PREDIKSI LoS PADA BAYI PREMATUR DENGAN MENGGUNAKAN MACHINE LEARNING DI RSIA BUNDA AISYAH KOTA TASIKMALAYA

Fauzi, Rasti Shindy (2025) PREDIKSI LoS PADA BAYI PREMATUR DENGAN MENGGUNAKAN MACHINE LEARNING DI RSIA BUNDA AISYAH KOTA TASIKMALAYA. Diploma thesis, POLITEKNIK KESEHATAN TASIKMALAYA.

[img] Text
Orisinalitas.pdf

Download (224kB)
[img] Text
Halaman Pengesahan KTI.pdf

Download (591kB)
[img] Text
Persetujuan Publikasi.pdf

Download (391kB)
[img] Text
Halaman Judul.pdf

Download (535kB)
[img] Text
BAB I.pdf

Download (161kB)
[img] Text
BAB II.pdf
Restricted to Repository staff only

Download (277kB)
[img] Text
BAB III.pdf
Restricted to Repository staff only

Download (269kB)
[img] Text
BAB IV.pdf
Restricted to Repository staff only

Download (731kB)
[img] Text
BAB V.pdf
Restricted to Repository staff only

Download (137kB)
[img] Text
Daftar Pustaka.pdf
Restricted to Repository staff only

Download (263kB)
[img] Text
Lampiran.pdf
Restricted to Repository staff only

Download (728kB)

Abstract

Background: Preterm infants are highly vulnerable to health complications, often requiring prolonged and intensive care. Accurately predicting Length of Stay (LoS) is essential for improving resource allocation and healthcare planning. However, RSIA Bunda Aisyah in Tasikmalaya currently lacks a specific LoS prediction model for preterm infants. Research Methods: This study applied a quantitative retrospective design using 501 medical records of preterm infants hospitalized between January 2022 and December 2024. Data were analyzed using Orange Data Mining with three regression-based machine learning algorithms: Linear Regression, k-Nearest Neighbors (kNN), and Random Forest Regression. Feature selection employed the RReliefF method, while model performance was evaluated through Stratified 20-Fold Cross Validation. Results: The analysis identified key clinical factors influencing LoS, including infant age, gestational age, clinical diagnoses (P07.0, P07.1, P22.0, P61.4, P90), and the type of medical interventions received. Non-clinical factors such as infant gender, payment status, and maternal medical history also showed notable influence. Among the tested models, the kNN algorithm outperformed others with a Mean Absolute Error (MAE) of 1.278 days, Root Mean Squared Error (RMSE) of 3.219 days, and R² value of 0.506. Visualization with box and scatter plots demonstrated strong model performance in general cases, though limitations remained in predicting extreme outliers. Conclusion: The k-NN algorithm proved to be the most effective model for predicting LoS among preterm infants and offers valuable support for hospitals in optimizing neonatal care, managing capacity, and planning resources more efficiently. Keywords: Length of Stay, preterm infants, prediction, machine learning, Orange, neonatal care, hospital resource planning Bibliography: 47 (2012-2024)

Item Type: Thesis (Diploma)
Subjects: R Medicine > RA Public aspects of medicine
R Medicine > RC Internal medicine
R Medicine > RJ Pediatrics
Divisions: Jurusan RMIK > D3 Rekam Medis & Informasi Kesehatan
Depositing User: Mhs Shindy Rasti Fauzi
Date Deposited: 02 Sep 2025 08:21
Last Modified: 02 Sep 2025 08:21
URI: http://repo.poltekkestasikmalaya.ac.id/id/eprint/6917

Actions (login required)

View Item View Item