I’m not gonna lie, I’m new here(like the post count of 1 wouldn’t give me away
). I stumbled on to this forum because of an assignment I have to make for school using Eviews. Now, I’m not the kind of guy that says, dear stranger, please make my homework for me but what I do want is ask for a little bit of help.The assignment is the following: We need to analyze a time series(uni or bivariate) using techniques we’ve seen in the course. This techniques are: Unit root, Q stat, correlograms, AR, MA, ARMA, ARIMA, SARMA, SARIMA models, the box jenkisn approach to forecasting, Information criteria(such as SIC and AIC to validate) and GARCH models for the univariate part. For bivariate we have seen cross correlograms and cointegration techniques(granger causality and Error correcting models).
Guidelines are: brief discussion of what the time series present, analyzing the series, adding a forecast, add only important figures/tables and give a “what have you learnt” opinion.
Every other year there were extensive PC courses using Eviews to make this assignment, but this year there are too many people enrolled in this course and they cancelled the PC courses(+ there are 3 people with no previous experience with eviews(I come from another school and had no econometric courses) including me), my work experience with Eviews is not that good, not to mention very bad and the course was given on a very theoretical way instead of a “learn to work with eviews” way.
So this weekend I sat behind my PC and try to make this assignment(it’s due Friday) but after only a few steps I hit a road block.
The series I tried to analyze was the unemployment rate(got my numbers from the saint Louis fed databank), my initial idea was to analyse these serie and then after a full analysis throw in the unemployment rate of another country and see if you could use this to forecast the other rate(because bivariate will award more points than univariate). It was easy to calculate that this rate was non stationary, the differences of this serie is stationary (unit root confirms both). Where do I go from here for a further analysis? I thought to take a look at the correlograms and than define an AR or MA process? And if I define an AR or MA process, do I have to look at the correlogram of the normal serie or the differenced serie? Or does any of you have a good "how to interpret ACF and PCF in a correlogram" tutorial?
Or do any of you guys have an idea of a time series that would be perfect for this kind of assignment? All suggestions are welcome ofcourse
Thanks