Maghfiroh, Nisfu Lailatul (2018) Pendekatan partial Least Square Regression pada pemodelan persamaan struktural. Undergraduate thesis, Universitas Islam Negeri Maulana Malik Ibrahim.
|
Text (Fulltext)
14610063.pdf - Accepted Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (2MB) | Preview |
Abstract
INDONESIA:
Pemodelan persamaan stuktural (SEM) adalah pemodelan yang mengukur hubungan antara variabel laten dan indikator secara simultan. SEM berbasis varian, Partial Least Square (PLS), model yang bebas distribusi (free distribution) dan fleksibel. Peneliti ingin menganalisis hubungan antara variabel laten eksogen terhadap variabel laten endogen, pendekatan yang dilakukan menggunakan Partial Least Square Regression (PLSR). Variabel laten diestimasi menggunakan metode Singular Value Decomposition (SVD). SVD merupakan metode dekomposisi suatu matriks ke dalam beberapa komponen yang berkaitan erat dengan nilai-nilai singularnya. Data yang digunakan dalam contoh kasus pada penelitian ini berupa data sekunder hasil survei pada Aziz (2017) yaitu populasi dosen tetap (PNS dan Non PNS) Fakultas Sains dan Teknologi semester ganjil 2016/2017 UIN Maulana Malik Ibrahim Malang. Berdasarkan penelitian yang telah dilakukan pada pendekatan PLSR terhadap SEM dapat disimpulkan bahwa hasil estimasi variabel laten menggunakan SVD menghasilkan model struktural, η=ξΓ+ζ dan model pengukuran, X=ξΛ_X^T+δ dan Y=ηΛ_Y^T+ε. Hasil analisis data menunjukkan bahwa Beban Kinerja Dosen (BKD) dengan semua aspek penyusunnya (Pengajaran dan Penunjang lainnya) mampu mempengaruhi secara positif dan signifikan terhadap Indeks Kepuasan Mahasiswa (IKM) melalui kompetensi pedagogik, profesional, kepribadian, dan sosial dosen dengan menggunakan metode SVD dengan indikator reflektif. Pada penelitian selanjutnya, diharapkan dapat dilanjutkan untuk dikembangkan atau dibandingkan dengan partial least square path-modeling atau metode lainnya.
ENGLISH:
Structural Equation Modeling (SEM) is a model that measures the relation between latent variables and indicator simultaneously. Variant-based SEM, Partial Least Square (PLS), (free distribution) and flexible models. The researcher wants to analyze the relation between exogenous latent variables and endogenous latent variables, the approach taken using Partial Least Square Regression (PLSR). The latent variable is estimated using the Singular Value Decomposition (SVD) method. SVD is a decomposition method of a matrix into several components that are closely related to its singular values. The data used in this case example are secondary data of survey result on Aziz (2017) that is population of permanent lecturer (PNS and Non PNS) Faculty of Science and Technology semester odd 2016/2017 UIN Maulana Malik Ibrahim Malang. Based on the research that has been done on the PLSR approach to SEM, it can be concluded that the results of estimating latent variables using SVD produce a structural model, η=ξΓ+ζ and the measurement model, X=ξΛ_X^T+δ and Y=ηΛ_Y^T+ε. The results of the data analysis show that Lecturer Performance Load (BKD) with all its constituent aspects (Teaching and other Support) is able to positively and significanly influence the Student Satisfaction Index (IKM) through pedagogic, professional, personality, and social lecturer competencies by using SVD method with reflective indicator. For the next research, the writer hopes can continue to be developed or compared with partial least square path-modeling or other methods.
Item Type: | Thesis (Undergraduate) | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
Supervisor: | Aziz, Abdul and Khudzaifah, Muhammad | |||||||||
Contributors: |
|
|||||||||
Keywords: | SEM; PLSR; SVD; IKM; BKD | |||||||||
Departement: | Fakultas Sains dan Teknologi > Jurusan Matematika | |||||||||
Depositing User: | Heni Kurnia Ningsih | |||||||||
Date Deposited: | 03 May 2019 09:31 | |||||||||
Last Modified: | 03 May 2019 09:31 | |||||||||
URI: | http://etheses.uin-malang.ac.id/id/eprint/14113 |
Downloads
Downloads per month over past year
Actions (login required)
View Item |