(motivation, problem description and indication of what will be covered in subsequent sections)
You will submit a short report/article on a forecasting problem. A suggestion template can be found by clicking above (Forecasting Exercise).
The submitted document should contain an abstract that summarises your problem and findings, which should not be longer than 150 words. The introductory section should focus on the problem, which you chose, and its motivation. In this section, your objectives should be clearly stated, including your forecasting horizon and the set of criteria that you will use to evaluate alternative solutions. The data and its preliminary analysis that will support your choice of forecasting methods and benchmarks (e.g. naïve forecasts), which should be described in the subsequent section. Your results section should include alternative forecasts, their evaluation, and the chosen alternative. These can be presented in tables, and you may use similar formats as in articles of your reading list in Moodle or in our textbooks. Finally, you will acknowledge any limitations in your study and briefly draw some conclusions.
As highlighted above, the problem and data are of your choice. The Library has databases from which you may extract data, links to different sources are also provided in this module, but you can also collect your own data. It is important, however, that you acknowledge any sources or references that you use in your coursework, including software. Referencing should be consistent and follow a standard format, the library site provides information on the different formats.
Your report and analysis should be your own unaided work. You are reminded that academic misconduct as defined in the University Disciplinary Code covers all types of assessment at Cass.
The submission should be a Word or a PDF file, maximum length of 5 pages (minimum font size 11, single-spaced). You may submit a separate file with an Appendix, or spreadsheet with calculations, but all relevant information should be summarized in the report or article.
If you have queries, email email@example.com.
This session provides an overview of business analytics and forecasting in business contexts. A range of forecasting problems are discussed. Basic statistical tools, which have been covered in previous modules are revised, so that you can use them in the next sessions.
Lecture slides l can be found below in two formats. Note that the handout format only contains the lecture, homework and appendix are included in the bigger file. Data from an example and for the homework are in a separate folder. There is also a very short article that illustrates the need to check data and assumptions.
Recommended reading, examples and exercises:
· Chapters 1 and 2 from Ord & Fildes (2013)
· Chapter 1 from Camm et al. (2016)
Lecture2: Time series forecasting 1
In this session, we focus on how we can compare different forecasts. We consider locally constant time series, and different ways to forecast their level. Long-term trends are also analysed and different ways of capturing them are discussed. In the lab, you will have the opportunities to present your findings from your homework and to develop a forecasting exercise using Excel (see last section of lecture slides).
Note that homework is a preparation for the next session, when statistical software will be introduced, and you should be able to compare your results to those obtained using a statistical software as well as other forecasting methods.
Recommended options for reading and examples as exercises:
· Ord & Fildes (2013) – sections: 2.7, 2.8; 3.1 to 3.3
· Albright & Winston (2016) – sections 12.2 to 12.6
· Camm et al. (2016) – sections 8.1 to 8.3
Lecture3: Time series forecasting 2
In this session we review simple exponential smoothing and move on to consider different approaches to trend forecasting. In the computer lab, you will have the opportunity to continue working on the Netflix sales data, but rather than using Excel, you will be using specialist statistical software. The instructions and data can be found below.
Ord & Fildes (2013) – Chapter 3
Albright & Winston (2016) – section 12.7
Ord & Fildes (2013) – 3.1, 3.3, 3.7, 3.8, 3.10, 3.13
Albright & Winston (2016) – examples 12.2, 12.4
Lecture4: Time series forecasting 3
This session addresses the different components of the time series. You will learn how to identify and forecast different types of seasonal time series and will further explore the Time Series Modeller in SPSS. Practical applications will be discussed and below you will find recommended reading and exercises.
Recommended Reading: Albright & Winston (2016) – section 12.8; Ord & Fildes (2016) – chapter 4
Recommended Exercises from Ord &Fildes (2013): 4.3, 4.4, 4.5, 4.6, 4.11
Lecture 5: Time series forecasting 4
This session covers multiple regression with time series models and also introduces ARIMA models, thus completing the set of “traditional methods” for forecasting.
Recommended Reading: Ord & Fildes – chapters 6 and 8