Bisono, Eva Firdayanti and Fernanda, Jerhi Wahyu and Nurkhalim, Ratna Frenty and Jayanti, Krisnita Dwi Length Of Stay (LOS) Prediction of Type 2 Diabetes Mellitus Using Classification and Regression Tree (CART). -. (Unpublished)
![[thumbnail of Similarity]](http://eprints.iik.ac.id/style/images/fileicons/text.png)
LENGTH OF STAY (LOS) PREDICTION OF TYPE 2 DIABETES MELLITUS USING CLASSIFICATION AND REGRESSION TREE (CART).pdf - Draft Version
Download (1MB)
![[thumbnail of Laporan Akhir Penelitian B. Eva]](http://eprints.iik.ac.id/style/images/fileicons/text.png)
Prediksi Prevalensi Diabetes Tipe 2 menggunakan Artificial Neural Network (2) - Eva Firdayanti Bisono IIK BW.pdf - Other
Download (3MB)
Abstract
The prediction of LOS in type 2 patients and the influencing factors can be used as a basis for managing comorbidities and the risk of complications in patients. Predictions can be made using machine learning methods such as Classification and Regression Tree (CART). This study aims to analyse the factors that influence the LOS of type 2 DM patients. The research data was obtained from the Hospital Information System in the period 2019 to 2021 and obtained data for 541 type 2 DM patients. The study variables consisted of the dependent variable, namely LOS of DM patients type 2 and the independent variables consisted of gender, age, complications, comorbidities and urban status of type 2 DM patients. The average LOS of type 2 DM patients was 3.39 days with a median of 3 days. The results of the analysis using CART with 10-fold cross validation concluded that the morbidity variable was the variable that most dominantly influenced the LOS of type 2 DM patients. Accuracy, precision, recall, and F1 scores were respectively 0.704, 0.814, and 0.755.
Item Type: | Article |
---|---|
Kata Kunci: | Classification and Regression Tree, Length of Stay, Prediction, Type 2 Diabetes Mellitus |
Subjects: | R Medicine > R Medicine (General) |
Divisions: | Karya Dosen > Rekam Medis & Informasi Kesehatan |
Depositing User: | admin |
Date Deposited: | 15 Jun 2024 02:32 |
Last Modified: | 15 Jun 2024 02:32 |
URI: | http://eprints.iik.ac.id/id/eprint/392 |