soft.fuzzy.unsupervised.cluster.online package
Submodules
soft.fuzzy.unsupervised.cluster.online.ecm module
Implements the Evolving Clustering Method as described in the DENFIS paper.
- class soft.fuzzy.unsupervised.cluster.online.ecm.LabeledClusters(clusters, labels)
Bases:
tuple- clusters
Alias for field number 0
- labels
Alias for field number 1
- soft.fuzzy.unsupervised.cluster.online.ecm.apply_evolving_clustering_method(dataset: SupervisedDataset, config: Config) LabeledClusters
Apply the Evolving Clustering Method to the input data.
- Parameters:
dataset – The given dataset.
config – The configuration settings.
- Returns:
The resulting clusters.
- soft.fuzzy.unsupervised.cluster.online.ecm.general_euclidean_distance(vector_1: Tensor, vector_2: Tensor) Tensor
The general Euclidean distance metric as described in the DENFIS paper.
- Parameters:
vector_1 – A two-dimensional Tensor, where rows are number of observations and columns are features.
vector_2 – A one-dimensional Tensor, where the columns are features.
- Returns:
A one-dimensional Tensor, where the columns are the distances from each identified cluster thus far.