a. Find the pdf of x = [a b] a 2-d vector, where a is a Bernoulli random variable and
b is a Gaussian random variable. Assume, θ is the parameter for Bernoulli which gives
the probability of a = 1, that is p(a=1) = θ. Assume, b follows a Gaussian distribution with
mean m and variance σ
2
. The covariance of x is [1]
[ θ(1-θ) 0;
0 σ
2
]
b. Let the pdf of part (a) be p(x). Now assume that there are N iid samples drawn from
this pdf. Find θ that maximizes the joint probability q(x) of these N samples. Once you
determine q(x), use ln q(x) for computing θ.
Comments
Leave a comment