Parameter range from complex inequality

Click For Summary
SUMMARY

The discussion focuses on determining the parameter range for phi1 and phi2 in the autoregressive process defined by the equation X(t) + phi1*X(t-1) + phi2*X(t-2) = Epsilon(t). The characteristic polynomial F(z) = z^2 + phi1*z + phi2 indicates that the process is stationary if the roots z are within the unit circle. The user initially struggled with complex parameters but concluded that only the roots can be complex, simplifying the process of finding the allowed parameter range for phi1 and phi2.

PREREQUISITES
  • Understanding of autoregressive processes in time series analysis
  • Familiarity with characteristic polynomials and their roots
  • Knowledge of complex numbers and their properties
  • Experience with stability conditions in statistical models
NEXT STEPS
  • Study the implications of the unit circle on the stability of autoregressive models
  • Learn about the Routh-Hurwitz criterion for stability in control systems
  • Explore the use of complex analysis in determining the roots of polynomials
  • Investigate numerical methods for solving polynomial equations
USEFUL FOR

Statisticians, data scientists, and researchers working with time series analysis or autoregressive models who need to ensure model stability and parameter selection.

DanMat
Messages
2
Reaction score
0

Homework Statement


Hi Guys,
I try to find the range for parameters phi1 and phi2 were the autoregressive process below is stationary.
We have the process X(t)+phi1*X(t-1)+phi2X(t-2)=Epsilon(t) (1)

Homework Equations


We get the characteristic polynomial F(z)=z^2+phi1*z+phi2 (2)
The process is stationary if the roots z are within the unit circle.

The Attempt at a Solution


I can off course easily find the roots of the polynom:
z1=-phi1/2+sqrt(phi1^2/4-phi2) (3)
z2=-phi1/2-sqrt(phi1^2/4-phi2) (4)

Now we need to find the range for phi1 and phi2, such that the absolute value of z1 and z2 is <1. Since the phi's can be complex, this is a bit tricky and I'm stuck here. I tried substituting phi1=a+bi and phi2=c+di but I can't get rid of the squareroot on the RHS of (3) and (4). Any good ideas, how to solve this?
 
Physics news on Phys.org
Solved it. It's not the phi's that can get complex, only the roots. Then it's "easily" possible to find the allowed parameter range.
 

Similar threads

  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
1
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
Replies
10
Views
3K
Replies
39
Views
6K
Replies
4
Views
2K
Replies
9
Views
3K