Main Content

plotsomplanes

Plot self-organizing map weight planes

Syntax

plotsomplanes(net)

Description

plotsomplanes(net) generates a set of subplots. Each ith subplot shows the weights from the ith input to the layer’s neurons, with the most negative connections shown as black, zero connections as red, and the strongest positive connections as yellow.

The plot is only shown for layers organized in one or two dimensions.

This plot supports SOM networks with hextop and gridtop topologies, but not tritop or randtop.

This function can also be called with standardized plotting function arguments used by the function train.

Examples

Plot SOM Weight Planes

x = iris_dataset;
net = selforgmap([5 5]);
net = train(net,x);

Figure Neural Network Training (20-Jul-2024 16:53:14) contains an object of type uigridlayout.

plotsomplanes(net)

Figure SOM Input Planes (plotsomplanes) contains 4 axes objects. Axes object 1 with title Weights from Input 1 contains 25 objects of type patch. Axes object 2 with title Weights from Input 2 contains 25 objects of type patch. Axes object 3 with title Weights from Input 3 contains 25 objects of type patch. Axes object 4 with title Weights from Input 4 contains 25 objects of type patch.

Version History

Introduced in R2008a