Pengembangan Model Hybrid Fuzzy Goal Programming Hidden Markov Dengan Feature Selection

Syalsabilla, Alya Fitri (2025) Pengembangan Model Hybrid Fuzzy Goal Programming Hidden Markov Dengan Feature Selection. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003241043-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Stunting merupakan permasalahan kesehatan masyarakat yang krusial di Provinsi Nusa Tenggara Timur (NTT), yang ditandai oleh prevalensi yang tinggi serta ketimpangan antarwilayah. Penelitian ini bertujuan untuk mengembangkan model Fugomarkov sebagai pendekatan analitis hibrida dalam mengidentifikasi determinan multidimensi stunting, mengoptimalkan pencapaian target penurunan prevalensi, serta memetakan dinamika status wilayah secara spasial dan temporal. Model Fugomarkov mengintegrasikan Multiview Attribute Fuzzy C-Means (MVFCM) untuk seleksi fitur berbasis keanggotaan fuzzy lintas dimensi, Fuzzy Goal Programming (FGP) sebagai metode optimasi multi-tujuan dalam penentuan kombinasi intervensi yang mendekati target yang diharapkan, serta Hidden Markov Model (HMM) untuk klasifikasi state laten dan analisis transisi wilayah. Hasil penelitian menunjukkan bahwa MV-FCM mampu mengidentifikasi determinan stunting yang dominan pada berbagai tingkat wilayah, yaitu aspek sanitasi dan lingkungan dengan indikator akses air bersih pada tingkat desa, aspek kesehatan dengan indikator tenaga medis pada tingkat kecamatan, serta aspek perlindungan sosial dengan indikator Anggaran Pendapatan dan Belanja Desa (APB Desa) pada tingkat kabupaten. FGP menghasilkan kombinasi optimal intervensi sanitasi dan lingkungan, khususnya pembangunan sumur gali dan embung, yang efektif dalam mencapai target penurunan prevalensi stunting pada tingkat desa di Nusa Tenggara Timur. Temuan pada tingkat kecamatan mengindikasikan perlunya peningkatan ketersediaan bidan dan dokter spesialis lainnya, serta penguatan kapasitas fiskal melalui efektifitas penggunaan APB Desa pada tingkat kabupaten. Selanjutnya, HMM memetakan dinamika spasial wilayah melalui pengelompokan state laten dan analisis transisi status, sehingga model hibrida Fugomarkov mampu menghasilkan rekomendasi kebijakan yang adaptif, kontekstual, dan berbasis bukti.
====================================================================================================================================
Stunting remains a critical public health challenge in the Province of East Nusa Tenggara (NTT), Indonesia, characterized by persistently high prevalence rates and substantial interregional disparities. This study aims to develop the Fugomarkov model as a hybrid analytical approach to identify the multidimensional determinants of stunting, optimize the achievement of prevalence reduction targets, and map the spatial–temporal dynamics of regional status. The Fugomarkov framework integrates Multiview Attribute Fuzzy C-Means (MV-FCM) for feature selection based on cross-dimensional fuzzy membership information, Fuzzy Goal Programming (FGP) as a multi-objective optimization method for determining intervention combinations that approximate desired targets, and Hidden Markov Models (HMM) for latent state classification and regional transition analysis. The
results indicate that MV-FCM effectively identifies dominant stunting determinants across multiple spatial levels, namely sanitation and environmental factors represented by access to clean water at the village level, health system capacity
indicated by the availability of medical personnel at the sub-district level, and social protection aspects represented by Village Budget Allocation (APB Desa) at the
district level. FGP yields an optimal combination of sanitation and environmental interventions, particularly the development of dug wells and water reservoirs, which effectively supports stunting prevalence reduction at the village level in East Nusa Tenggara. Findings at the sub-district level further highlight the need to increase the availability of midwives and other medical specialists, as well as to strengthen fiscal capacity through the effective utilization of village budget allocations at the district level. Furthermore, HMM captures spatial dynamics through latent state clustering and transition analysis, enabling the hybrid Fugomarkov model to generate adaptive, contextual, and evidence-based policy recommendations.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Stunting, Fuzzy Goal Programming, Hidden Markov Model, Multiview Feature Selection, Policy Modeling Stunting, Fuzzy Goal Programming, Hidden Markov Model, Multiview Feature Selection, Policy Modeling
Subjects:	Q Science > QA Mathematics > QA274.7 Markov processes--Mathematical models.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Alya Fitri Syalsabilla
Date Deposited:	08 Jan 2026 03:10
Last Modified:	08 Jan 2026 03:10
URI:	http://repository.its.ac.id/id/eprint/129370

Actions (login required)

View Item