## MR vs ARIMA (Multiple Regression vs Auto Regressive Integrated

### MR vs ARIMA (Multiple Regression vs Auto Regressive Integrated

Dear Colleagues,

Can anyone help in pointing out to me  sources of information to carry out a
comparison between the MR (Multiple Regression) and the ARIMA (Auto
Regressive Integrated Moving Average) techniques and to draw a comparison

Best Regards

Mahmoud

### MR vs ARIMA (Multiple Regression vs Auto Regressive Integrated

Quote:

> Dear Colleagues,

> Can anyone help in pointing out to me  sources of information to carry out a
> comparison between the MR (Multiple Regression) and the ARIMA (Auto
> Regressive Integrated Moving Average) techniques and to draw a comparison

Box, George E. P., et al., Time Series Analysis: Forecasting & Control
(3rd Edition)

Hi All

I got a very strange result.
I use a set of data(32 data) and try to estimate the parameter of
ARIMA(5,0,0) using Minitab and Matlab. (Minitab is just for
benchmarking).

The Matlab Result :
a: [1 -0.6461 -0.1922 -0.1807 -0.1281 0.1769]
b: []
c: 1
d: 1
f: []
da: [0 0.1650 0.1961 0.1930 0.2049 0.1821]

The Minitab Result of Parameters are:
p0=323.92; p1=0.4735; p2=0.1418; p3=0.1416; p4=-0.0207; p5=-0.4921

1) There a big different between the result of parameter.
2) and of course the result of forecasting aslo difference.

Below is how i use the Matlab:
y=[259;262;218;372;669;435;396;536;600;654;586;570.9166667;445.0833333
;501.5;257;500.5833333;257;342.5;437;350.4166667;289.8333333;339.08333
33;393.75;378;528.1666667;628.4166667;531;545.8333333;671.1666667;436;
310.25;286]
y = iddata(y);
fb = ar(y,5,'burg');

Is this because i use the wrong step in Matlab ?
OR
Is Matlab's AR(5) = Minitab's ARIMA(5,0,0) ?

Any idea ?