PREDIKSI MORTALITAS PADA PASIEN INSTALASI PERAWATAN INTENSIF MENGGUNAKAN DEEP LEARNING

UTAMI, RAHMADINA MAULIA and Firdaus, Firdaus and Tutuko, Bambang (2024) PREDIKSI MORTALITAS PADA PASIEN INSTALASI PERAWATAN INTENSIF MENGGUNAKAN DEEP LEARNING. Undergraduate thesis, Sriwijaya University.

	Text RAMA_56201_09011182025003.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (3MB) | Request a copy
	Text RAMA_56201_09011182025003_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (5MB) | Request a copy
	Text RAMA_56201_09011182025003_0221017801_0012016003_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (702kB)
	Text RAMA_56201_09011182025003_0221017801_0012016003_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (534kB) | Request a copy
	Text RAMA_56201_09011182025003_0221017801_0012016003_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (634kB) | Request a copy
	Text RAMA_56201_09011182025003_0221017801_0012016003_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_56201_09011182025003_0221017801_0012016003_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (184kB) | Request a copy
	Text RAMA_56201_09011182025003_0221017801_0012016003_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (196kB) | Request a copy
	Text RAMA_56201_09011182025003_0221017801_0012016003_07_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (858kB) | Request a copy

Abstract

Mortality prediction for patients in intensive care units (ICUs) is crucial for care planning, resource allocation, and timely medical decision-making. To enhance the accuracy of these predictions, artificial intelligence (AI) technology has shown promising results. This study aims to develop deep learning methods to improve the accuracy of mortality predictions for ICU patients, addressing missing values and class imbalance in multivariate time-series data. The research methodology includes literature review, data collection from the MIMIC-IV database, data preprocessing, model development, testing, validation, and result analysis. The preprocessing stage involves feature selection, data pivoting, data filtering, data encoding, data imputation, and data balancing. The deep learning methods applied include GRU, LSTM, RNN, CNN, Bi-LSTM, Stacked LSTM, and 1D-MSNet. Data imputation is performed using linear interpolation and XU-Net, while undersampling and SMOTE techniques are used to address class imbalance. The study results show that the 1D-MSNet method achieves 99% accuracy on training data and 96% on validation and testing data, with an AUC value approaching 100%, indicating very high predictive capability. In conclusion, deep learning methods, particularly 1D-MSNet, are effective in improving the accuracy of mortality predictions for ICU patients. The use of XU-Net is effective in handling missing values, and SMOTE techniques are effective in addressing class imbalance, significantly enhancing model performance. Keywords : Mortality prediction, deep learning, MIMIC-IV

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Mortality prediction, deep learning, MIMIC-IV
Subjects:	Q Science > Q Science (General) > Q334-342 Computer science. Artificial intelligence. Algorithms. Robotics. Automation.
Divisions:	09-Faculty of Computer Science > 56201-Computer Systems (S1)
Depositing User:	Rahmadina Maulia Utami
Date Deposited:	28 Aug 2024 05:57
Last Modified:	28 Aug 2024 05:57
URI:	http://repository.unsri.ac.id/id/eprint/156292

Actions (login required)

View Item