PERBANDINGAN VERIFIKASI TANDA TANGAN DENGAN MENGGUNAKAN JARINGAN SARAF TIRUAN BACKPROPAGATION DAN SUPPORT VECTOR MACHINE

• Main Authors: , Barry Caesar Oktariyadi, , Drs. Agus Harjoko, M.Sc., Ph.D
• Format: Theses and Dissertations NonPeerReviewed
• Published: [Yogyakarta] : Universitas Gadjah Mada 2014
• Subjects: ETD
• Online Access: https://repository.ugm.ac.id/130236/; http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=70651
• Institution: Universitas Gadjah Mada

Signature represents biometric feature useful to verify individual�s identity. The study presents the implementation of the signature identification with Support Vector Machine and compared with neural network backpropagation of 100 signature samples. Feature extraction using vertical splitting and horizontal splitting to get the value of the angle and distance as the characteristic value of the signature which would then be in the process. Artificial neural network classification methods Backpropagation earning Neural Network (ANN-BP) and support vector machine (SVM). The Artificial Neural Network with Backpropagation learning method consists of 100 input nodes, 3 hidden layer and 2 output nodes, while its learning function uses resilient backpropagation. Support Vector Machine (SVM) is implemented using one-againt-one method. ANN-BP has an accuracy in the verification of 98.5%, while SVM 94.5% for Traced signatures, ANN with back propagation method is able to verify with the accuracy of 82%, while for the new SVM signatures could not be verified with an accuracy of 43%. The JST with the backpropagation method is faster for the learning process than the SVM.