UDiNus Repository

Optimasi Jaringan Syaraf Tiruan Backpropagation Menggunakan Particle Swarm Optimization untuk Deteksi Penderita Penyakit Jantung

OKTAWANDARI, HARGIANTI HENNI (2014) Optimasi Jaringan Syaraf Tiruan Backpropagation Menggunakan Particle Swarm Optimization untuk Deteksi Penderita Penyakit Jantung. Skripsi, Fakultas Ilmu Komputer.

[img]
Preview
PDF - Published Version
Download (3970b) | Preview

    Abstract

    According to data from the Student Association of Epidemiology School of Public Health, Hasanuddin University (HIMAPID), in Indonesia, with a population of around 237 million people, there were 6.6 million babies born each year and 48,800 of them are already bearing heart disease. To classify whether a person is indicated for heart disease, this current research applied Artificial Neural Network model of Backpropagation and Backpropagation Neural Network models that its weight are optimized using Particle Swarm Optimization with the aim of minimizing the mean squared error (MSE). Both of those methods will be sought through the highest accuracy using Confusion Matrix and the impairment of RMSE for then applied to the Graphical User Interface (GUI). The data that used are from UCI Machine Learning respository with the observation of 270 participants consisting of 13 variables were determined to do the detection. The test results show the accuracy of BPNN + PSO is higher than BPNN, ie, 87.7% and RMSE values decreased from 1.0862 becomes 0.3403.

    Item Type: Article
    Subjects: T Technology > Teknik Informatika > INF Informatika
    Universitas Dian Nuswantoro > Fakultas Ilmu Komputer > Teknik Informatika > INF Informatika
    Semantik 2013 > INF Informatika
    Divisions: Library of Congress Subject Areas > T Technology > Teknik Informatika > INF Informatika
    Fakultas Ilmu Komputer > Teknik Informatika > INF Informatika
    Semantik 2013 > INF Informatika
    Depositing User: Psi Udinus
    Date Deposited: 12 Sep 2014 09:37
    Last Modified: 22 Nov 2014 06:05
    URI: http://eprints.dinus.ac.id/id/eprint/5428

    Actions (login required)

    View Item