Time Series - MAT 3208
Course overview
The analysis of real-world data that have been observed at different time points leads to unique problems in statistical modeling and inference. The time series analysis explains specific ways of analyzing a sequence of data points collected over an interval of time. This course will introduce modern time series analysis techniques and theories which could use to model real-world applications.Introduction and objective of time series analysis. Some simple time series models: stationary models and the autocorrelation function, Estimation and elimination of trend and seasonal components, Testing the estimated noise sequence. Stationary process: Basics properties and forecasting methods, An introduction to ARMA models.
Textbook
Robert H. Shumway and David S. Stoffer. Time Series Analysis and Its Applications With R Examples Fourth Edition. Springer Publication
Learning objectives
By the end of the semester, students should be able to:
- Describe the important features of the time series pattern.
- Explain how the past affects the future or how two-time series can “interact”.
- Determine a model that describes the pattern of the time series.
- Forecast future values of the series.
Lecturer in charge
Ranjan Dissanayake - ranjand82@gmail.com
Instructor
Nethmi Herath - nethmichanika@gmail.com
Pre-requisites
Students should have a working knowledge of the R computing software. Students should also have an understanding of basic probability and statistical inference.
Teaching methodology
This course will introduce new material primarily through prepared slides and hands-on demonstrations. Students will be expected to work both individually and collaboratively (to the extent possible given the current conditions); course content and evaluation will emphasize the communication of ideas and the ability to think critically more so than a specific pathway or method. Other areas of this website provide an overview of the topics to be covered, including links to weekly reading assignments, lecture materials, computer labs, and homework assignments.
Evaluation
- Pre preparation Worksheets (10%)
- Homework (10%)
- quiz (10%)
- Mid Examination (15%)
- End Examination (25%)
- Final project (30%)
Grading Rubric
In general, allocated marks for each and every problem will graded under the following rubric.
- Preparation / Attempt / Start you will get - 25%
- Steps / Process / Logical Flow you will get - 50%
- Answer / Completion you will get - 25%
Important Dates
| Description | Date/ Time |
|---|---|
| Pre preparation Worksheets | Wednesday 11.59 AM |
| Homework | Sunday 11.59 PM |
| Quiz | Thursday 8.00 AM |
| Mid Examination | TBA |
| End Examination | TBA |
| Group / Individual Project | Every other Thursday at 9.30 AM |
Note: All the grades will be uploaded to the grade book (click here) within one week after your submission due date.
Computer Labs
Computer Labs
| Date | Topic | Lab Material | Homework | Due Date | Key |
|---|---|---|---|---|---|
| 15 Sep | Introduction and objective of Time Series Analysis | lab01 | lab01 | 18 Sep | Key |
| 06 Oct | Basic Time Series Objects | lab02 | lab02 | 09 Oct | Key |
| 20 Oct | Time Series Statistical Models | lab03 | lab03 | 23 Oct | Key |
| 27 Oct | Time Series Statistical Models | lab04 | lab04 | 30 Oct | Key |
| 3 Nov | AR and MA model estimation and forecasting | lab05 | lab05 | 6 Nov | Key |
Homework
Homework
The homework for each week is listed with the Computer Labs.
Homework format
Please submit your homework as an Rmarkdown document (.Rmd), which will allow you to combine text, equations, and R code into a pdf or html file. The easiest way to do so is to use the built-in capabilities of RStudio. For those unfamiliar with Rmarkdown, there is a nice introduction here on the R Markdown: The Definitive Guide. There is also help available in RStudio. Please use the template below when submitting your homework. Email your file(s) to the course mail timeseries.mat3208@gmail.com weekly.
Homework templates
Final Project
Final Project
As part of the class, each student will have to write a complete, publishable (<20 page) paper using the time series analysis techniques learned in class. See below for details on the structure of the paper.
Due Dates
Project proposal due Fri Jan 29th 11:59 pm PST Project methods due Thurs Feb 18th 11:59 pm PST Final paper due Fri Mar 12 11:59pm PST Presentations Mar 9th and 11th during class time Peer review due Fri Mar 19th 11:59pm PST
Data sets
Students are encouraged to use their own data and the paper may form a chapter for their thesis/dissertation. However some students might not have their own data. Students may also use data from the instructors, public datasets or datasets included in R libraries.
Project Plan
due Fri Jan 29th 11:59 pm PST
Write a 1-2 page description of your project idea that includes
the question(s) of interest the data you will use your general approach to analyzing the data. Things to consider:
- what type of time series models do you expect to use?
- univariate or multivariate
- one model or multiple models (and multi-model inference)
- covariates?
- linear, non-linear, or non-parametric
- Guassian or non-Gaussian
Project Methods
due Thurs Feb 18th 11:59 pm PST
-
Write a draft methods section for your project. This is HW #5. The section should include a mathematical description of your model sufficient for someone else to understand and fit that model.
-
Show that you can fit a pilot version of your proposed model. In other words, show that you can fit your proposed model to some data.
Preparation of final papers
references
References
Textbooks with specific R examples
The main class reference is the ATSA Lab Book
Holmes, E. E., M. D. Scheuerell, and E. J. Ward. Applied Time Series Analysis for Fisheries and Environmental data.
In addition, we will use these as references for the class.
Cowpertwait PSP, Metcalfe AV. 2009. Introductory Time Series with R. Springer, New York.
Hyndman RJ, Athanasopoulos G. 2018. Forecasting: Principles and Practice. eBook.
Some classic textbooks that you may find helpful
Box GEP, Jenkins GM, Reinsel GC. 2008. Time Series Analysis: Forecasting and Control. John Wiley & Sons, Hoboken, New Jersey.
Brockwell PJ, Davis RA. 2010. Introduction to Time Series and Forecasting. Springer, New York.
Durbin J, Koopman SJ. 2012. Time Series Analysis by State Space Methods. Oxford University Press, Oxford.