EC6001 Time Series

SMU MSE EC6001 Time Series Econometrics Term 2 AY21-22

Individual Project Instructions

Select a time series of interest to you (your “forecast series”), and any number of potential

predictors. The only restriction to your choice is that you have at least 40 observations in each

of your series. Justify your choice of predictors on theoretical grounds, either citing and briefly

describing a formal theory, or using informal arguments. Provide all relevant background and

descriptive analyses of your series that you think will be helpful to your reader. Your predictors

may be of a different frequency to the series you are trying to predict. The objective of the

project will be to forecast a particular series of your choice, and to evaluate the forecasts.

Instructions will be given to you one step at a time, as the course progresses. There is

nothing to hand in for now, but you are encouraged to start on your project immediately, and

to work on it as we proceed along in the course. Due date for the final report is Monday

April 11 via eLearn. Report format: no more than 20 pages, including all tables, figures, and

references, but not including computer code (data and computer code should be provided

separately). Use Times New Roman font, 11 or 12 pt, 1.5 line spacing. The limited number of

pages means that you will have to decide carefully what information to report, and what to

omit.

You will not be evaluated on how well you predict the series – your models might fail

abjectly – but on how go about choosing your model, how you deal with special features of

your data, how you evaluate your forecasts, and the clarity of your written report.

Part 1 Reserve roughly the last quarter of your sample for forecasting and evaluation (call this

your “forecast sample”). On the remaining “estimation” (or “training”) sample, find the most

appropriate ARIMA(p,d,q) model for your forecast series (you may include deterministic

trends and seasonal dummies in your model, where appropriate).

Added 10 Feb 2022

Part 2 As part of the model identification process in Part 1, you should formally test your

series to see if it is integrated, and of what order. Test your predictors as well for unit roots.

Which are integrated and to what order? Are the I(1) variables cointegrated?

Added 24 Feb 2022

Part 3 (Final Part) Choose two plausible forecasting models, and generate one-step ahead

forecasts of your selected variable over the forecast sample, updating your dataset and re-

estimating your model at each period. Present, evaluate, and compare the predictive accuracy

of your forecasts.

Added 17 Mar 2022

你可能感兴趣的:(算法)