasting: the forecast package for R. Last examples.
  • Mon., Dec. 7: Multiple choice Final Exam; email answers to Yawei by noon Tuesday, December 8.
  • Tues., Dec. 8: Email Final Exam answers by noon today.