Antecedent Set Estimator

class pyfume.EstimateAntecendentSet.AntecedentEstimator(x_train, partition_matrix)

Bases: object

Creates a new antecedent estimator object.

Parameters

  • x_train – The input data.

  • partition_matrix – The partition matrix of the input data (generated by a clustering the data).

determineMF(mf_shape='gauss', merge_threshold=1.0)

Estimates the parameters of the membership functions that are used as antecedents of the fuzzy rules.

Parameters

  • mf_shape – The desired shape of the fuzzy sets. The user can choose from ‘gauss’ (gaussian), ‘gauss2’ (double gaussion) or ‘sigmf’ (sigmoidal) (default = gauss).

  • merge_threshold – Threshold for the merging of fuzzy sets for the GRABS approach. By default no merging takes place (default = 1.0).

Returns

mu, sigma; if gauss2: mu1, sigma1, mu2, sigma2)

Return type

A list with the estimated parameters of the membership functions (format if gauss