Download as pdf or txt
Download as pdf or txt
You are on page 1of 2

Assignment 2

28 February 2020

Introduction to probabilistic models for time series analysis


Due by March 10, 2020

This assigmment is aimed to help your study of the material presented in Lecture 3 (lecture notes are posted
on BBoard).
Assignments of this kind (“on the theory”) are not evaluated, and they are not mandatory.
Again, the intention if to provide you additional material for helping your distance learning, and to encourage
you to ask me (email me, we can arrange “virtual office hours”) if you find anything unclear.
Note:
• Some of the questions are of the kind you may find in the written proof in the final exam.
• Other questions involve using R, and are just further exercizes for you in learning R.

1. Question 1.
Consider a univariate time series (Yt ).
• Give the definition of the autocovariance function and of the autocorrelation function.
• What changes if the time series (Yt ) is stationary?

2. Question 2.(Written proof, May 2019 ).


Consider a univariate time series (Yt ).
• When is (Yt ) stationary? When is it strictly stationary?
• Plot the autocorrelation function for a stationary time series (Yt ) for which V (Yt ) = 1, Cov(Yt , Yt+h ) = .7
for h = 1 and zero for h = 2, 3, . . ..

3. Question 3.
i.i.d
Consider a Gaussian white noise process (Yt ), with Yt ∼ N (1, 1). Compute the mean function, the variance
functions and the autocovariance function.

4. Question 4. (If you are not familiar yet with ARMA models, you may skip this question).
Consider an autoregressive process of order p, AR(p), with p = 1. This means that (Yt ) satisfies the equation
i.i.d
Yt = αYt−1 + t , t ∼ N (0, σ 2 ).

• Assume that α = .7 and σ 2 = 1. Plot the autocorrelation function of (Yt ) and comment briefly.
Hint: An AR(p) process is a special case of the more general class of ARM A(p, q) models.
The R function arima provides tools for the analysis of ARMA process. In particular, the function ARMAacf
computes the theoretical autocorrelation function for an ARMA process.

1
?arima
?ARMAacf
out = ARMAacf(ar=.7, lag.max = 10)
plot(out)
1.0
0.8
0.6
out

0.4
0.2
0.0

2 4 6 8 10

Index

• Repeat the exercize, assuming α = −0.7.

5. Question 5.
Consider a random vector (X, Y ) such that
   2 
µx σ σ
(X, Y ) ∼ N , x .
µy σ σy2

• What is the marginal distribution of X?


• What is the conditional distribution of Y | X = x?
(See Appendix A in Petrone, Petris, Campagnoli textbook).

You might also like