Pemodelan Faktor-Faktor yang Memengaruhi Kasus Pneumonia Balita di Jawa Timur Menggunakan Geographically Weighted Log-Normal Regression

Shofiatun, Nur (2025) Pemodelan Faktor-Faktor yang Memengaruhi Kasus Pneumonia Balita di Jawa Timur Menggunakan Geographically Weighted Log-Normal Regression. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003211042-Undergraduate_Theses.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Pneumonia merupakan infeksi saluran pernapasan akut yang menyerang jaringan paru-paru dan banyak menyerang balita karena sistem kekebalan tubuh mereka yang masih lemah. Pada tahun 2019, WHO mencatat 740.180 kematian balita akibat pneumonia, yaitu setara 14% dari total kematian anak di bawah usia lima tahun. Di Jawa Timur, persentase kasus pneumonia balita meningkat dari 74,7% pada 2022 menjadi 79,7% pada 2023. Meskipun dapat dicegah melalui imunisasi, gizi baik, dan lingkungan sehat, pneumonia tetap menjadi masalah kesehatan yang cukup serius. Penelitian ini bertujuan untuk mengetahui karakteristik data serta memodelkan dan mengidentifikasi faktor-faktor yang memengaruhi persentase kasus pneumonia balita di Provinsi Jawa Timur tahun 2023. Variabel prediktor yang digunakan, yaitu persentase imunisasi dasar lengkap pada balita, persentase pemberian vitamin A pada balita, persentase kepala keluarga dengan pengelolaan kualitas udara rumah tangga, persentase rumah tangga dengan bahan bakar utama untuk memasak adalah kayu, dan persentase rumah tangga dengan jenis lantai tanah. Uji Kolmogorov-Smirnov menunjukkan data berdistribusi log-normal, sedangkan uji Breusch-Pagan menunjukkan adanya heterogenitas spasial, sehingga dilakukan pemodelan spasial dengan menggunakan metode Geographically Weighted Log-Normal Regression. Pemodelan dilakukan menggunakan dua fungsi pembobot, yaitu Adaptive Bisquare Kernel dan Adaptive Tricube Kernel. Hasil menunjukkan bahwa kedua model memiliki kinerja yang hampir sama karena selisih nilai AICc keduanya sangat kecil, sehingga kedua model merupakan model terbaik dalam memodelkan persentase kasus pneumonia balita di Jawa Timur tahun 2023. Kedua model menghasilkan tiga kelompok kabupaten/kota berdasarkan variabel prediktor signifikan terhadap persentase pneumonia balita, yaitu kelompok 1 dipengaruhi oleh persentase vitamain A, kelompok 2 oleh persentase rumah tangga dengan bahan bakar memasak kayu, dan kelompok 3 tanpa variabel signifikan. Pola pengelompokan GWLNR dengan Adaptive Tricube dan Bisquare Kernel serupa, kecuali di Kabupaten Sumenep, di mana pada Adaptive Tricube Kernel wilayah masuk kelompok 1, sedangkan pada Adaptive Bisquare Kernel masuk kelompok 3.
===========================================================================================================================================
Pneumonia is an acute respiratory infection that affects lung tissue and commonly affects children under five due to their weak immune systems. In 2019, WHO recorded 740,180 deaths of children under five due to pneumonia, accounting for 14% of total deaths in that age group. In East Java, the percentage of pneumonia cases in children under five increased from 74.7% in 2022 to 79.7% in 2023. Although preventable through immunization, proper nutrition, and a healthy environment, pneumonia remains a serious public health issue. This study aims to explore the characteristics of the data, model, and identify the factors influencing the percentage of pneumonia cases in children under five in East Java Province in 2023. The predictor variables used are the percentage of children under five receiving complete basic immunization, the percentage receiving vitamin A, the percentage of household heads managing indoor air quality, the percentage of households using firewood as the main cooking fuel, and the percentage of households with earthen floors. The Kolmogorov-Smirnov test showed that the data follows a log-normal distribution, while the Breusch-Pagan test indicated the presence of spatial heterogeneity. Therefore, spatial modeling was conducted using the Geographically Weighted Log-Normal Regression method with two kernel functions: Adaptive Bisquare Kernel and Adaptive Tricube Kernel. The results showed that both models performed similarly, as indicated by a very small difference in AICc values, making them both suitable for modeling the percentage of pneumonia cases in East Java in 2023. Both models produced three groups of districts/cities based on significant predictor variables for the percentage of pneumonia in infants, namely group 1 influenced by the percentage of vitamin A, group 2 by the percentage of households using wood as cooking fuel, and group 3 without significant variables. The GWLNR clustering patterns using the Adaptive Tricube and Bisquare Kernels are similar, except in Sumenep District, where the region is classified into group 1 using the Adaptive Tricube Kernel, while it is classified into group 3 using the Adaptive Bisquare Kernel.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Geographically Weighted Log-Normal Regression (GWLNR), Jawa Timur, Log-Normal Regression (LNR), pneumonia balita
Subjects:	Q Science
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Nur Shofiatun
Date Deposited:	31 Jul 2025 02:56
Last Modified:	31 Jul 2025 02:56
URI:	http://repository.its.ac.id/id/eprint/124413

Actions (login required)

View Item