SIGNATURE ANALYSIS FOR PERSONAL CHARACTERISTICS PREDICTION USING CONVOLUTIONAL NEURAL NETWORK (CNN) METHOD

Antivirus : Jurnal Ilmiah Teknik Informatika

View Publication Info
 
 
Field Value
 
Title SIGNATURE ANALYSIS FOR PERSONAL CHARACTERISTICS PREDICTION USING CONVOLUTIONAL NEURAL NETWORK (CNN) METHOD
PREDIKSI KARAKTERISTIK PERSONAL MENGGUNAKAN ANALISIS TANDA TANGAN DENGAN MENGGGUNAKAN METODE CONVOLUTIONAL NEURAL NETWORK (CNN)
 
Creator Mawaddah, Udkhiati
Armanto, Hendrawan
Setyati, Endang
 
Subject classification
graphology
hand signature
CNN
grafology
tanda tangan
klasifikasi
CNN
 
Description Graphology is the study of handwriting that can describe the characteristics of a writer and his emotional disposition. Knowing the characteristics of prospective applicants is very important for the Human Resource Development (HRD) that responsible for selecting employees in their fields. HRD often experienced the Mistaken when in the process of hiring employees who identify the candidate employee signature to lose both time and costs in that company. This research using 7 signature features which are divided into two algorithms respectively, 5 signature features consisting are Curved Start, End Streak, Shell, Middle Streaks, Underline and Identification Structure Algorithm consist 2 signature features are Dot Structure and Streaks disconnected. The evaluation results obtained a training data accuracy value of 0.7333, training data loss of 0.7693, test data accuracy of 0.7778, and test data loss of 0.8377 which can be concluded that the results of the two data is underfitting. Thus, we must concern to collecting other dataset which has features similarity in every classes.
Grafologi ilmu yang mempelajari tentang tulisan tangan yang dapat mengetahui gambaran karakteristik seorang penulis dan disposisi emosional. Dengan mengetahui karakteristik calon pelamar dapat membantu pekerjaan Human Resource Development (HRD) yang bertanggung jawab dengan pemilihan calon karyawan yang sesuai dengan bidangnya. Kesalahan yang sering dialami oleh HRD ketika dalam proses perekrutan karyawan yang mengidentifikasi tanda tangan calon karyawan yang dapat mengalami kerugian baik waktu dan biaya di perusahaan tersebut. Penelitian ini menggunakan 7 fitur tanda tangan yang terbagi menjadi dua algoritma diantaranya 5 fitur tanda tangan yang terdiri dari awal kurva, coretan akhir, cangkang, coretan di tengah dan garis bawah tanda tangan yang diproses menggunakan metode Convolutional neural network (CNN) dan klasifikasi 2 fitur tanda tangan yang terdiri dari struktur titik dan tanda tangan terpisah menggunakan metode Algoritma Identifikasi Struktur. Hasil evaluasi didapatkan nilai akurasi data training sebesar 0.7333, loss data training sebesar 0.7693, akurasi data test sebesar 0.7778 dan loss data test sebesar 0.8377 yang dapat disimpulkan bahwa hasil penelitian ini masih tergolong underfitting. Hal ini dikarenakan masih butuh banyak dataset yang lebih banyak jumlah dan variannya yang mempunyai ciri-ciri yang ada kemiripan masing-masing kelasnya.
 
Publisher Universitas Islam Balitar
 
Date 2021-06-10
 
Type info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
Peer-reviewed Article
 
Format application/pdf
 
Identifier https://ejournal.unisbablitar.ac.id/index.php/antivirus/article/view/1526
10.35457/antivirus.v15i1.1526
 
Source Antivirus : Jurnal Ilmiah Teknik Informatika; Vol. 15 No. 1 (2021): Mei 2021; 123-133
Antivirus : Jurnal Ilmiah Teknik Informatika; Vol 15 No 1 (2021): Mei 2021; 123-133
2527-337X
1978-5232
 
Language eng
 
Relation https://ejournal.unisbablitar.ac.id/index.php/antivirus/article/view/1526/1040
 
Rights Copyright (c) 2021 Antivirus : Jurnal Ilmiah Teknik Informatika
https://creativecommons.org/licenses/by-sa/4.0
 

Contact Us

The PKP Index is an initiative of the Public Knowledge Project.

For PKP Publishing Services please use the PKP|PS contact form.

For support with PKP software we encourage users to consult our wiki for documentation and search our support forums.

For any other correspondence feel free to contact us using the PKP contact form.

Find Us

Twitter

Copyright © 2015-2018 Simon Fraser University Library