During execution, the network's inputs are set and
each node computes its activation which is a measure
of how closely its stored pair matches the network's
inputs. If p is the pair stored in node n, then the
activation for n is defined to be the ratio of the number
of p's asserted inputs matched by the network's inputs
to p's number of asserted inputs. This measure is
identical to the measure of conceptual similarity of
[20]. The node with highest activation wins and the
output of its stored pair becomes the output of the
network. In case of a tie, priority is first given to the
node whose pair is more specific, then to the one that
has the highest priority value, and finally to the one
whose cover count is highest. This process is easily
handled by the broadcast and gather facility