Pemodelan Kasus Jumlah Balita Underweight di Provinsi Jawa Timur Menggunakan Bootstrap Aggregating Multivariate Adaptive Generalized Poisson Regression Splines (Bagging MAGPRS)

Hartanto, Andrew Putra (2025) Pemodelan Kasus Jumlah Balita Underweight di Provinsi Jawa Timur Menggunakan Bootstrap Aggregating Multivariate Adaptive Generalized Poisson Regression Splines (Bagging MAGPRS). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003211016-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Malnutrisi pada balita merupakan salah satu tantangan utama dalam mencapai Sustainable Development Goals (SDGs), terutama terkait dengan tujuan kedua yaitu memastikan akses terhadap pangan yang aman, bergizi, dan berkelanjutan. Pada tahun 2023 persentase balita underweight di Jawa Timur mencapai 6,8%, meningkat dibandingkan tahun 2022 yang sebesar 4,69%. Salah satu langkah preventif dalam menekan peningkatan jumlah kasus underweight pada balita di Jawa Timur adalah melakukan pemodelan jumlah kasus balita underweight menggunakan metode Bootstrap Aggregating Multivariate Adaptive Generalized Poisson Regression Spline (Bagging MAGPRS). Penelitian ini menggunakan metode MARS dengan estimator Generalized Poisson, membentuk metode MAGPRS karena hubungan antara variabel respon dan prediktor bersifat nonlinier dan data mengalami kasus overdispersi, selanjutnya diterapkan teknik bootstrap aggregating untuk meningkatkan kestabilan dan akurasi model regresi. Tujuan penelitian ini adalah untuk mengidentifikasi faktor-faktor yang memengaruhi jumlah kasus balita underweight, membangun model prediksi menggunakan metode MAGPRS, dan membandingkan kinerja antara model MAGPRS dan Bagging MAGPRS. Hasil pemodelan menggunakan MAGPRS menunjukkan bahwa lima variabel yang berpengaruh terhadap jumlah kasus balita underweight adalah persentase keluarga dengan akses sanitasi layak, pengelolaan air dan makanan rumah tangga, balita menderita diare yang dilayani, pemberian ASI eksklusif, serta kunjungan neonatal lengkap. Model terbaik MAGPRS diperoleh dengan kombinasi Basis Function (BF) = 28, Maximum Interaction (MI) = 2, dan Minimum Observation (MO) = 3, dengan nilai Generalized Cross Validation (GCV) sebesar 37.899,94 dan koefisien determinasi (R²) sebesar 0,9943. Penerapan metode Bagging MAGPRS menunjukkan peningkatan kinerja model yang signifikan, dengan nilai GCV menurun menjadi 2.642,683 dan R² meningkat menjadi 0,9996 yang menandakan peningkatan akurasi dan stabilitas prediksi. Dengan demikian, pendekatan bagging MAGPRS terbukti lebih optimal dalam memodelkan dan memprediksi jumlah kasus balita underweight di Jawa Timur.
=======================================================================================================================================
Malnutrition in toddlers is one of the main challenges in achieving the Sustainable Development Goals (SDGs), especially related to the second goal, namely ensuring access to safe, nutritious, and sustainable food. In 2023, the percentage of underweight toddlers in East Java reached 6.8%, an increase compared to 2022 which was 4.69%. One of the preventive measures in suppressing the increase in the number of underweight cases in toddlers in East Java is to model the number of underweight toddler cases using the Bootstrap Aggregating Multivariate Adaptive Generalized Poisson Regression Spline (Bagging MAGPRS) method. This study uses the MARS method with the Generalized Poisson estimator, forming the MAGPRS method because the relationship between the response and predictor variables is nonlinear and the data experiences overdispersion cases, then the bootstrap aggregating technique is applied to increase the stability and accuracy of the regression model. The purpose of this study was to identify factors that influence the number of underweight toddler cases, build a prediction model using the MAGPRS method, and compare the performance between the MAGPRS and Bagging MAGPRS models. The results of modeling using MAGPRS show that the five variables that influence the number of cases of underweight toddlers are the percentage of families with access to proper sanitation, household water and food management, toddlers suffering from diarrhea who are served, exclusive breastfeeding, and complete neonatal visits. The best MAGPRS model was obtained with a combination of Basis Function (BF) = 28, Maximum Interaction (MI) = 2, and Minimum Observation (MO) = 3, with a Generalized Cross Validation (GCV) value of 37,899.94 and a coefficient of determination (R²) of 0.9943. The application of the MAGPRS Bagging method showed a significant increase in model performance, with the GCV value decreasing to 2,642.683 and R² increasing to 0.9996, indicating an increase in prediction accuracy and stability. Thus, the MAGPRS bagging approach is proven to be more optimal in modeling and predicting the number of cases of underweight toddlers in East Java.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Bagging MAGPRS, Balita, Jawa Timur, Poisson, Underweight, East Java, MAGPRS Bagging, Poisson, Toddlers, Underweight
Subjects:	Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics) R Medicine > RJ Pediatrics > RJ101 Child Health. Child health services
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Andrew Putra Hartanto
Date Deposited:	24 Jul 2025 07:39
Last Modified:	24 Jul 2025 07:39
URI:	http://repository.its.ac.id/id/eprint/121290

Actions (login required)

View Item