Pemodelan Angka Kesakitan Demam Berdarah Dengue di Pulau Jawa dan Bali Menggunakan Metode Hibrida MARS-Random Forest

Wijayanti, Evyana (2026) Pemodelan Angka Kesakitan Demam Berdarah Dengue di Pulau Jawa dan Bali Menggunakan Metode Hibrida MARS-Random Forest. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221080-Undergaduate_thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Demam Berdarah Dengue (DBD) merupakan penyakit menular yang masih menjadi masalah kesehatan masyarakat di Indonesia. Penelitian ini bertujuan mengidentifikasi faktor-faktor yang memengaruhi angka kesakitan DBD di kabupaten/kota Pulau Jawa dan Bali menggunakan metode Multivariate Adaptive Regression Splines (MARS) dan random forest, membangun model hibrida MARS-random forest, serta membandingkan performa ketiga metode. Data yang digunakan merupakan data sekunder tahun 2024 pada 128 kabupaten/kota di Pulau Jawa dan Bali, menjadi 97 kabupaten/kota setelah preprocessing. Variabel respon berupa angka kesakitan DBD per 100.000 penduduk, sedangkan variabel prediktor terdiri atas kepadatan penduduk (X1), persentase penduduk miskin (X2), sanitasi layak (X3), akses air minum layak (X4), jumlah fasilitas kesehatan (X5), kepemilikan jamban milik sendiri (X6), dan curah hujan tahunan (X7). Evaluasi model menggunakan Mean Squared Error (MSE), Root Mean Squared Error (RMSE), dan Mean Absolute Error (MAE). Model MARS terbaik diperoleh pada parameter BF sebanyak 21, MI sebesar 1, dan MO sebesar 1, dengan tiga variabel yang dipertahankan, yaitu X3, X6, dan X7. Model Random Forest terbaik menggunakan mtry = 2 dan n-tree = 500. Model hibrida MARS-random forest dibangun menggunakan tiga fungsi basis MARS sebagai variabel masukan, dengan parameter terbaik mtry = 1 dan n-tree = 500. Berdasarkan evaluasi data uji, random forest menghasilkan kesalahan prediksi terendah (MSE = 6822,702; RMSE = 82,5997; MAE = 60,0867), diikuti model hibrida (MSE = 9383,311; RMSE = 96,8675; MAE = 70,5195), sedangkan MARS menghasilkan kesalahan prediksi tertinggi. Dengan demikian, random forest merupakan model terbaik memprediksi angka kesakitan DBD di kabupaten/kota Pulau Jawa dan Bali tahun 2024.
=======================================================================================================================================
Dengue Hemorrhagic Fever (DHF) is an infectious disease that remains a public health problem in Indonesia. This study aims to identify the factors affecting the DHF incidence rate in regencies/cities in Java and Bali using the Multivariate Adaptive Regression Splines (MARS) and random forest methods, to build a MARS-random forest hybrid model, and to compare the performance of the three methods. The data used are secondary data from 2024 covering 128 regencies/cities in Java and Bali, reduced to 97 regencies/cities after preprocessing. The response variable is the DHF incidence rate per 100,000 population, while the predictor variables consist of population density (X1), percentage of poor population (X2), proper sanitation (X3), access to proper drinking water (X4), number of health facilities (X5), ownership of private latrines (X6), and annual rainfall (X7). Model evaluation was conducted using Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE). The best MARS model was obtained with parameters of 21 basis functions (BF), an interaction degree (MI) of 1, and a minimum observation (MO) of 1, retaining three variables, namely X3, X6, and X7. The best random forest model used mtry = 2 and n-tree = 500. The MARS-random forest hybrid model was built using the three MARS basis functions as input variables, with the best parameters being mtry = 1 and n-tree = 500. Based on the test data evaluation, Random Forest produced the lowest prediction error (MSE = 6822.702; RMSE = 82.5997; MAE = 60.0867), followed by the hybrid model (MSE = 9383.311; RMSE = 96.8675; MAE = 70.5195), while MARS produced the highest prediction error. Thus, random forest is the best model for predicting the DHF incidence rate in regencies/cities in Java and Bali in 2024.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Angka kesakitan DBD, Demam Berdarah Dengue, Hibrida MARS-Random Forest, Multivariate Adaptive Regression Splines (MARS), Random Forest, Dengue Hemorrhagic Fever, DHF incidence rate, Hybrid MARS-Random Forest, Multivariate Adaptive Regression Splines (MARS), Random Forest
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Evyana Wijayanti
Date Deposited:	01 Aug 2026 01:51
Last Modified:	01 Aug 2026 01:51
URI:	http://repository.its.ac.id/id/eprint/140852

Actions (login required)

View Item