skip to main content

Penggunaan Jaringan Syaraf Tiruan Backpropagation Untuk Seleksi Penerimaan Mahasiswa Baru Pada Jurusan Teknik Komputer Di Politeknik Negeri Sriwijaya

*Maria Agustin  -  Program Studi Sistem Informasi,
Toni Prahasto  -  Jurusan Teknik Mesin, Fakultas Teknik

Citation Format:

Data availability of new studentsat the Polytechnic State Srivijaya high enough, so the need fora method to analyze the data. Artificial neuralnetwork  is  an  information  processing  system  that  has  characteristics  similar  to  biological  neural  networks,  neural  network  sare  used  topredict because ofthe ability of a good approach to ketidak linearan. This study  will design the software selection admission of new studentswith a  backpropagation  neural network  methods.    From the analysis of  backpropagation  neural networks  with  one  hidden layer  with  the number of neurons 50, 1000 iteration sand the activation functiont an sigproduce regression of 0.4822. Backpropagation neural network withtwo  hidden  layers  with  the  number  of  neurons  50,  4000  iterations  with  tansig  activation  function,  resulting  in  regression  of  0.7944. Backpropagation  neural networks  with 3  hidden layer  with  the number of neurons  35,  5000  iterations, resulting in  regression  of  0.8563. Based  on  the  results  of  this  analysis,  backpropagation  neural  networks  quite  effectively  used  for  selection  of  candidates  for  student admission.

Keywords: Selection, Backpropagation, Regression

Note: This article has supplementary file(s).

Fulltext |  Research Instrument
Type Research Instrument
  Download (825KB)    Indexing metadata

Article Metrics:

Article Info
Section: Research Articles
Language : ID

Last update:

No citation recorded.

Last update: 2021-09-20 23:59:06

  1. Early detection application of bipolar disorders using backpropagation algorithm

    Fitriati D.. International Conference on Electrical Engineering, Computer Science and Informatics (EECSI), 2019. doi: 10.23919/EECSI48112.2019.8977102