Segmentasi Wilayah Pada Faktor Yang Memengaruhi Status Gizi Balita Di Pulau Jawa: Pendekatan Path Modeling Segmentation Tree Partial Least Square (PATHMOX-PLS)

Alfahima, Nada Lubna (2026) Segmentasi Wilayah Pada Faktor Yang Memengaruhi Status Gizi Balita Di Pulau Jawa: Pendekatan Path Modeling Segmentation Tree Partial Least Square (PATHMOX-PLS). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221152-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Masalah gizi pada balita, khususnya stunting, wasting, dan underweight, merupakan tantangan kesehatan masyarakat yang serius di Pulau Jawa. Kompleksitas determinan status gizi yang melibatkan faktor ibu, rumah tangga, pemberian ASI, akses layanan kesehatan, dan morbiditas menuntut pendekatan analisis yang mampu menangkap hubungan struktural secara simultan sekaligus mengakomodasi heterogenitas antar wilayah. Penelitian ini bertujuan menganalisis pengaruh kelima faktor tersebut terhadap status gizi balita serta mengidentifikasi segmentasi wilayah berdasarkan perbedaan struktural model menggunakan pendekatan Path Modeling Segmentation Partial Least Square (PATHMOX-PLS). Data yang digunakan merupakan data sekunder agregat tingkat kabupaten/kota dari Profil Kesehatan Provinsi dan publikasi Dalam Angka tahun 2024-2025, mencakup 105 kabupaten/kota di Provinsi Jawa Barat, Jawa Tengah, DI Yogyakarta, dan Jawa Timur. Hasil evaluasi model global PLS-SEM menunjukkan bahwa hanya faktor ibu yang berpengaruh signifikan terhadap status gizi balita dengan koefisien jalur sebesar -0,323, sedangkan faktor rumah tangga, pemberian ASI, akses layanan kesehatan, dan morbiditas tidak terbukti signifikan pada tingkat populasi agregat. Nilai goodness of fit sebesar 0,413 mengindikasikan bahwa model global mampu menjelaskan data secara memadai. Analisis PATHMOX mengidentifikasi heterogenitas struktural yang signifikan berdasarkan dua variabel segmentasi, yaitu status wilayah dan provinsi, sehingga terbentuk tiga segmen terminal: wilayah kota (n=25), kabupaten di Jawa Barat (n=18), dan kabupaten di Jawa Tengah, DIY, dan Jawa Timur (n=62). Model lokal pada setiap segmen menunjukkan peningkatan kemampuan prediksi yang substansial dibandingkan model global, dengan nilai R² berkisar antara 0,329 hingga 0,662. Pada segmen wilayah kota, faktor rumah tangga merupakan determinan dominan (γ = -0,631), sementara di kabupaten Jawa Barat faktor ibu menjadi penentu utama (γ = -0,727) disertai peran signifikan pemberian ASI eksklusif (γ = -0,442). Di kabupaten Jawa Tengah, DIY, dan Jawa Timur, faktor morbiditas terbukti signifikan (γ = -0,313). Temuan ini mengonfirmasi adanya heterogenitas struktural yang bermakna dalam determinan status gizi balita antarwilayah di Pulau Jawa, sehingga intervensi gizi perlu dirancang secara kontekstual sesuai karakteristik masing-masing segmen wilayah.
===================================================================================================================================
Child malnutrition, particularly stunting, wasting, and underweight, remains a critical public health challenge in Java. The complexity of nutritional status determinants encompassing maternal factors, household conditions, breastfeeding practices, healthcare access, and morbidity requires an analytical approach capable of simultaneously capturing structural relationships while accommodating inter-regional heterogeneity. This study aims to analyze the influence of these five factors on toddler nutritional status and to identify regional segmentation based on structural model differences using the Path Modeling Segmentation Partial Least Square (PATHMOX-PLS) approach. Secondary data aggregated at the district/city level were obtained from Provincial Health Profiles and Regional Statistics publications for 2024-2025, covering 105 districts and cities across West Java, Central Java, the Special Region of Yogyakarta, and East Java. Results from the global PLS-SEM model indicate that only the maternal factor had a significant effect on toddler nutritional status, with a path coefficient of -0,323, while household conditions, breastfeeding, healthcare access, and morbidity were not statistically significant at the aggregate population level. A goodness-of-fit value of 0,413 confirmed adequate model fit. PATHMOX analysis identified significant structural heterogeneity based on two segmentation variables, area status and province, yielding three terminal segments: urban areas (n=25), districts in West Java (n=18), and districts in Central Java, Yogyakarta, and East Java (n=62). Local models for each segment demonstrated substantially improved predictive performance compared to the global model, with R² values ranging from 0,329 to 0,662. In urban areas, the household factor was the dominant determinant (γ = -0,631), whereas in West Java districts the maternal factor was the primary driver (γ = -0,727), accompanied by a significant role of exclusive breastfeeding (γ = -0,442). In districts across Central Java, Yogyakarta, and East Java, the morbidity factor was found to be significant (γ = -0,313). These findings confirm meaningful structural heterogeneity in the determinants of toddler nutritional status across regions in Java, underscoring the need for nutritional interventions tailored to the specific characteristics of each regional segment.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Heterogenitas, PATHMOX, PLS-SEM, Pulau Jawa, Status Gizi Balita, Heterogeneity, Java, PATHMOX, PLS-SEM, Toddler Nutritional Status
Subjects:	Q Science > QA Mathematics Q Science > QA Mathematics > QA278.3 Structural equation modeling. Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Nada Lubna Alfahima
Date Deposited:	31 Jul 2026 01:49
Last Modified:	31 Jul 2026 01:49
URI:	http://repository.its.ac.id/id/eprint/140381

Actions (login required)

View Item