Skip to main content

modules.clustering.NBClustering

NBClustering Objects​

class NBClustering(Experiment)

This module computes Naive Bayes Clusters. This version only works for Bernoulli variables.

This version also computes

__init__​

 | __init__(count_variables, num_clusters=15, epsilon=0.01, min_support=0.01, soft=False, lmbda=1.0, supplementary_variables=None, basic_statistics=None, custom_statistics=None)

Attributes:

  • count_variables - List[str] A list with the variables to take into account.
  • num_clusters - The number of clusters to compute.
  • min_support - Variable values with lower support will be ignored.

do​

 | do(data)

Performs clustering of the dataset.

Attributes:

  • data - Should be a Stream.

asvtd​

 | asvtd(X, k)

Learn an approximate pair M, omega @param X: the dataset @param k: the number of clusters

EM​

 | EM(X, M, omega, Eps=0.001, verbose=False)

Implementation of EM to learn a NBM with binary variables @param X: the dataset @param M: the centers of the mixture @param omega: the mixing weights @param Eps: the stopping criterion @param verbose: whether to show or not the error