plotFuzzyClusters

Plot data clusters for fuzzy-c-means clustering

Since R2025a

Syntax

plotFuzzyClusters(data,U)

plotFuzzyClusters(data,U,Name=Value)

h = plotFuzzyClusters(___)

Description

plotFuzzyClusters(data,U) plots fuzzy clusters using input data data and fuzzy partition matrix U. To compute U, first cluster the data using the fcm function.

The plot contains an N_f-by-N_f array of axes, where N_f is the number of features in data. In the plot:

• Each data point is classified into the cluster for which the data point has the highest membership value.

• The nondiagonal axes show 2-D plots for each pairwise feature combination.

• The diagonal axes show the marginal cluster membership values of the data points with respect to each feature.

plotFuzzyClusters(data,U,Name=Value) specifies options using one or more name-value arguments. For example, plotFuzzyClusters(data,U,ClusterCenters=centers) plots the clustered data along with the computed cluster centers specified by centers.

h = plotFuzzyClusters(___) returns a tiled chart layout that you can use to modify the plot properties. Use this syntax to customize the plot appearance or when plotting fuzzy clusters in an App Designer app.

Examples

Plot Fuzzy Clusters

Load the data to cluster. The three columns in clusterDemo correspond to the three features in the data.

load clusterDemo.dat

Cluster the data into three clusters using the fcm function.

options = fcmOptions(...
    NumClusters=3,...
    Verbose=false);
[centers,U,objFun] = fcm(clusterDemo,options);

Plot the clusters using the fuzzy partition matrix.

plotFuzzyClusters(clusterDemo,U)

Because there are three features in the data, the plot shows a 3-by-3 grid of axes. The nondiagonal axes show the classified data points for each pair of features. The diagonal axes show the marginal cluster membership values across each feature range.

By default, plotFuzzyClusters plots each data point. Alternatively, you can interpolate the data points and create surface plots.

plotFuzzyClusters(clusterDemo,U,PlotType="surface")

Plot Fuzzy Clusters and Cluster Centers

Load the data to be clustered.

load fcmdata.dat

Cluster the data into two clusters.

options = fcmOptions(...
    NumClusters=2,...
    Verbose=false);
[centers,U] = fcm(fcmdata,options);

Plot the clustered data along with the cluster centers.

plotFuzzyClusters(fcmdata,U,ClusterCenters=centers)

In each plot, the cluster centers are shown as diamonds.

Customize Fuzzy Cluster Plot

Load the data to cluster. For this example, create a random data set.

data = rand(100,2);

Cluster the data into two clusters.

options = fcmOptions(...
    NumClusters=2,...
    Verbose=false);
[centers,U] = fcm(data,options);

Plot the clusters and return a handle to the resulting tiled chart layout.

h = plotFuzzyClusters(data,U);

You can customize the plot by modifying its properties using the handle, which is useful when adding a fuzzy cluster plot to an App Designer app.

For this example, modify the plot title.

h.Title.String = "FCM Clustering Results";

Input Arguments

data — Data set to be clustered

N_d-by-N_f matrix

Data set to be clustered, specified as an N_d-by-N_f matrix, where N_d is the number of data points and N_f is the number of features in each data point.

U — Fuzzy partition matrix

N_c-by-N_d matrix

Fuzzy partition matrix returned by the fcm function, specified as an N_c-by-N_d matrix. Element U(i,j) indicates the degree of membership μ_ij of the jth data point in the ith cluster.

Name-Value Arguments

Specify optional pairs of arguments as Name1=Value1,...,NameN=ValueN, where Name is the argument name and Value is the corresponding value. Name-value arguments must appear after other arguments, but the order of the pairs does not matter.

Example: PlotType="surface" plots a surface plot.

ClusterCenters — Cluster centers

[] (default) | N_c-by-N_f matrix

Cluster centers returned by the fcm function, specified as an N_c-by-N_f matrix, where N_c is the number of clusters and N_f is the number of data features.

To omit the cluster centers from the plot, set ClusterCenters to [].

Dependencies

• The number of rows in ClusterCenters must match the number of rows in U.

• The number of columns in ClusterCenters must match the number of columns in data.

PlotType — Plot type

"scatter" (default) | "surface"

Plot type, specified as one of these values:

• "scatter" — Scatter plot of data points specified in data

• "surface" — Surface plot with interpolated regions between specified data points

ShowLegend — Option to display legend

false (default) | true

Option to display a legend, specified as one of these values:

• true — Display a legend.

• false — Do not display a legend.

Output Arguments

h — Plot handle

TiledChartLayout object

Plot handle, returned as a TiledChartLayout object.

Version History

Introduced in R2025a