The UtilityFunctions submodule

This submodule implements UtilityFunctions.R from the original R implementation in Python.

It contains functions useful for preprocessing training data.

mhn.full_state_space.UtilityFunctions.KL_div(p: ndarray, q: ndarray) → float

Computes the Kullback–Leibler divergence between two probability distributions.

Parameters:

• p (np.ndarray) – Probability distribution p.

• q (np.ndarray) – Probability distribution q.

Returns:

The KL-divergence of p and q.

Return type:

float

mhn.full_state_space.UtilityFunctions.data_to_pD(data: ndarray) → ndarray

Calculates the probability distribution for the different events from a given binary mutation matrix.

Parameters:

data (np.ndarray) – A numpy array or matrix representing the mutation data.

Returns:

The probability distribution of the different events.

Return type:

np.ndarray

mhn.full_state_space.UtilityFunctions.finite_sample(p_th: ndarray, k: int) → ndarray

Generates a random sample given a probability distribution and returns the probability distribution for the new sample.

Parameters:

• p_th (np.ndarray) – Probability distribution of events (the distribution of a true Theta).

• k (int) – The number of samples to generate.

Returns:

The probability distribution of events from the generated samples.

Return type:

np.ndarray

mhn.full_state_space.UtilityFunctions.state_to_int(x: ndarray) → int

Interprets a binary array as a binary number and returns the corresponding integer value.

Parameters:

x (np.ndarray) – Binary array, typically representing a row of the mutation matrix in this context.

Returns:

The integer value corresponding to the binary array.

Return type:

int