Pemodelan Jumlah Kasus Tuberkulosis di Pulau Kalimantan Menggunakan Metode Bootstrap Aggregating Multivariate Adaptive Generalized Poisson Regression Splines (Bagging MAGPRS)

Pratiwi, Andhini (2026) Pemodelan Jumlah Kasus Tuberkulosis di Pulau Kalimantan Menggunakan Metode Bootstrap Aggregating Multivariate Adaptive Generalized Poisson Regression Splines (Bagging MAGPRS). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221191-Undergraduate_thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Tuberkulosis merupakan penyakit infeksi menular yang masih menjadi permasalahan kesehatan masyarakat di Indonesia, termasuk di Pulau Kalimantan. Jumlah kasus tuberkulosis dipengaruhi oleh berbagai faktor lingkungan, kependudukan, fasilitas kesehatan, dan sosial ekonomi, serta memiliki karakteristik sebagai data cacah yang sering mengalami overdispersi dan hubungan nonlinier antarvariabel. Kondisi tersebut menyebabkan metode regresi konvensional kurang optimal sehingga diperlukan pendekatan pemodelan yang lebih fleksibel dan stabil. Selanjutnya, teknik bootstrap aggregating diterapkan untuk meningkatkan kestabilan dan akurasi model regresi. Penelitian ini bertujuan untuk memodelkan jumlah kasus tuberkulosis dan membandingkan kinerja metode Multivariate Adaptive Generalized Poisson Regression Splines (MAGPRS) dengan Bootstrap Aggregating (Bagging MAGPRS). Hasil pemodelan menggunakan MAGPRS menunjukkan bahwa seluruh variabel prediktor berpengaruh terhadap jumlah kasus tuberkulosis, yaitu kepadatan penduduk, rasio puskesmas, rasio rumah sakit, persentase rumah tangga yang memiliki akses terhadap sanitasi layak, persentase rumah tangga yang memiliki akses terhadap air minum layak, dan persentase penduduk berusia 15 tahun ke atas yang merokok dalam sebulan terakhir. Model terbaik MAGPRS diperoleh pada kombinasi Basis Function (BF) = 24, Maximum Interaction (MI) = 3, dan Minimum Observation (MO) = 1 dengan nilai Generalized Cross Validation (GCC) sebesar 40.324,38 dan koefisien determinasi (R2) sebesar 0,933. Penerapan metode bagging MAGPRS menunjukkan peningkatan kinerja model yang ditunjukkan oleh penurunan nilai GCV menjadi 5.519,065 dan peningkatan nilai R2 menjadi 0,99 yang mengindikasikan bahwa model memiliki akurasi dan stabilitasi prediksi yang lebih baik sehingga pendekatan bagging MAGPRS dapat dikatakan lebih efektif dalam memodelkan dan memprediksi jumlah kasus tuberkulosis di Pulau Kalimantan.
=====================================================================================================================================
Tuberculosis is an infectious disease that remains a major public health problem in Indonesia, including on the island of Kalimantan. The number of tuberculosis cases is influenced by various environmental, demographic, healthcare facility, and sosioeconomic factors, and exhibits characteristics of count data that often experience overdispersion and nonlinear relationships among variables. These conditions make conventional regression methods less optimal, thereby requiring more flexible and stable modeling approaches. Furthermore, the bootstrap aggregating technique is applied to improve the stability and accuracy of the regression model. This study aims to model the number of tuberculosis cases and compare the performance of the Multivariate Adaptive Generalized Poisson Regression Splines (MAGPRS) method with Bootstrap Aggregating (Bagging MAGPRS). The modelling results using MAGPRS indicate that all predictor variables influence the number of tuberculosis cases, namely population density, the ratio of community health centers, the ratio of hospitals, the percentage of households with access to adequate sanitation, the percentage of households with access to safe drinking water, and the percentage of the population aged 15 years and older who have smoked in the past month. The best MAGPRS model was obtained with a combination of Basis Function (BF) = 24, Maximum Interaction (MI) = 3, and Minimum Observation (MO) = 1, with a Generalized Cross-Validation (GCV) value of 40,324.38 and a coefficient of determination (R²) of 0.933. The application of the MAGPRS bagging method showed an improvement in model performance, as indicated by a decrease in the GCV value to 5,519.065 and an increase in the R² value to 0.99. This indicates that the model has better prediction accuracy and stability, suggesting that the MAGPRS bagging approach is more effective in modeling and predicting the number of tuberculosis cases on the island of Kalimantan.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Bagging, MAGPRS, Poisson, Pulau Kalimantan, Tuberkulosis, Bagging, East Kalimantan, MAGPRS, Poisson, Tuberculosis
Subjects:	Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Andhini Pratiwi
Date Deposited:	04 Aug 2026 07:36
Last Modified:	04 Aug 2026 07:36
URI:	http://repository.its.ac.id/id/eprint/140992

Actions (login required)

View Item