Partial Least Square Dengan Skema PCA Dan Importance-Performance Analysis Dalam Kasus Gizi Buruk Balita Di Jawa Timur

Alfasanah, Zulfani (2026) Partial Least Square Dengan Skema PCA Dan Importance-Performance Analysis Dalam Kasus Gizi Buruk Balita Di Jawa Timur. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003232022-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Structural Equation Modeling (SEM) merupakan metode yang efektif untuk mengestimasi hubungan kausal antar variabel laten, namun pendekatan Covariance-Based SEM (CB-SEM) sering terkendala oleh asumsi distribusi dan ukuran sampel. Oleh karena itu, penelitian ini mengembangkan Partial Least Squares (PLS-SEM) dengan skema pembobotan Principal Component Analysis guna meningkatkan stabilitas dan efisiensi estimasi, serta mengintegrasikannya dengan Importance-Performance Analysis (IPA) untuk penentuan prioritas kebijakan. Skema PCA memungkinkan penentuan bobot indikator melalui komponen utama pertama tanpa proses iteratif, sedangkan koefisien jalur diestimasi menggunakan regresi OLS. Aplikasi pada kasus gizi buruk balita di Jawa Timur menunjukkan bahwa konstruk sosial ekonomi dibentuk secara seimbang oleh pengeluaran per kapita dan rata-rata lama sekolah, pelayanan kesehatan dan lingkungan didominasi oleh cakupan imunisasi dasar lengkap, serta gizi buruk paling kuat direpresentasikan oleh prevalensi underweight. Hasil model struktural menunjukkan bahwa sosial ekonomi berpengaruh signifikan terhadap pola asuh, ketahanan pangan, serta pelayanan kesehatan dan lingkungan, sementara pengaruh ketiga konstruk tersebut terhadap gizi buruk tidak signifikan secara statistik. Analisis IPA mengidentifikasi sosial ekonomi sebagai faktor prioritas utama karena memiliki kepentingan tinggi tetapi kinerja rendah. Temuan ini memberikan kontribusi metodologis dalam pengembangan PLS berbasis PCA serta rekomendasi kebijakan yang lebih terarah dalam penanganan gizi buruk balita.
======================================================================================================================================
Structural Equation Modeling (SEM) is an effective method for estimating causal relationships among latent variables; however, the Covariance-Based SEM (CB-SEM) approach is often constrained by distributional assumptions and sample size requirements. Therefore, this study develops a Partial Least Squares (PLS-SEM) framework using a Principal Component Analysis weighting scheme to enhance estimation stability and efficiency, and integrates it with Importance–Performance Analysis (IPA) for policy prioritization. The PCA scheme determines indicator weights through the first principal component without requiring iterative procedures, while path coefficients are estimated using Ordinary Least Squares. Application to the case of undernutrition among children under five in East Java reveals that the socio-economic construct is equally represented by per capita expenditure and average years of schooling, health and environmental services are dominated by complete basic immunization coverage, and malnutrition status is most strongly reflected by underweight prevalence. Structural model results show that socio-economic status has significant positive effects on parenting practices, food security, and health and environmental services, whereas the effects of these constructs on malnutrition are not statistically significant. IPA results identify socio-economic as the highest priority factor due to high importance but low performance. These findings contribute methodologically to the development of PCA-based PLS and provide more targeted policy recommendations for addressing child malnutrition.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Importance-Performance Analysis, Partial Least Square, Principal Component Analysis, Structural Equation Modeling, Importance-Performance Analysis, Partial Least Square, Principal Component Analysis, Structural Equation Modeling
Subjects:	Q Science > QA Mathematics > QA278.3 Structural equation modeling.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Zulfani Alfasanah
Date Deposited:	29 Jan 2026 04:17
Last Modified:	29 Jan 2026 04:17
URI:	http://repository.its.ac.id/id/eprint/130949

Actions (login required)

View Item