A simple, illustrative implementation of a single-layered perceptron in Java. When a pattern is impressed on the perceptron the

activation of the network is adjusted according to an activation formula and a given bias value. To adjust the way the perceptron

reacts to a given input, a learning algorithm (the delta rule) is implemented to adjust the weights connecting the neurons of the

perceptron, which are initially set to 0. This rule keeps adjusting the weights until the resulting output for a given input

corresponds to a supplied correct output, resulting in a perceptron trained to react to a certain perception in a certain way.

The input patterns and the teaching output are hard coded as integer matrices.-A simple, illustrative implementation of a single-layered perceptron　in Java. When a pattern is impressed on the perceptron the

activation of the network is adjusted according to an activation formula and a given bias value. To adjust the way the perceptron

reacts to a given input, a learning algorithm (the delta rule) is implemented to adjust the weights connecting the neurons of the

perceptron, which are initially set to 0. This rule keeps adjusting the weights until the resulting output for a given input

corresponds to a supplied correct output, resulting in a perceptron trained to react to a certain perception in a certain way.

The input patterns and the teaching output are hard coded as integer matrices.