Model Geographically And Temporally Weighted Log Logistic 3-Parameter Regression ( Studi Kasus: Indeks Keparahan Kemiskinan di Provinsi Jawa Timur Tahun 2022-2024 )

Huda, Nur (2026) Model Geographically And Temporally Weighted Log Logistic 3-Parameter Regression ( Studi Kasus: Indeks Keparahan Kemiskinan di Provinsi Jawa Timur Tahun 2022-2024 ). Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003241040-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (13MB) | Request a copy

Abstract

Geographically and Temporally Weighted Log Logistic 3-Parameter Regression (GTWLL(3)R) merupakan model pengembangan dari LL(3)R yang secara simultan mampu menangkap heterogenitas spasial dan temporal pada indeks keparahan kemiskinan. Dengan menggunakan indeks keparahan kemiskinan provinsi Jawa Timur tahun 2022-2024, parameter lokal diestimasi menggunakan Metode Maximum Likelihood Estimation (MLE) melalui matriks pembobot fixed gaussian kernel berdasarkan jarak spasial dan temporal, dengan optimasi menggunakan algoritma Newton Raphson. Kinerja model dievaluasi menggunakan Akaike Information Criterion corrected (AICc). Hasil penelitian menunjukkan bahwa GTWLL(3)R lebih baik serta mampu memberikan penjelasan yang lebih kuat dibandingkan model LL(3)R dan GWLL(3)R dengan memperoleh nilai AICc terendah sebesar 18,31159. Koefisien yang diestimasi bervariasi antar kabupaten/kota dan periode waktu, sehingga menunjukkan pola pengaruh prediktor yang berbeda terhadap indeks keparahan kemiskinan. Berdasarkan variabel prediktor yang signifikan, kabupaten/kota tersebut diklasifikasikan ke dalam tiga klaster. Temuan ini menunjukkan bahwa integrasi LL(3)R ke dalam kerangka GTWLL(3)R memberikan pendekatan yang lebih fleksibel dan akurat untuk menganalisis dinamika kemiskinan spasial-temporal, serta menyediakan bukti yang lebih kuat untuk kebijakan penanggulangan kemiskinan yang lebih terarah.
========================================================================================================================================
Geographically Weighted and Temporally Weighted Log-Logistic 3-parameter Regression (GTWLL(3)R) is an extension of the LL(3)R model that can simultaneously capture spatial and temporal heterogeneity in the poverty severity index. Using the poverty severity index for East Java Province from 2022 to 2024, local parameters were estimated using the Maximum Likelihood Estimation (MLE) method via an fixed Gaussian kernel weighting matrix based on spatial and temporal distances, with optimization using the Newton-Raphson algorithm. Model performance was evaluated using the Akaike Information Criterion corrected (AICc). The results indicate that the GTWLL(3)R model outperforms the LL(3)R and GWLL(3)R models, providing a stronger explanation and achieving the lowest AICc value of 18.31159. The estimated coefficients vary across districts/cities and time periods, revealing distinct patterns of predictor influence on the poverty severity index. Based on significant predictor variables, the districts/cities were classified into three clusters. These findings indicate that integrating LL(3)R into the GTWLL(3)R framework provides a more flexible and accurate approach for analyzing the spatial-temporal dynamics of poverty, as well as providing stronger evidence for more targeted poverty alleviation policies.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Distribusi Log-Logistik 3-Parameter, GTWLL(3)R, Indeks Keparahan Kemiskinan
Subjects:	H Social Sciences > HA Statistics
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Nur Huda
Date Deposited:	20 Jul 2026 07:24
Last Modified:	20 Jul 2026 07:24
URI:	http://repository.its.ac.id/id/eprint/135732

Actions (login required)

View Item