Amanda, Rizki (2026) Penerapan Model Deepsurv dalam Analisis Survival dan Prediksi Mortalitas pada Pasien Hemodialisis. Undergraduate thesis, UPN Veteran Jawa Timur.
|
Text (Cover)
22083010045.-cover.pdf Download (968kB) | Preview |
|
|
Text (Bab 1)
22083010045.-bab1.pdf Download (225kB) | Preview |
|
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Bab 2)
22083010045.-bab2.pdf Restricted to Repository staff only until 7 July 2029. Download (973kB) |
|
|
Text (Bab 3)
22083010045.-bab3.pdf Restricted to Repository staff only until 7 July 2029. Download (1MB) |
|
|
Text (Bab 4)
22083010045.-bab4.pdf Restricted to Repository staff only until 7 July 2029. Download (2MB) |
|
|
Text (Bab 5)
22083010045.-bab5.pdf Download (198kB) | Preview |
|
|
Text (Daftar Pustaka)
22083010045.-daftarpustaka.pdf Download (183kB) | Preview |
|
|
Text (Lampiran)
22083010045.-lampiran.pdf Restricted to Repository staff only until 7 July 2029. Download (429kB) |
Abstract
Chronic Kidney Disease (CKD) is one of the leading causes of mortality in Indonesia. Nationally, deaths caused by kidney failure exceed 42,000 cases annually, while data from the Surabaya Health Office reported 308 CKD patients as of June 2024. The high mortality risk among hemodialysis patients highlights the need for a predictive system capable of analyzing patient survival and mortality risk at an individual level to support clinical decision making. This study aims to analyze survival outcomes and predict mortality risk among hemodialysis patients at RSUD Haji Surabaya using the DeepSurv method. The dataset was divided into a training set comprising 80% (243 patients) and a testing set comprising 20% (61 patients). DeepSurv was selected because of its ability to model non linear relationships and complex interactions among clinical variables that are difficult to capture using conventional survival models. Results from conventional methods showed that only dialysis frequency per month had a statistically significant effect on patient survival, whereas the DeepSurv model indicated that all clinical variables contributed to mortality risk prediction. Based on model evaluation, DeepSurv achieved a C-index of 0.901 on the training set and 0.931 on the testing set, outperforming the Cox Proportional Hazard model (0.852) and Cox Spline model (0.866). Furthermore, DeepSurv obtained the lowest Integrated Brier Score (IBS) of 0.0557, indicassssting the smallest prediction error among the compared models. The study also implemented the prediction results in a Graphical User Interface (GUI) to provide real-time mortality probability estimates and patient risk stratification. These findings demonstrate that DeepSurv delivers superior predictive performance compared with conventional survival analysis methods for predicting mortality risk in hemodialysis patients.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | Q Science > QA Mathematics Q Science > QA Mathematics > QA76.6 Computer Programming Q Science > QA Mathematics > QA76.87 Neural computers R Medicine > R Medicine (General) T Technology > T Technology (General) T Technology > T Technology (General) > T58.6-58.62 Management Information Systems |
||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Data Science | ||||||||||||
| Depositing User: | Rizki Amanda Amanda | ||||||||||||
| Date Deposited: | 07 Jul 2026 06:38 | ||||||||||||
| Last Modified: | 07 Jul 2026 07:16 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/54763 |
Actions (login required)
 |
View Item |