prism.mixture module

class prism.mixture.BetaBinomialMixture(n_components=1, max_iter=10000, tol=0.001, seed=None, verbose=False)[source]

Bases: object

Beta-binomial mixture model for PRISM core algorithm.

bic()[source]
Returns:Bayesian Information Criterion (BIC) value of the model.
dispersions_
fit(n, k, headers)[source]
get_counts()[source]
Returns:Fingerprint pattern counts of fingerprint epiloci used for fitting the model.
get_depths()[source]
Returns:Depths of fingerprint epiloci used for fitting the model.
get_dispersions()[source]
Returns:Cluster dispersions.
get_headers()[source]
Returns:Headers of fingerprint epiloci used for fitting the model.
get_means()[source]
Returns:Cluster means.
get_n_components()[source]
Returns:Number of clusters.
get_n_dimensions()[source]
Returns:Number of dimensions.
get_weights()[source]

Returns the list of cluster weights. Note that a cluster weight is computed as a sum of posterior probabilities that each of the data point will be assiged to that cluster.

Returns:Cluster weights.
means_
predict_proba(n, k)[source]

Returns the posterior probabilities of each fingerprint epiloci for each cluster.

Parameters:
  • n (list) – Depths of fingerprint epiloci.
  • k (list) – Fingerprint pattern counts of fingerprint epiloci.
Returns:

Posterior probabilities of each fingerprint epiloci.

prism.mixture.trigamma(x)[source]