Nandel Syofneri
Universitas Putra Indonesia YPTK Padang

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

Found 1 Documents
Search

Implementasi Metode Backpropagation untuk Memprediksi Tingkat Kelulusan Uji Kopetensi Siswa Nandel Syofneri; Sarjon Defit; Sumijan
Jurnal Informasi dan Teknologi 2019, Vol. 1, No. 4
Publisher : Rektorat Universitas Putra Indonesia YPTK Padang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37034/jidt.v1i4.13

Abstract

Vocational High School (SMK) 2 Pekanbaru is a Vocational School in Industrial Technology. At present there are 2400 students with 14 majors. In students the level of will in students is still low. Resulting in a low graduation rate for students. This happened because of the difficulty in predicting the level of competency examination passing at SMK Negeri 2 Pekanbaru. The purpose of this study is to assist in predicting the passing level of competency exams so as to produce predictions of student graduation. The method used is the Backpropagation method. With this method data processing can be done using input values and targets that you want to produce. So that it can predict the graduation of students in the expertise competency test. Furthermore, the data to be managed is a recapitulation of the average vocational values majoring in computer network engineering from semester 1 to semester 5 with aspects of knowledge on the target students of 2017 Academic Year and 2018 Academic Year obtained from the sum of all subjects in each semester. The results of calculations using the Backpropagation method with the Matlab application will be predictive in producing grades for students' graduation rates in the future. So that the accuracy value will be obtained in the prediction. With the results of testing the accuracy of prediction student competency tests with patterns 5-4-1 reaching 85%, with patterns 5-6-1 reaching 95%, patterns 5-8-1 reaching 70%, patterns 5-10-1 reaching 85% % and with 5-12-1 patterns it reaches 85%. Of the five patterns, the best accuracy rate of 5-6-1 is 95%. The prediction results using the Bacpropagation method can become knowledge in the next year. So that the system parameters used in testing can be recognized properly.