Move adjacent neurons by precomputed alpha*(1-((i-j)^2/[r]^2)) in radpower[|i-j|]
Move neuron i towards biased (b,g,r) by factor alpha
Search for biased BGR values description: finds closest neuron (min dist) and updates freq finds best neuron (min dist-bias) and returns position for frequently chosen neurons, freq[i] is high and bias[i] is negative bias[i] = _gamma*((1/this._networkSize)-freq[i])
Original distance equation: dist = n.b - b; if (dist < 0) dist = -dist; a = n.g - g; if (a < 0) a = -a; dist += a; a = n.r - r; if (a < 0) a = -a; dist += a; a = (n.a - al); if (a < 0) a = -a; dist += a;
Insertion sort of network and building of netindex[0..255] (to do after unbias)
Main Learning Loop
Generated using TypeDoc
sampling factor 1..30