Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Prosiding Seminar Nasional Ilmu Sosial dan Teknologi (SNISTEK)
Resky Novaliandy
Universitas Sanata Dharma
Computer Science & IT Economics, Econometrics & Finance Social Sciences Other

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Klasifikasi Aksara Jawa Cetak Menggunakan Jaringan Syaraf Tiruan Backpropagation Resky Novaliandy; Anastasia Rita Widiarti
Prosiding Vol 4 (2022): SNISTEK
Publisher : LPPM Universitas Putera Batam

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Script is a visual symbol that is usually printed on paper, stone, wood, cloth and others. One example of a script is the Javanese Nglegena script. Nglegena Javanese script can be found in various forms, one of which is in the form of images. Image is a digital representation of an object. By using the image of the character, it is done as an effort to fight against the character. Identification can be done digitally by using one of the classifications, namely artificial-backpropagation network. Backpropagation neural network is a good classification that can update its weight value. This research uses image data of Javanese Nglegena script with a total data of 635 images. Each image is preprocessed in the form of image resizing and thinning. Then the feature extraction used is in the form of 7 moment invariant values, the value of image length and image width so that the number of inputs for each data is 9 features. Then the 9 feature values ​​become input for classifying the backpropagation-artificial network. The best results for testing 16 types of Javanese Nglegena script get a training accuracy of 92.1% and a test accuracy of 76.1%. Where the best architecture uses a hidden 2-layer network with 9 input neurons, 55 neurons in layer 1, 65 neurons in layer 2 and 4 layers, logsig activation function and the trainlm training method.
Co-Authors Anastasia Rita Widiarti