SMCLA Algorithm Documentation

πŸ” Overview

The Soft MCLA (sMCLA) is a consensus clustering algorithm that combines soft membership matrices by treating clusters as vectors and grouping them into meta-clusters. This is the soft version of the MCLA ensemble algorithm.

βš™οΈ Class Definition

Class Name: SMCLA

class SMCLA:
    def __init__(self, n_clusters: int):
        self.n_clusters = n_clusters

Implements the soft MCLA algorithm that aggregates memberships by clustering cluster-vectors themselves and re-assigning to objects.

πŸ“‹ Parameters

Parameter

Type

Default

Description

n_clusters

int

β€”

Number of consensus clusters to output

πŸš€ Usage Examples

from soft_clustering._smcla._smcla import SMCLA
import numpy as np

# Simulate 3 soft clusterings
soft1 = np.random.dirichlet(np.ones(3), size=100)
soft2 = np.random.dirichlet(np.ones(3), size=100)
soft3 = np.random.dirichlet(np.ones(3), size=100)

model = SMCLA(n_clusters=3)
labels = model.fit_predict([soft1, soft2, soft3])
print("Consensus Labels:", labels)

πŸ“₯ Input / πŸ“€ Output

  • Input to fit_predict(soft_memberships):

    • soft_memberships (list of np.ndarray): List of soft clustering matrices (shape N x K)

  • Returns:

    • labels (np.ndarray): Consensus clustering labels (length N)

πŸ”§ Methods

  • __init__(self, n_clusters: int): Initializes the model with number of consensus clusters.

  • fit_predict(self, soft_memberships: List[np.ndarray]): Transposes all cluster matrices, clusters the vectors, and aggregates to produce final labels.

–

πŸ› οΈ Implementation Notes

  • Clusters are treated as feature vectors for meta-clustering.

  • KMeans is applied to these vectors to create meta-clusters.

  • Each data point is re-assigned based on aggregated contributions from cluster-vectors.

  • Final assignment is the max-weighted cluster.

πŸ“š Reference

  1. Punera, K., & Ghosh, J. (2008). Consensus-Based Ensembles of Soft Clusterings.