1

INDIAN INSTITUTE OF TECHNOLOGY MADRAS

Department of Chemical Engineering
CH5350 Applied Time-Series Analysis

Assignment 1
Due: Monday, August 22, 2016

1.

[Distributions]
(a) Assume that the arrival times of the Institute bus at your hostel follows an exponential
distribution with an average rate of = 4 buses / hour. Then (i) if you arrived at the
bus stop at 7:50 AM, what is the expected time of the next bus?, (ii) a walker-by told
you that the last bus left at 7:45 AM, when is the next bus expected to arrive?
(b) Temperature on the campus between 7:00 AM - 8:00 AM is known to follow a Gaussian
distribution. The statistical properties of the temperature for this duration vary with the
month. During August, the average temperature is = 27.3 C with = 2.2 C. If
PrakritiMitr decides to walk to classes only if Pr(T > 25) is at most 0.2 in the morning
hours, determine if he will walk to the class in August.

2.

[Density function]
A random variable has the p.d.f.
f (x) =

c(x + 0.5 x), x [0, 1],

0,

otherwise

Determine (i) the constant c, (ii) E(1/ X), (iii) p.d.f. of Y = X 2 . Verify your answer to part
(iii) empirically using R (use the supplementary notes for constructing the empirical distribution
of Y ).
3.

[Joint density]
If two random variables have joint density

x/y ey

K e
y
f (x, y) =

x > 0, y > 0
elsewhere

Find (i) the value of K (ii) marginal density of X, (iii) the probability Pr(0 < X < 1,
0.2 < Y < 0.4) (iv) conditional expectation E(Y |X). Carry out numerical integration (use
integrate in R) wherever necessary.
4.

[Joint distribution]
(a) The joint cumulative distribution function of two continuous random variables is given by
1
F (x, y) = xy(x + y),
6

0 x 2, 0 y 1

2
Determine (i) the c.d.f. of y and (ii) the joint p.d.f. of x and y. Plot the three functions
using R.
(b) Show that if two random variables with joint Gaussian distribution are uncorrelated, they
are also independent.
5.

[Expectations]
(a) Show that for two RVs X and Y that have a joint Gaussian distribution, the conditional
expectation E(Y |X) is a linear function of X.
(b) Given three zero-mean, unit variance, independent random variables X1 , X2 and X3 ,
compute E(X13 (X22 + 3X3 ))

6.

[Correlations in R]
(a) The covariance between two RVs is estimated from their samples x[n] and y[n] as

yx =

N
1 X
(y[k] y)(x[k] x
)
N
k=1

where x
and y are the sample means of X and Y , respectively and N is the sample size.
Write a function in R to calculate this sample covariance matrix given samples of two
random variables. Test your code on the case X N (1, 2) and Y = 2X 2 + 4X (use the
rnorm routine to generate N = 1000 samples of X) by comparing the resulting covariance
matrix with the the values obtained from cov or the var command in R. Finally, show by
means of simulation that the estimate
yx tends to the theoretical value as N .
(b) For the partial correlation example in class, X = 2Z + 3V, Y = Z + W , do the following
based on 200 observations of X and Y generated from a random sample of Z, V and
W , all from a zero-mean, unit-variance Gaussian distribution. Verify the relation between sample partial correlation and regression coefficients, (ii) compute the theoretical
semi-partial correlation between X and Y given Z, i.e., correlation between X and
conditioned Y (Y |Z). Is this the same as partial correlation between X and Y ? Compare
your answer with the estimate obtained in R.