Perceptron in pure PHP: standalone example

Classic Rosenblatt perceptron (standalone run)

A minimal implementation of the classic perceptron: weighted sum, threshold activation, and training with the Rosenblatt update rule.

Example of use


                
<?php

include 'code.php';

$X = [
    [0, 0],
    [0, 1],
    [1, 0],
    [1, 1],
];

$y = [0, 0, 0, 1];

$perceptron = new Perceptron(2, 0.1);
$perceptron->train($X, $y, 20);

echo 'Perceptron (Rosenblatt learning rule) for logical AND:' . PHP_EOL;
echo '-----------' . PHP_EOL;
foreach ($X as $sample) {
    echo '[' . implode(', ', $sample) . '] => ' . $perceptron->predict($sample) . PHP_EOL;
}

echo PHP_EOL . 'Weights and bias:' . PHP_EOL;
echo '-----------' . PHP_EOL;
echo 'w1 = ' . $perceptron->getWeights()[0] . PHP_EOL;
echo 'w2 = ' . $perceptron->getWeights()[1] . PHP_EOL;
echo 'b = ' . $perceptron->getBias() . PHP_EOL;

How one perceptron makes a decision

The animation shows the signal flow: inputs are multiplied by weights, summed with bias, passed through a step function, and converted into a binary output for the AND rule.

Δw1 = lr · error · x1 Δw2 = lr · error · x2 Δb = lr · error

Epoch (epoch)

Sample (sample)

Target y (target y)

Error (error)

#	x	y	ŷ	w1, w2, b

Result: Memory: 0.008 Mb Time running: 0.001 sec.


                Perceptron (Rosenblatt learning rule) for logical AND:
-----------
[0, 0] => 0
[0, 1] => 0
[1, 0] => 0
[1, 1] => 1

Weights and bias:
-----------
w1 = 0.2
w2 = 0.1
b = -0.2