EMAIL: PASSWORD:
Front Office
UPT. PERPUSTAKAAN
Institut Teknologi Sepuluh Nopember Surabaya


Kampus ITS Sukolilo - Surabaya 60111

Phone : 031-5921733 , 5923623
Fax : 031-5937774
E-mail : libits@its.ac.id
Website : http://library.its.ac.id

Support (Customer Service) :
timit_perpus@its.ac.id




Welcome..guys!

Have a problem with your access?
Please, contact our technical support below:
LIVE SUPPORT


Moh. Fandika Aqsa


Davi Wahyuni


Tondo Indra Nyata


Anis Wulandari


Ansi Aflacha




ITS » PhD Theses » S3 - Statistika
Posted by dee@its.ac.id at 08/06/2015 21:14:38  •  1327 Views


ROBUST MODEL-BASED CLUSTERING DENGAN DISTRIBUSI T MULTIVARIAT DAN MINIMUM MESSAGE LENGTH APLIKASI PADA PENGUKURAN KERAWANAN SOSIAL

ROBUST MODEL-BASED CLUSTERING WITH MULTIVARIATE t DISTRIBUTION AND MINIMUM MESSAGE LENGTH Applied in Social Vulnerability Assessment

Author :
SIAGIAN, TIODORA HADUMAON ( 1309301701 )




ABSTRAK

Pengukuran kerawanan merupakan solusi yang efektif untuk mengurangi resiko dan kerugian dampak bencana alam. Banyak pengukuran kerawanan lebih menekankan pada aspek kerawanan biofisik dan fisik saja dan mengabaikan kerawanan bidang sosial. Kurangnya perhatian pada kerawanan sosial umumnya adalah karena kesulitan dalam pengukuran. Social Vulnerability Index SoVI adalah metode yang populer digunakan untuk mengukur kerawanan sosial. SoVI memiliki beberapa kelemahan antara lain dalam akurasi dan pengujian validitasnya mengandung subjektivitas dalam proses transformasi agregasi dan pembobotan dan tidak dapat menangani masalah outlier. Metode model-based clustering dapat digunakan untuk mengukur kerawanan sosial. Metode modelbased clustering lebih disukai belakangan ini karena menggunakan prinsip-prinsip statistik dapat menentukan jumlah kelompok dalam data dan dapat mengatasi masalah outlier. Metode model-based clustering didasarkan pada model finite mixture yang mengasumsikan data dihasilkan dari beberapa distribusi probabilitas yang masing-masing mewakili kelompok yang berbeda. Disertasi ini mengembangkan metode RMBC-MML yaitu metode model-based clustering dengan metode Maximum Penalized Likelihood untuk estimasi parameter dan ukuran Minimum Message Length untuk pemilihan model pada model finite mixture t multivariat. Metode RMBC-MML terbukti secara deskriptif menghasilkan ketepatan estimasi parameter yang sangat baik dan dapat mengatasi masalah singularitas pada estimasi matriks kovarians. Aplikasi algoritma RMBCMML mengidentifikasi ada 3 kelompok dalam data kerawanan sosial tahun 2010. Sebagian besar kabupatenkota di Indonesia berada pada tingkat kerawanan sosial menengah dan sekitar 402 persen berada pada tingkat kerawanan sosial tinggi. Sebuah peta tipologi wilayah dibuat berdasarkan kelompok-kelompok terbentuk. Saat sumber daya terbatas peta tipologi wilayah ini dapat digunakan untuk memprioritaskan kabupatenkota dengan tingkat kerawanan sosial yang relatif tinggi


ABSTRACT

Vulnerability assessments are effective solutions to reduce risk and losses of the impact of natural hazards. Many vulnerability assessments put more emphasis on the biophysical and physical vulnerability aspects leaving the social dimension of vulnerability poorly addressed. Social vulnerability was largely disregarded mainly because it is difficult to quantify. Social Vulnerability Index SoVI is a popular method to assess social vulnerability. SoVI has some weaknesses such as in the accuracy and validity testing containing subjectivity in the process of transformation aggregation and weighting and cannot handle outlier problem. Model-based clustering method can be used to assess social vulnerability. This method is much preferred recently because it uses statistical principles can determine number of clusters in the data and can solve outlier problem. Modelbased clustering method which based on finite mixture models has an assumption that data are generated from several probability distributions each representing a different cluster. This disertation develops RMBC-MML method a model-based clustering method which using Maximum Penalized Likelihood method for parameter estimation and Minimum Message Length criterion for model selection on mixtures of multivariate t distributions. The RMBC-MML method has proven to yield accurate parameter estimation descriptively and can overcome singularity problem on covariance matrix estimation. Application of the RMBCMML algorithm identifies there are 3 clusters in the 2010 social vulnerability data. The majority of districts in Indonesia are in moderate level of social vulnerability and about 4.02 percent is in high level of social vulnerability. A typology map based on the formed clusters is created. When resources are limited this typology map can be used to prioritize those districts with relative high of social vulnerability level.



KeywordsRobust Model-Based Clustering; Model Finite Mixture; Distribusi t Multivariat; Minimum Message Length; Kerawanan Sosial; Bencana Alam; Indonesia
 
Subject:  Studi wilayah;Analisis diskriminan
Contributor
  1. DR. Purhadi, M. Sc
  2. DR. Suhartono, M. Sc
Date Create: 13/02/2014
Type: Text
Format: pdf
Language: Indonesian
Identifier: ITS-PhD-13104150000129
Collection ID: 13104150000129
Call Number: RDSt 519.53 Sia r


Source
PhD Thesis of Statistics, RDSt 519.53 Sia r, 2015

Coverage
ITS Community

Rights
Copyright @2015 by ITS Library. This publication is protected by copyright and per obtained from the ITS Library prior to any prohibited reproduction, storage in a re transmission in any form or by any means, electronic, mechanical, photocopying, reco For information regarding permission(s), write to ITS Library




[ Download - Open Access ]

  1.  ITS-PhD-37800-1309301701-abstract_id.pdf - 1791 KB
  2.  ITS-PhD-37800-1309301701-abstract_en.pdf - 1791 KB
  3.  ITS-PhD-37800-1309301701-conclusion.pdf - 707 KB




 Similar Document...




! ATTENTION !

To facilitate the activation process, please fill out the member application form correctly and completely

Registration activation of our members will process up to max 24 hours (confirm by email). Please wait patiently

POLLING

Bagaimana pendapat Anda tentang layanan repository kami ?

Bagus Sekali
Baik
Biasa
Jelek
Mengecewakan





You are connected from 54.196.31.117
using CCBot/2.0 (http://commoncrawl.org/faq/)



Copyright © ITS Library 2006 - 2018 - All rights reserved.
Dublin Core Metadata Initiative and OpenArchives Compatible
Developed by Hassan