gemseo / mlearning / core

unsupervised module

Unsupervised machine learning algorithm

The unsupervised module implements the concept of unsupervised machine learning models, where the data has no notion of input or output.

This concept is implemented through the MLUnsupervisedAlgo class, which inherits from the MLAlgo class.

class gemseo.mlearning.core.unsupervised.MLUnsupervisedAlgo(data, transformer=None, var_names=None, **parameters)[source]

Bases: gemseo.mlearning.core.ml_algo.MLAlgo

Unsupervised machine learning algorithm.

Inheriting classes should overload the MLUnsupervisedAlgo._fit() method.

Constructor.

Parameters

  • data (Dataset) – learning dataset

  • transformer (dict(str)) – transformation strategy for data groups. If None, do not scale data. Default: None.

  • var_names (list(str)) – names of the variables to consider.

  • parameters – algorithm parameters

ABBR = 'MLUnupervisedAlgo'
learn(samples=None)[source]

Train machine learning algorithm on learning set.

Parameters

  • names (list(str)) – learning variables. Default: None.

  • samples (list(int)) – training samples (indices). Default: None.