A Tutorial on Clustering Algorithms


Introduction | K-means | Fuzzy C-means | Hierarchical | Mixture of Gaussians | Links


Fuzzy C-Means - Interactive demo

This applet requires Java Runtime Environment version 1.3 or later. You can download it from the Sun Java website.

GETTING STARTED

  • Choose how many data and clusters you want and then click on the Initialize button to generate them in random positions.
  • Move data and clusters along x-axis as you like by clicking and dragging.
  • Set a value of Fuzzines (1.2 - 2) and Accuracy (0.01 - 0.3) coefficients.
  • Click on Start to begin the simulation. During simulation data and clusters positions are fixed.
  • Go on using either Step or Run until the end of the simulation. Current number of steps is shown.
  • Use the Reset button to go back to the initial configuration. Now you can move existing data and clusters or generate new ones and then begin another simulation.

NOTE: The value of the membership function is computed only in the points where there is a datum. The tracing of the function is then obtained with a linear interpolation of the previously computed values. As a result, you get a broken line that is slightly different from the real membership function. In particular the peak of the drawn function could not correspond to the real one. This is why in some cases peaks and centroids are placed in different positions.


Back to Fuzzy C-means