Skip to content

komm.DiscreteMemorylessSource

Discrete memoryless source (DMS). It is defined by an alphabet $\mathcal{X}$ and a probability mass function (pmf) $p_X$. Here, for simplicity, the alphabet is always taken as $\mathcal{X} = [0 : |\mathcal{X}|)$. The pmf $p_X$ gives the probability of the source emitting the symbol $X = x$.

Parameters:

  • pmf (ArrayLike | int) –

    Either the source pmf $p_X$, in which case the element at position $x \in \mathcal{X}$ must be equal to $p_X(x)$, or an integer $n \geq 1$, in which case a uniform distribution over $n$ symbols is used.

Examples:

>>> komm.DiscreteMemorylessSource([1/2, 1/4, 1/8, 1/8])
DiscreteMemorylessSource(pmf=[0.5, 0.25, 0.125, 0.125])

>>> komm.DiscreteMemorylessSource(4)
DiscreteMemorylessSource(pmf=[0.25, 0.25, 0.25, 0.25])

cardinality int cached property

The cardinality $|\mathcal{X}|$ of the source alphabet.

entropy()

Returns the source entropy $\mathrm{H}(X)$. See komm.entropy for more details.

Parameters:

  • base (LogBase) –

    See komm.entropy. The default value is $2.0$.

Examples:

>>> dms = komm.DiscreteMemorylessSource([1/2, 1/4, 1/8, 1/8])
>>> dms.entropy()
np.float64(1.75)
>>> dms.entropy(base=4)
np.float64(0.875)

__call__()

Returns random samples from the source.

Parameters:

  • shape (int | tuple[int, ...]) –

    The shape of the output array. If shape is an integer, the output array will have shape (shape,). The default value is (), which returns a single sample.

Returns:

  • NDArray[integer] –

    An array of shape shape with random samples from the source.

Examples:

>>> dms = komm.DiscreteMemorylessSource([0.5, 0.4, 0.1])
>>> dms()
array(1)
>>> dms(10)
array([0, 1, 1, 0, 2, 1, 1, 0, 0, 0])
>>> dms((2, 5))
array([[2, 1, 1, 0, 0],
       [1, 0, 1, 1, 1]])