Rhomah, Azizatu (2012) Studi copula gumbel family 2-dimensi dalam identifikasi struktur dependensi. Undergraduate thesis, Universitas Islam Negeri Maulana Malik Ibrahim.
|
Text (Fulltext)
08610068.pdf - Accepted Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (6MB) | Preview |
Abstract
INDONESIA:
Hubungan antara beberapa variabel atau disebut dependensi memiliki cara pengukuran yang sama baik untuk linier maupun non linier. Dependensi linier akan lebih mudah dalam pengukuran dependensinya, sedangkan untuk non linier akan sulit karena korelasi memiliki kelemahan. Oleh karena itu dependensi non linier dijelaskan dalam copula yang digunakan untuk memodelkan dependensi antar variabel acak. Jika dalam korelasi pearson distribusinya diasumsikan normal dan diketahui, maka copula dapat mengidentifikasi jika distribusi diasumsikan tidak normal dan bahkan tidak diketahui. Dalam hal ini copula yang akan diidentifikasi adalah family copula Archimedian, yaitu copula Gumbel yang dapat mengidentifikasi dependensi positif.
Copula Gumbel tersebut dianalisis menggunakan teori-teori copula untuk diidentifikasi sifat-sifat, transformasi ke dalam distribusi marginal uniform [0,1], identifikasi distribusi tersebut berdasarkan sifat dasar copula dan dependensi antara dua variabel. Pada penelitian ini diperoleh nilai dependensi Kendall τ, yaitu τç = θ-1/θ. Melalui simulasi dengan membangkitkan data acak 2 variabel yang berdistribusi Gamma (4,3) dan lognormal (3,2) diperoleh τ= 0.5440 untuk θ= 1.5396
ENGLISH:
Relationship among severally variable or so called dependence has to make the point good same measurement for linear and also non linear. Linear Dependence will a lot easier in measurement dependence, meanwhile for non linear will be hard since correlation have weakness. Therefore non linear dependence is worded deep copula which is utilized to model dependence among random variable. If in Pearson's correlation its distribution assumed by normal and is known, therefore copula can identify if distribution is assumed not normal and even unknown. In this case copula what do will be identified is family copula Archimedian, which is copula Gumbel one that gets to identify dependence positive.
Copula Gumbel that analyzed utilizes copula's theories to been identified characters, transformation into uniform's marginal distribution [0,1] identify that distribution bases copula's basic character and dependence among two variables. Via simulation arouse random data two variables that gets Gamma distribution (4,3) and lognormal (3,2) acquired τ= 0.5440 for θ= 1.5396.
Item Type: | Thesis (Undergraduate) | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
Supervisor: | Rozi, Fachrur and Abidin, Munirul | |||||||||
Contributors: |
|
|||||||||
Keywords: | Copula Gumbel Family; Struktur Dependensi; Copula Gumbel Family; Dependence structure | |||||||||
Subjects: | 01 MATHEMATICAL SCIENCES > 0101 Pure Mathematics > 010111 Real and Complex Functions (incl. Several Variables) 01 MATHEMATICAL SCIENCES > 0101 Pure Mathematics > 010199 Pure Mathematics not elsewhere classified |
|||||||||
Departement: | Fakultas Sains dan Teknologi > Jurusan Matematika | |||||||||
Depositing User: | Ahmad Zaini | |||||||||
Date Deposited: | 29 May 2017 10:24 | |||||||||
Last Modified: | 29 May 2017 10:24 | |||||||||
URI: | http://etheses.uin-malang.ac.id/id/eprint/6748 |
Downloads
Downloads per month over past year
Actions (login required)
View Item |