Design of a compact cluster structure by using genetic algorithms
Hong Lan Jin, Yoshikazu Miyanaga, Koji Tochinai · 2002
Cluster nodes produced in the experiments of our handwritten Japanese character recognition involve the enormous increasement of computation quantity. It is necessary to delete unnecessary cluster nodes to get the compact cluster structure. A compress algorithm using genetic algorithms (GA) is employed in this report to reduce redundant cluster nodes. The results of the handwritten Japanese character recognition using GA are also shown.