This is a visualization of how different clustering algorithms learn clusters for some different datasets. You can also experiment with how hyperparameter settings affect the algorithms.
Select Dataset ?
Set Hyperparameters ?
Visualizer contains five generated two-dimensional datasets that can be used to demonstrate how different clustering
algorithms learn clusters for different problems. Each algorithm learnsx clusters in different ways, which is visualized
in this demonstrator.
The visualization shows the decision boundaries for the clusters an algorithm learns on the dataset.
The small circles show the examples in the dataset, and the colored areas show which cluster each point belongs
to. The small red squares show important concepts of an algorithm, such as centroids in K-means and Mean-Shift or
core instances in DBSCAN.
Click ► to expand the Set Hyperparameters section.
The hyperparameters (configuration of algorithms) are set so each algorithm learn good clusters on each dataset.
You can change the hyperparameters to see how different configurations affect how algorithms learn clusters.
Clustering is based on calculating distances between examples, and there are a wide range of different distance
functions that can be used. For numerical attributes, it is common to use the Euclidean distance (or L2-norm).
Two other common distance functions are Manhattan distance (L1-norm) and Chebyshev distance (L∞-norm).
Consider the figure below. We shall calculate the distance between the two points A and B. The
length of the two sides are denoted as x and y.
The formulas for calculating the different distance metrics are then:
About Web Clustering Demonstrator
Web Clustering Demonstrator is a demonstrator for clustering algorithms running purely on the client browser. All algorithms
have all the functionality of state-of-the-art implementations. The main purpose of this demonstrator is to be used as a
tool when teaching and explaining clustering and clustering related concepts.
Web ML Demonstrator is developed by Johan Hagelbäck, senior lecturer at Linnaeus University in Kalmar, Sweden. Contact details
for the developer is here.