Fitriatien, Sri Rahmawati (2015) Aplikasi Jaringan Syaraf Tiruan Kohonen Self Organizing MAPS Dan Learning Vector Quantization Pada Data Kualitas Air Kali Surabaya. Masters thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
1213201031-Master_Theses.pdf - Published Version Download (7MB) | Preview |
Preview |
Text
1213201031-Presentation.pdf - Published Version Download (2MB) | Preview |
Preview |
Text
1213201031-Paper-1213201031-paperpdf.pdf - Published Version Download (1MB) | Preview |
Preview |
Text
1213201031-Paper-1213201031-publicationpdf.pdf - Published Version Download (1MB) | Preview |
Abstract
Kali Surabaya adalah sumber air baku yang digunakan masyarakat
Surabaya untuk memenuhi kebutuhan sehari-hari. Kondisi air permukaan Kali
Surabaya mengalami penurunan kualitas air yang dirasakan semakin hari semakin
meningkat akibat sebagian besar limbah cair hasil dari kegiatan manusia dibuang
ke saluran yang bermuara di Kali Surabaya. Limbah tersebut berasal dari
permukiman, industri, pertanian, peternakan dan lain-lain. Indikator kimia
pencemaran air limbah cair yang digunakan yaitu BOD, COD dan DO. Tujuan dari
penelitian tesis ini adalah melakukan pengelompokan seluruh titik pantau kualitas
air Kali Surabaya dengan jumlah cluster yang terbentuk dimulai dari 2 hingga 4
cluster berdasarkan kategori pembagian status mutu air. Dari pengelompokan titik
pantau ini kemudian dilakukan penetapan status mutu kualitas air selama 14 periode
dimulai dari bulan Januari 2010 hingga Juni 2013.
Jaringan syaraf tiruan merupakan sistem pemroses informasi seperti
pemroses pada otak manusia. Jaringan syaraf tiruan telah banyak digunakan dalam
banyak aplikasi, salah satunya adalah clustering. Dalam tesis ini, metode Kohonen
Self Organizing Maps dan Learning Vector Quantization digunakan untuk
menyelesaikan masalah clustering titik pantau kualitas air Kali Surabaya pada
setiap waktu pantau. Untuk penentuan cluster terbaik menggunakan Davies-
Bouldin Index (DBI) sebagai validasi cluster. Penentuan status mutu air Kali
Surabaya di setiap titik pantau dilakukan dengan metode Indeks Pencemaran yang
divalidasi dengan uji distribusi normal.
Dari implementasi dan uji coba program dapat diperoleh simpulan bahwa
algoritma Kohonen-SOM dan LVQ dapat mengenali pola dan mampu mencocokan
anggota kelompok titik pantau dengan parameter learning rate minimal 0,000001
diperoleh nilai mean square error pada jaringan LVQ lebih kecil dibandingkan
dengan jaringan Kohonen-SOM. Berdasarkan Indeks Pencemaran, status mutu air
Kali Surabaya pada Januari 2010-Juni 2013 berada pada status mutu air Tercemar
Ringan. ========== Surabaya’s river is primary water source needed for Surabaya’s people for
their daily demand. Meanwhile, the quality for water of Surabaya’s river is more
decreased from day to day because the most liquid waste of human activities are
thrown into canal to Surabaya’s river. The waste are from the settlement, industry,
agriculture, animal husbandry, etc. Chemical indicator of liquid waste pollution that
is BOD, COD and DO. The purposes of this research are to cluster all observation
points of the quality of water in Surabaya’s river with the cluster number started
from 2 clusters to 4 clusters. Based on this cluster result of observation points, it is
then fixed the quality of water for 14 period started from January 2010 to June 2013.
Artificial neural networks are information processing systems such as
processing in the human brain. Artificial neural networks have been widely used in
many applications, one of which is clustering. In this thesis, the method of Kohonen
Self Organizing Maps and Learning Vector Quantization clustering is used to solve
the problems of water quality monitoring points Surabaya at any time to monitor.
To determine the best cluster using the Davies-Bouldin Index (DBI) as the cluster
validation. Determination of the status of water quality at any point Surabaya
monitoring was conducted by the Pollution Index test is validated by the normal
distribution.
From implementation and test programs can be concluded that the
algorithm-SOM Kohonen and LVQ can recognize patterns and able to match the
group members monitoring points with a minimal learning rate parameter value
0.000001 mean square error values obtained in LVQ network is smaller than the
Kohonen -SOM. Based on the Pollution Index, the water quality status Kali
Surabaya in January 2010-June 2013 are in the floaty polluted water quality status.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Kohonen self organizing maps; learning vector quantization; status mutu air; indeks pencemaran; Kohonen self organizing maps; learning vector quantization; quality of water level; pollution index |
| Subjects: | Q Science > QA Mathematics > QA278.55 Cluster analysis Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA9.58 Algorithms |
| Divisions: | Faculty of Mathematics and Science > Mathematics > 44101-(S2) Master Thesis |
| Depositing User: | - Davi Wah |
| Date Deposited: | 24 May 2018 07:06 |
| Last Modified: | 06 Mar 2019 06:23 |
| URI: | http://repository.its.ac.id/id/eprint/51944 |
Actions (login required)
 |
View Item |