Artificial neural network

Workings of a network

Cells

A network consists of cells that have either positive or negative connections to other cells

Calculus

Sigmoid function

This is a function that returns values between 0 and 1, and gives around 0.5 when given a zero:

Logistic function
hyperbolic tangent
arctangent function
Gudermannian function
Error function
Generalised logistic function
Smoothstep function

Calculus

Sigmoid function - Logistic function

function sigmoid(input) {
	return 1
	    / (1 + Math.pow(Math.E, -input));
}

Calculus

Communications between the cells

// Calculate

const outputCell1 = sigmoid(
    weigthsCell1[0]
    + weigthsCell1[1] * inputs[0]
    + weigthsCell1[2] * inputs[1]
    + weigthsCell1[3] * inputs[2]
);
const outputCell2 = sigmoid(
    weigthsCell2[0]
    + weigthsCell2[1] * inputs[0]
    + weigthsCell2[2] * inputs[1]
    + weigthsCell2[3] * inputs[2]
);

const inputs = [
    1,
    0,
    1,
];

const weigthsCell1 = [
    3,
    5,
    -2,
    -10,
];

const weigthsCell2 = [
    -7,
    -1,
    -4,
    -14,
];