PENERAPAN METODE BACKPROPAGATION DAN ICZ-ZCZ PADA PENGENALAN POLA TULISAN TANGAN AKSARA BIMA
THE APPLICATION OF BACKPROPAGATION AND ICZ-ZCZ METHODS ON HANDWRITING PATTERN RECOGNITION OF BIMA SCRIPT
The Bima Script as known as Aksara Bima is one of Bima’s local heritage that needs to be preserved. Based on an online questionnaire of 81 respondent from Bima, there were 66.7% of people who were not familiar with the Bima's Script and 45.7% of people did not even know the existence of the Bima's Script . One of the ways to preserving the Bima script is building a pattern recognition. This research proposes to build a machine learning model that is able to recognize the Handwriting of Bima Script through Zoning feature extraction, Image Centroid Zone (ICZ) and Zone Centroid Zone combined with Backpropagation Neural Network (BPNN) classification. Result of the test using ICZ reached an accuracy up to 87.03% and the result using ZCZ reached and accuracy up to 88.64%, The best performance obtained accuracy up to 89.89% by applying Hidden size = 2, 128 neurons, 0.02 learning rate, error limit 0.001, 1000 epochs, and 9:1 training:testing data.