Slides-04Time Series Analysis using ARMA models: Part 2
Univariate Time Series Analysis: ARIMA models
Copyright By Assignmentchef assignmentchef
Building ARIMA models
Autoregressive Process
Consider the following AR(2) process:
yt = 2 0.5yt1 + 0.3yt2 + t , t N (0, 1) , T = 100
The characteristic equation is given by
1 + 0.5z 0.3z2 = 0
The characteristic roots are
0.52 + 4 0.3
0.52 + 4 0.3
This is a stationary series as the characteristic roots are larger than
1 in absolute value.
Univariate Time Series Analysis: ARIMA models
Building ARIMA models
Autoregressive Process
Note that stationarity could also be concluded from
i=1 i = 0.5 + 0.3 = 0.2 < 1i=1 |i | = 0.5 + 0.3 = 0.8 < 1PropertiesI The expected value of the series is given byE (yt) = 2 /(1 + 0.5 0.3) = 1.67I The variance is given by(1 + 0.3) (1 + 0.5 0.3) (1 0.5 0.3)Univariate Time Series Analysis: ARIMA modelsBuilding ARIMA modelsAutoregressive ProcessI The ACF is given by1 = 0.5 /(1 0.3) = 0.71432 /(1 0.3) + 0.3 = 0.65713 = 0.5 0.6571 + 0.3 0.7143 = 0.54294 = 0.5 0.5429 + 0.3 0.6571 = 0.4686I The PACF is given by11 = 0.7143kk = 0 k > 2
Univariate Time Series Analysis: ARIMA models
Building ARIMA models
Autoregressive Process
Figure 37 : Theoretical ACF and PACF of generated AR(2) process
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Univariate Time Series Analysis: ARIMA models
Building ARIMA models
Autoregressive Process
Figure 38 : Dynamic impact of a shock t on y
t-5 t t+5 t+10 t+15 t+20 t+25
CS: assignmentchef QQ: 1823890830 Email: [email protected]
Reviews
There are no reviews yet.