numbasom 0.0.1

A Fast Self-Organizing Map Python Library Implemented in Numba

NumbaSOM

A Fast Self-Organizing Map Python Library Implemented in Numba.

If you need a fast and simple to use SOM library implemented as a 2D lattice or torus, check this out. It utilizes online rather than batch training.

Install

pip install numbasom

How to use

Train a SOM on 1000 random 3-dimensional vectors:

import numpy as np
from numbasom import SOM, u_matrix, plot_u_matrix

Load some data

data = np.random.randn(100,3)

Initialize the library

som = SOM(som_size=(20,20))

Train

lattice = som.train(data, num_iterations=1000)

Data scaling took: 0.338414 seconds.
SOM training took: 1.049132 seconds.

Display the value in the first row and first column of the lattice

lattice[1::6,1]

array([[0.72234145, 0.20699402, 0.48731189],
       [0.49514563, 0.18585944, 0.64291118],
       [0.39127316, 0.13052984, 0.4517087 ],
       [0.40195937, 0.42461381, 0.14699424]])

Make U-matrix

um = u_matrix(lattice)

Plot U-matrix

plot_u_matrix(um)