Pengenalan pola huruf latin tulisan tangan menggunakan jaringan saraf tiruan adaptive resonance theory 2

• Main Author: Perpustakaan UGM, i-lib
• Format: Article NonPeerReviewed
• Published: [Yogyakarta] : Universitas Gadjah Mada 1997
• Subjects: Jurnal i-lib UGM
• Online Access: https://repository.ugm.ac.id/18565/; http://i-lib.ugm.ac.id/jurnal/download.php?dataId=1357
• Institution: Universitas Gadjah Mada

ABSTRACT This paper deals with an application of an artificial neural network Adaptive Resonance Theory 2 (ANN ART 2) for hand writing latin character recognition. An adaptive resonance theory 2 was used for this clustering problem. The influence of vigilance parameter variation upon the number of clusters formed during the training phase was investigated. The result shows that if the vigilance parameter increases then the number of clusters formed will also increase. The experiment was continued by testing the trained ANN to recognize hand writing latin characters. For the vigilance parameter value of 0.95 and 95 number of data training, the ANN can recognize them in the range between 22.2 % and 83.3 %. When the vigilance parameter is 0.95 and the number of data training is 240, the ANN can recognize hand writing latin characters in the range between 50 % and 92.3 %.