EC313 Lecture 14
J James Reade
24/03/2015
Introduction
- Project basics.
- Presentation basics.
- Data manipulation: Excel, R, etc.
- Forecast by simulation.
- Stop early for questions.
Project Basics
- Construct models:
- ARIMA, ETS, time series decomposition.
- Check for structural breaks, check for outliers.
- Make use of
isat and Arima (plus explanatory variables).
- Construct basic forecasts on training datasets.
- Do 1 and 2 in presentation!
Presentation Basics
- Idea: Present what you have so far.
- What will you forecast? Why is it important?
- Prepare thoroughly, but don’t prepare “too much”:
- I.e. don’t try and cram talk too full.
- Maximum 5 minutes, so maximum 5 slides.
- Please be punctual tomorrow, we don’t have much time.
- Suggested slides:
- Introduction (including idea).
- Literature/context.
- Data (collection, characteristics, plot?).
- Results (can just be proposed model, some interim results).
- Conclusions.
- Starter for 10: http://rpubs.com/jjreade/ec313-slides-template
- Data plus lecture slide material: https://www.dropbox.com/sh/l8kxmgsvr7yo1gl/AACV7CwdydfZUz-Z44Bep9qOa?dl=0
- Useful stuff from the question mark:
Markscheme for Project
- I will create this during the coming week.
- Two main ingredients to project:
- Forecasting the variable chosen.
- Writing up the modelling and forecasting of that variable.
- Credit will be awarded for depth of project reflecting volume of work:
- For projects with more “economic” variables this may represent:
- A wider range of forecasts, including forecasting explanatory variables.
- A deeper review of other forecasters in your chosen area.
Writing the Project
- Recommend writing it in an .Rmd file:
- Can nest code much more easily within .Rmd file.
- Knitting to
.doc file I’ve found to be a bit unreliable.
- Better to knit to
.html, and for sharing, to .pdf.
- Writing style: Code in the text, or elsewhere?
Feedback!