STUDI KOMPARASI METODE BACKPROPOGATION DENGAN METODE LEARNING VECTOR QUANTIZATION (LVQ) PADA PENGENALAN CITRA TANDA TANGAN

Annisa Solehatul Jannah, NIM.: 15650043 (2020) STUDI KOMPARASI METODE BACKPROPOGATION DENGAN METODE LEARNING VECTOR QUANTIZATION (LVQ) PADA PENGENALAN CITRA TANDA TANGAN. Skripsi thesis, UIN SUNAN KALIJAGA YOGYAKARTA.

Preview	Text (STUDI KOMPARASI METODE BACKPROPOGATION DENGAN METODE LEARNING VECTOR QUANTIZATION (LVQ) PADA PENGENALAN CITRA TANDA TANGAN) 15650043_BAB-I_IV-atau-V_DAFTAR-PUSTAKA.pdf - Published Version Download (7MB) | Preview
	Text (STUDI KOMPARASI METODE BACKPROPOGATION DENGAN METODE LEARNING VECTOR QUANTIZATION (LVQ) PADA PENGENALAN CITRA TANDA TANGAN) 15650043_BAB-II_sampai_SEBELUM-BAB-TERAKHIR.pdf - Published Version Restricted to Registered users only Download (3MB) | Request a copy

Abstract

Signatures are more often used for ratification compared to others. The sale transaction, deposit in a bank, such as a withdrawal ora storage funds, the conclusion of land purchase and other management processes all required the validity evidenced by administrative documents. The purpose of this study was to made an biometric signature system with backpropagation and LVQ-type neural networks, and to measure the accuracy of backpropagation and LVQ. Based on that, the reseachers would develop an application to identify signature by using backpropagation and LVQ algorithm and the results of signatures description identifications with backpropagation method compared to LVQ method. The backpropagation and LVQ materials that will be used was the result from the pra-process and extraction of the descriptions. Pra-processing of the descriptions will be done by applying resizing, noise reduction, and edge detection. The results of system test the accuracy level of system in recognizing signature images was 20.0% for backpropogation and 90.0% for LVQ. The best network architecture that used at signatures description identifications by backpropagation method was with the learning rate variations at 0.8 and the tolerance number at 0.6. On the other hand the learning vector quantization method results was with the learning rate variations at 0.5, and the learning rate change at 0.5

Share this knowledge with your friends :

Actions (login required)

View Item