Toggle menu

Gaussian Mixture Model (GMM)

gmm.GMM( K, itr=100, name="gmm", category=None, mode="learn" )

gmm.GMM is a module for unsupervised classification based on a Gaussian mixture model. It computes the probabilities that each data element is classified into each class and the means of the distributions and sends them to the connected modules.


Parameter Type Description
K int Number of clusters
itr int Number of iterations
name string Module name
category array Correct class labels
mode string Choose from learning mode (“learn”) or recognition mode (“recog”)



# import necessary modules
import serket as srk
import gmm
import numpy as np

data = np.loadtxt( "data.txt" )  # load data
data_category = np.loadtxt( "category.txt" )  # load correct labels

# define the modules
obs = srk.Observation( data )  # send the observation to the connected module
gmm1 = gmm.GMM( 10, catogory=data_category )  # classify into ten classes

# construct the model
gmm1.connect( obs )  # connect obs to gmm1

gmm1.update()  # train gmm1