New at pd
10-11-2001, 05:13 PM
Good afternoon everyone,
There is a problem in the ASM for which I don't quite believe the answer. The question is as follows:
You are given the time series:
y_t = 100,105,110,90,100,95,100, which is modeled as an ARMA process
y_t -mu = -0.5*(y_(t-1) -mu)+.1*e_(t-1)+ e_t
Calculate the sum of squares error functions.
So, let w_t = y_t - mu, so the time series is represented as 0,5,10,-10,0,-5,0
By algebra,
e_t = w_t + .5*w_(t-1) -.1*e_(t-1)
starting with the assumption that e_1 = 0, and proceeding, I get
e_2=5
e_3=12
e_4=-6.2
e_5=-4.38
e_6=-4.562
e_7=-2.0438
So the sum of squares function is 251.61.
The solution says that e_7 = 2.9562, thus making the sum of squares function 256.18.
Confirmation that either one is correct would be greatly appreciated. Thanks.
There is a problem in the ASM for which I don't quite believe the answer. The question is as follows:
You are given the time series:
y_t = 100,105,110,90,100,95,100, which is modeled as an ARMA process
y_t -mu = -0.5*(y_(t-1) -mu)+.1*e_(t-1)+ e_t
Calculate the sum of squares error functions.
So, let w_t = y_t - mu, so the time series is represented as 0,5,10,-10,0,-5,0
By algebra,
e_t = w_t + .5*w_(t-1) -.1*e_(t-1)
starting with the assumption that e_1 = 0, and proceeding, I get
e_2=5
e_3=12
e_4=-6.2
e_5=-4.38
e_6=-4.562
e_7=-2.0438
So the sum of squares function is 251.61.
The solution says that e_7 = 2.9562, thus making the sum of squares function 256.18.
Confirmation that either one is correct would be greatly appreciated. Thanks.