$E[X/Y] = E[X] / E[Y]$

$E[X + Y] = E[X] + E[Y]$

$E[X^2] = (E[X])^2$

$E[XY] = E[X]E[Y]$

False

True

As a random variable that can only take on values along the imaginary axis.

As a random variable whose outcomes are rotation matrices.

As a single real-valued random variable with a complex probability measure.

As an ordered pair of real-valued random variables, $Z = X + iY$.

$E[Z] = \sqrt{(E[X])^2 + (E[Y])^2}$

$E[Z] = E[X] + iE[Y]$

$E[Z] = E[X \cdot Y]$

$E[Z] = E[X] - iE[Y]$

$E[Z^*] = E[Z]$

$E[Z^*] = (E[Z])^*$

$E[Z^*] = -E[Z]$

$E[Z^*] = 1/E[Z]$

$P(X \ge a) \le E[X]$

$P(X \ge a) \le E[X] / a$

$P(X \ge a) \le Var(X) / a^2$

$P(X \ge a) \le a / E[X]$

It only applies to discrete random variables.

It requires the random variable to be non-negative.

The bound it provides is often not very tight (i.e., it can be very loose).

It is computationally very expensive to calculate.

The product of two independent random variables.

A complex random variable.

A random sum of random variables.

A continuous-time stochastic process.

Poisson random variables

Rademacher random variables

Uniform random variables

Gaussian random variables

The signal's maximum and minimum frequencies.

The in-phase (I) and quadrature (Q) components.

The signal's average power and its peak power.

The signal's start and end times.

most likely outcome of a single measurement.

exact, deterministic value of the observable before measurement.

eigenvalue of the system's wavefunction.

average of a large number of measurements of that observable on identical systems.

The 'E-step' calculates the expected value of the log-likelihood function, with respect to the conditional distribution of latent variables.

It calculates the expected value of the model parameters.

It directly minimizes the expected error on a test dataset.