Prerequisite: ANN | Self Organizing Neural Network (SONN)
In the Self Organizing Neural Network (SONN), learning is performed by shifting the weights from inactive connections to active ones. The neurons which were won are selected to learn along with their neighborhood neurons. If a neuron does not respond for a specific input pattern, then learning will not be performed in that particular neuron.
Self-Organizing Neural Network Learning Algorithm:
- Initialize synaptic weights to random values in a specific interval like, [-1, 1] or [0, 1].
- Assign topological neighborhood parameters.
- Define learning rate (say, 0.1).
Step 1: Until termination condition is reached, do loop: Steps 2-8
This is a criteria to finding similarity between two sets of samples. The nodes (neurons) in the network are evaluated to determine the most likely input vector according to its weights
Where: is the number of neurons in the input layer,
is the number of neurons in the Kohonen layer.
The winning node is generally termed as the Best Matching Unit (BMU).
Where is the weight correction at iteration .
This process to update weight is based on the competitive learning rule:
Where is the learning rate.
Here the neighborhood function centered around the winner-takes-all neuron at iteration . Any neurons within the radius of the BMU are modified to make them more similar to the input vector.
Example with iterations:
Take an input vector of 2 – Dimension:
- The initial weight vectors, , are given by
- We find the winning (best-matching) neuron satisfying the minimum distance Euclidean criterion:
- Neuron 3 is the winner and its weight vector is updated following the competitive learning rule.
- The updated weight vector at iteration is calculated as:
- The weight vector of the winning neuron 3 becomes closer to the input vector in each iteration.