Pengenalan Pola Tulisan Tangan Aksara Sasak Menggunakan Metode Linear Discriminant Analysis dan Jaringan Syaraf Tiruan Jenis Backpropagation

Abstract

Sasak’s script is the one of Indonesia's cultural heritages which is endangered language category and it needs to be preserved so its not become extinct script. If there is no preservation, Sasak’s script will be close to extinction. Given the importance of conservation efforts, a reserch was carried out to recognize the Sasak’s script handwriting patterns. This reserch aims to see the optimal model for recognizing Sasak script patterns and to see how accurate the Linear Discriminant Analysis method as extraction features and Backpropagation Neural Network as a classification method for recognizing Sasak’s script patterns. The dataset in this reserch consisted of three variations amount of data which are 13500, 10800 and 2700. According to the final results of the study the best accuracy was obtained in the 10800 dataset, 92.20% with 92% precision and 92% recall with 250,691s computing time. The test is based on the best paramater of DCT coefficient test parameters, eigen value, number of neurons, number of hidden layers and learning rate that has been tested before.