Algorithm

Algorithm

• 1: Initialize the:
  • network weights $\mathbf{u}_k=\left(u_{k 1}, u_{k 2}, \cdots, u_{K I}\right)$
  • the learning rate $\eta$
  • neighbourhood radius $\kappa$ While stopping conditions are not met:
• 2: for each input pattern $p$:
  • compute the Euc distance $d_{k, p}$ between $\mathbf{z}_p$ and $\mathbf{u}_k$
  • Find the winning output unit $o_k$ (smallest $d_{k, p}$)
  • Update weights for the neighbourhood $\kappa_{k, p}$
• 3: Update learning rate the learning rate $\eta(t)$
• 4: Reduce neighbourhood radius $\kappa(t)$