ISOM 4540 Time Series Analysis and Forecasting Instructor Inchi Hu, Professor of ISOM, Room 4436, Tel: 2358-7734 Office hours: By appointment Email: imichu@ust.hk; Tutorials The weekly tutorial will go over assignments and supplementary materials. The tutorials will begin in the second week. Tutors: Ms. Jenny Choi Email: kailam@ust.hk Ms. Mengmeng AO Email: maoaa@ust.hk Homepage: http://lmes2.ust.hk All lecture notes, assignments, solutions, old exam papers, and important announcements will be posted on the homepage Course Objective The objective of this course is to equip students with various forecasting techniques and knowledge on modern statistical methods for analyzing time series data. The course consists of three parts: I. Univariate methods; II. Regression methods; III. ARIMA models. Intended Learning Outcomes Upon completion of the course, you should be able to Understand the fundamental advantage and necessity of forecasting in various situations. Know how to choose an appropriate forecasting method in a particular environment. Know how to apply various forecasting methods, which includes obtaining the relevant data and carrying out the necessary computation (running suitable statistical software, if necessary). Improve forecast with better statistical models based on statistical analysis 1 Teaching Method The method is lecturing aided by directed discussion. The context of the relevant concepts and methods will be presented first followed by the discussion of pre-designed questions and examples to explore the concepts and methods in depth. Course Materials: “Forecasting and Time Series”, 4th Edition, by Bowerman and O’Connell, Duxbury Additional materials (lecture notes and business cases) will be available from course website. Assessment scheme Your grade is based on the following components: The 1st exam (40%) covers Parts I & II; an in-class, closed-book, written exam of approximately 90 minutes long. The 2nd exam (40%) covers Part III; similar format to 1st exam. Assignments (15%) will be given every one to two weeks. There are two types of assignments: individual-based and group projects, where students work in groups. All assignments will be collected and returned by your TA during tutorial sessions. Participation (5%) is crucial to a lively and effective learning environment. Most students can get 3% by attending the class regularly without disturbing behaviors. For those students who actively participate classroom discussions and raising and answering questions that help the whole class to learn better, a full 5% will be awarded. In exceptional situations, when a student shows continuous and significant improvements, the instructor reserves the right to put more weight on the 2nd exam than the 1st exam. 2 Schedule Part I: Univariate methods Feb. 2, Setting the stage Feb. 7, Simple smoothing methods Feb. 9, Decomposition methods [HW1] Feb. 14, Holt’s and Winters’ methods Part II: Regression method Feb. 16, Review simple linear regression [HW2; Group] Feb. 21, correlation analysis & multiple regression Feb. 23, Marginal sums of squares Feb. 28, Use of variables in regression Mar 1, meaning of regression coefficients [HW3] Mar 6, Variable selection techniques Mar 8, Model selection in regression [HW4; Group] Mar 13, Residual analysis [Read Appendix B for matrix algebra of regression analysis] Mar 15, Detecting outliers in regression Mar 20, Autocorrelated errors in regression ☆ Thur. Mar. 22, Review for the 1st Examination ☆ Tue. Mar. 27, the 1st Examination Part III: Box-Jenkins method Mar 29 Introduction to Box-Jenkins method Apr. 10, 2nd order stationary processes Apr. 12, Identification of ARMA models. [HW5] Apr. 17, Parameter estimation in ARMA models Apr. 19, Diagnostic checking in ARMA models Apr 24, Forecast using ARMA models Apr 26, ARIMA model and forecasting [HW6; Group] May 3, Seasonal ARIMA models May 8, Seasonal ARIMA model building [HW7; Group] ☆ Thu. May 10, the 2nd examination 3 Academic Integrity Without academic integrity, there is no serious learning. Thus you are expected to hold the highest standard of academic integrity in the course. You are encouraged to study and do the project in groups. However, no cheating and plagiarism will be tolerated. Anyone caught cheating or plagiarism will fail the course. Please make sure you adhere to the HKUST Academic Honor Code at all time (see http://www.ust.hk/vpaao/integrity/). Learning Environment Maintaining an exciting learning environment is the responsibility of both the teacher and the students. I expect students to behave responsibly during class without disturbing your classmates. You are encouraged to raise and answer questions in classroom. Particularly welcomed are active participation in the classroom discussion and sharing your thoughts that enhance the learning. 4