Pemodelan Status Migran Risen Masuk Di Provinsi Jawa Timur Menggunakan Metode Rare Event Weighted Logistic Regression (RE-WLR)

Ghaliyah, Mawar Jannah (2026) Pemodelan Status Migran Risen Masuk Di Provinsi Jawa Timur Menggunakan Metode Rare Event Weighted Logistic Regression (RE-WLR). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221005-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Migrasi merupakan salah satu komponen demografi yang memengaruhi jumlah, komposisi, dan persebaran penduduk. Salah satu indikator mobilitas penduduk terkini adalah migrasi risen, yaitu perpindahan tempat tinggal yang ditunjukkan oleh perbedaan wilayah tempat tinggal saat pencacahan dengan lima tahun sebelumnya. Provinsi Jawa Timur termasuk dalam lima besar provinsi tujuan migran risen antarprovinsi di Indonesia. Penelitian ini bertujuan untuk mendeskripsikan karakteristik migran risen masuk dan memodelkan pengaruh faktor sosial demografi, pendidikan, ekonomi, serta akses teknologi informasi terhadap status migran risen masuk di Provinsi Jawa Timur. Data yang digunakan merupakan data SUSENAS KOR Tahun 2025 yang mencakup 93.047 penduduk berusia lima tahun atau lebih, terdiri atas 526 migran risen atau sebesar 0,57% dan 92.521 bukan migran risen atau sebesar 99,43%. Proporsi migran risen yang sangat kecil menunjukkan bahwa migran risen masuk merupakan kejadian langka sehingga membentuk data tidak seimbang. Data dibagi menjadi data training dan testing dengan proporsi 80:20 menggunakan teknik stratified sampling. Ketidakseimbangan kelas ditangani menggunakan metode Rare Event Weighted Logistic Regression (RE-WLR) melalui pembobotan, regularisasi, dan koreksi bias. Hasil uji independensi menunjukkan bahwa delapan dari empat belas variabel prediktor memiliki hubungan signifikan dengan status migran risen masuk. Model RE-WLR terbaik diperoleh pada skema rasio kelas 1:115 dengan parameter regularisasi λ=1. Setelah dilakukan backward elimination, variabel yang dipertahankan dalam model meliputi hubungan dengan KRT (X1), usia kawin pertama (X4), status penggunaan produk jasa keuangan (X8), akses internet (X10), status kepemilikan rumah (X11), dan kepemilikan usaha mikro kecil (X13). Model menghasilkan akurasi sebesar 80,66%, sensitivitas 44,76%, spesifisitas 80,86%, G-mean 60,16%, dan AUC 69,20%.
=================================================================================================================================
Migration is one of the demographic components that influences population size, composition, and distribution. One indicator of recent population mobility is recent migration, defined as a change in place of residence indicated by a difference between an individual’s place of residence at the time of enumeration and five years earlier. East Java Province is among the five major destination provinces for interprovincial recent migrants in Indonesia. This study aims to describe the characteristics of in-migrants and to model the effects of sociodemographic, educational, economic, and information technology access factors on recent in-migration status in East Java Province. The data were obtained from the 2025 National Socioeconomic Survey (SUSENAS) Core and covered 93,047 individuals aged five years or older, consisting of 526 recent migrants, or 0.57 percent, and 92,521 non-recent migrants, or 99.43 percent. The very small proportion of recent migrants indicates that recent in-migration is a rare event, resulting in imbalanced data. The data were divided into training and testing sets in an 80:20 proportion using stratified sampling. Class imbalance was addressed using Rare Event Weighted Logistic Regression (RE-WLR) through weighting, regularization, and bias correction. The independence tests showed that eight of the fourteen predictor variables were significantly associated with recent in-migration status. The best RE-WLR model was obtained using a class ratio scheme of 1:115 and a regularization parameter of λ=1. After backward elimination, the variables retained in the model were relationship to the household head (X1), age at first marriage (X4), use of financial service products (X8), internet access (X10), home ownership status (X11), and ownership of a micro or small enterprise (X13). The model achieved an accuracy of 80.66%, sensitivity of 44.76%, specificity of 80.86%, G-mean of 60.16%, and AUC of 69.20%.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Data Tidak Seimbang, Kejadian Langka, Migrasi Risen, Rare Event Weighted Logistic Regression, Imbalanced Data, Rare Events, Recent Migration, Rare Event Weighted Logistic Regression
Subjects:	H Social Sciences > HA Statistics H Social Sciences > HA Statistics > HA31.3 Regression. Correlation. Logistic regression analysis. H Social Sciences > HA Statistics > HA31.7 Estimation Q Science Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA401 Mathematical models.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Mawar Jannah Ghaliyah
Date Deposited:	29 Jul 2026 01:43
Last Modified:	29 Jul 2026 01:43
URI:	http://repository.its.ac.id/id/eprint/139367

Actions (login required)

View Item