ECON 5337
Review For Exam #1
March 7, 2016
Exam coverage: All lectures, Chapters 1-4, Chapters 6-7.
As a guide in your exam preparation, I have posted a recent midterm. As noted in class,
we have had fewer class periods than we had last semester, and as s
Points covered earlier
1) Sample mean
2)Sample std deviation
Newwwww
1)Predicted value minus the actual value gives us the value of the residual.
2)Fitted Y
3)Log Likelihood -> how likely is the data that i have observed given the model i have used.
It d
US Exports
Forecasting
Team Members
Nitin Shewale
Shaval Popli
Hardeep Chawla
Gajanan Ganji
Kunal Parashar
US Exports Background
According to data published in 2013
Worlds largest trading nation with exports of 2.3 trillion
USD.
Ratio of 5400-5600 jobs
1) When we have a model with no of variables, then we want to pin down the important variables.
Or find variables in the model which are insignificant, assumption is we prefer smaller model.
So for that we run a test associated with the hypothesis that a
Different SARMA(p,q)x(P,Q) models
The general model could be written as:
(1-1L-2L2-2L2)(1-sLs- 2sL2s - PsL2Ps)(Yt-)=
(1+1L+2L2+qLq) (1+sL+2sL2s+QsLQs)t.
Models
k
k,k
ARMA(models)
P=0, Q=0
_
AR(1)
Decays exponentially to 0, only
k,k= if k=1, =0 if k>1.
app
1) When we have a model with no of variables, then we want to pin down the important variables.
Or find variables in the model which are insignificant, assumption is we prefer smaller model.
So for that we run a test associated with the hypothesis that a
3) a) Regression model run on the variable y and on one lag of itself and a constant in Eview.
Results of the same are as shown below in the screenshot
3) b) Correlogram diagram of the residuals.
Lags used 20, please find the screenshot below for the res