Solution (source code)

= Solution

For these counts, start with any $0<\theta^{(0)}<1$. The conditional missing count from part (ii) has <mean>
$$
z_r=\mathbb E[Z\mid\mathbf x,\theta^{(r)}]
=\frac{340\theta^{(r)}}{7+4\theta^{(r)}}.
$$
The E-step objective is
$$
Q(\theta\mid\theta^{(r)})=(5+z_r)\log\theta+30\log(1-\theta)+\mathrm{constant}.
$$
Its strictly <concave> M-step is explicit:
$$
\boxed{\theta^{(r+1)}=\frac{5+z_r}{35+z_r},\qquad
z_r=\frac{340\theta^{(r)}}{7+4\theta^{(r)}}.}
$$
Equivalently, the update map is $F(\theta)=(35+360\theta)/(245+480\theta)$. Its fixed-point equation is
$$
480\theta^2-115\theta-35=0.
$$
The unique root in $(0,1)$ is the <maximum-likelihood estimate>
$$
\boxed{\widehat\theta=\frac{115+\sqrt{80425}}{960}\approx0.4152.}
$$
For example, starting at $1/2$ gives $0.443299$, then approximately $0.425065$, and subsequent iterates approach the displayed root. Convergence is also transparent without relying only on the general EM theorem: $F$ is increasing and
$$
F(\theta)-\theta=\frac{35+115\theta-480\theta^2}{245+480\theta}.
$$
This is positive below the feasible fixed point and negative above it. Monotonicity of $F$ prevents crossing that fixed point, so the iterates converge monotonically toward it from either side. The original <log-likelihood> is strictly <concave> and diverges to minus infinity at both endpoints, establishing that the fixed point is the unique global maximum.