krepresentatives 1.1.2

Python implementations of the k-Representatives and LSH-k-Representatives algorithms for clustering categorical data:

Different from k-Modes algorithm, k-Representatives and LSH-k-Representatives define the "representatives" that keep the frequencies of all categorical values of the clusters.

Installation:

Using pip:

pip install krepresentatives

Import the packages:

import numpy as np
from krepresentatives.kRepresentatives import kRepresentatives

Generate a simple categorical dataset:

X = np.array([[0,0],[0,1],[0,0],[1,1],[2,2],[2,3],[2,3]])
y = np.array([0,0,0,0,1,1,1])

k-Representatives:

kreps = kRepresentatives(X,y,n_init=5,n_clusters=2 ,verbose=3)
kreps.fit_predict()

Built-in evaluattion metrics:

kreps.CalcScore()

Out come:

kRepresentatives Init 0
Iter 0  Cost: 8.00  Move: 0  Num empty: 0  Timelapse: 0.00
Iter 1  Cost: 4.83  Move: 0  Num empty: 0  Timelapse: 0.00
Iter 2  Cost: 4.83  Move: 0  Num empty: 0  Timelapse: 0.00
kRepresentatives Init 1
Iter 0  Cost: 9.48  Move: 0  Num empty: 0  Timelapse: 0.00
Iter 1  Cost: 6.50  Move: 1  Num empty: 0  Timelapse: 0.00
Iter 2  Cost: 5.33  Move: 0  Num empty: 0  Timelapse: 0.00
Iter 3  Cost: 5.33  Move: 0  Num empty: 0  Timelapse: 0.00
kRepresentatives Init 2
Iter 0  Cost: 9.08  Move: 0  Num empty: 0  Timelapse: 0.00
Iter 1  Cost: 7.60  Move: 0  Num empty: 0  Timelapse: 0.00
Iter 2  Cost: 7.60  Move: 0  Num empty: 0  Timelapse: 0.00
kRepresentatives Init 3
Iter 0  Cost: 9.31  Move: 0  Num empty: 0  Timelapse: 0.00
Iter 1  Cost: 6.50  Move: 1  Num empty: 0  Timelapse: 0.00
Iter 2  Cost: 5.33  Move: 0  Num empty: 0  Timelapse: 0.00
Iter 3  Cost: 5.33  Move: 0  Num empty: 0  Timelapse: 0.00
kRepresentatives Init 4
Iter 0  Cost: 9.42  Move: 0  Num empty: 0  Timelapse: 0.00
Iter 1  Cost: 7.60  Move: 0  Num empty: 0  Timelapse: 0.00
Iter 2  Cost: 7.60  Move: 0  Num empty: 0  Timelapse: 0.00
Score:  4.833333333333334  Time: 0.0015569399999999956
Purity: 1.00 NMI: 1.00 ARI: 1.00 Sil:  0.52 Acc: 1.00 Recall: 1.00 Precision: 1.00

Parameters:

X: Categorical dataset
y: Labels of object (for evaluation only)
n_init: Number of initializations
n_clusters: Number of target clusters
max_iter: Maximum iterations
verbose:
random_state:

Outputs:

cluster_representatives: List of final representatives
labels_: Prediction labels
cost_: Final sum of squared distance from objects to their centroids
n_iter_: Number of iterations
epoch_costs_: Average time for an initialization

LSH-k-Representatives: To be updated

References:

[1] San, Ohn Mar, Van-Nam Huynh, and Yoshiteru Nakamori. "An alternative extension of the k-means algorithm for clustering categorical data." International journal of applied mathematics and computer science 14 (2004): 241-247. [2] To be updated