Scribd
Upload
22 views

HW 3

This document contains homework problems for a time series analysis course. It includes 4 problems: 1) Analyzing the autocorrelation and invertibility of an MA(2) process and expressing it as an infinite AR process. 2) Analyzing the stationarity and autocorrelations of an AR(2) process and expressing it using a difference equation and as an infinite MA process. 3) Identifying the order of an AR process based on given autocorrelations. 4) Finding coefficients for linear difference equations that generate sinusoidal and damped sinusoidal sequences.

Uploaded by

melof58728
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
22 views

HW 3

This document contains homework problems for a time series analysis course. It includes 4 problems: 1) Analyzing the autocorrelation and invertibility of an MA(2) process and expressing it as an infinite AR process. 2) Analyzing the stationarity and autocorrelations of an AR(2) process and expressing it using a difference equation and as an infinite MA process. 3) Identifying the order of an AR process based on given autocorrelations. 4) Finding coefficients for linear difference equations that generate sinusoidal and damped sinusoidal sequences.

Uploaded by

melof58728
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

STAT 520 HW 3 SPRING 2024

1. Consider the MA(2) process zt = at − at−1 + 29 at−2 , var[at ] = σa2 = 3.

(a) Find the variance and autocorrelation function of {zt }.


(b) Determine if the process is invertible,
P∞ and if so express the process in the infi-
nite autoregressive form, zt = j=1 πj zt−j + at , providing numerical values for
π1 , . . . , π5 .
(c) Show that πj − θ1 πj−1 − θ2 πj−2 = 0 for j > 0, where π−1 = 0. Find a general
expression for πj .

2. Consider the AR(2) process zt = 1.6zt−1 − .89zt−2 + at , σa2 = 2.

(a) Verify that {zt } is stationary, and find the variance γ0 .


(b) Solve the first 2 Yule-Walker equations for ρ1 and ρ2 , and then compute ρk for
k = 3, . . . , 8 using the recursive equation ρk = φ1 ρk−1 + φ2 ρk−2 .
(c) Give a general formula for ρk as the solution of the linear difference equation,
ρk − 1.6ρk−1 + .89ρk−2 = 0, k > 0. Do express your result in terms of real numbers
and real functions.
P
(d) Consider the infinite moving average form of the process, zt = at + ∞ j=1 ψj at−j .
Show that ψj − 1.6ψj−1 + .89ψj−2 = 0, j > 0, where ψ−1 = 0, and find a general
expression for ψj .

3. Assume that the stationary process {zt } is generated by a low order AR model with
p ≤ 4, and that {zt } has the following variance and autocorrelation function,

γ0 = 12, ρ1 = .8, ρ2 = .46, ρ3 = .152, ρ4 = −.0476.

(a) For each of the possible orders 1, 2, 3, solve the first sets of Yule-Walker equations
to obtain the corresponding φj ’s and σa2 .
(b) What is the actual AR order p of this process?

4. Suppose a sequence of real numbers {xt }∞


0 satisfy the linear difference equation,

xt + axt−1 + bxt−2 = 0, t > 1.

(a) Find the coefficients a, b, such that xt is sinusoidal with period 12,
xt = a1 cos(2πt/12) + a2 sin(2πt/12).
(b) Find the coefficients a, b, such that xt is damped sinusoidal with damping factor
.8 and period 12,
xt = (.8)t {a1 cos(2πt/12) + a2 sin(2πt/12)}.

Due February 12, 2024

You might also like