PEMETAAN PREFERENSI MAHASISWA BARU DALAM MEMILIH JURUSAN MENGGUNAKAN ARTIFICIAL NEURAL NETWORK (ANN) DENGAN ALGORITMA SELF ORGANIZING MAPS (SOM)

Hilmi, Muh Najib (2015) PEMETAAN PREFERENSI MAHASISWA BARU DALAM MEMILIH JURUSAN MENGGUNAKAN ARTIFICIAL NEURAL NETWORK (ANN) DENGAN ALGORITMA SELF ORGANIZING MAPS (SOM). Undergraduate thesis, FSM Universitas Diponegoro.

[img]
Preview
PDF
6Mb

Abstract

College is the highest educational institution and the role the intellectual life of the Indonesian people that the main purpose of academics. Not all colleges into their destination but only college that has a role, credibility and rank the best course of which it is their goal. This makes higher education marketing research approach to get attention and become the main goal of the academics in choosing a college. This research was conducted in order to determine with certainty attribute / emotional reasons academics in choosing college as their academic goals. The method used in this study were self-organizing maps with the Kohonen algorithm is a classification method. Kohonen SOM algorithm with learning rate used 0:05, 0.25, 0:50, 0.75, 0.95 and initialization of initial weight value and the value of the midpoint and 500 iterations with output 3 clusters are formed. Results clustering of SOM validated using Davies-Bouldin index with the best clustering results that DBI minimum (1.7802) with the learning rate is 0.95 and the cluster formed three clusters for the first cluster as many as six members, cluster-2 by 9 members and 3rd cluster as 5 members the results of clustering with top priority contained in the cluster to-2 with a mean (7.434) with the characteristics of each member is an emotional reason in choosing a major. Keywords: Self Organizing Maps, Kohonen algorithm, Learning Rate, Index Davies Bouldin, Cluster.

Item Type:Thesis (Undergraduate)
Subjects:H Social Sciences > HA Statistics
Divisions:Faculty of Science and Mathematics > Department of Statistics
ID Code:46558
Deposited By:INVALID USER
Deposited On:06 Oct 2015 10:04
Last Modified:06 Oct 2015 10:04

Repository Staff Only: item control page