Pengembangan Aplikasi Healthcare Intelligent System Untuk Pasien Dan Caregiver Alzheimer: Klasifikasi Alzheimer Menggunakan Deep Learning Model Convolutional Neural Network

Hiswara, I Gusti Agung Jaya (2025) Pengembangan Aplikasi Healthcare Intelligent System Untuk Pasien Dan Caregiver Alzheimer: Klasifikasi Alzheimer Menggunakan Deep Learning Model Convolutional Neural Network. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026211122-Undergraduate-Thesis_compressed.pdf Restricted to Repository staff only Download (3MB)

Abstract

Penyakit Alzheimer merupakan salah satu bentuk demensia yang menyebabkan penurunan progresif fungsi memori dan kognitif. Di Indonesia, rendahnya kesadaran masyarakat—dengan 86,3% populasi tidak memahami istilah Alzheimer—menghambat diagnosis dini dan memperberat beban pengasuh. Penelitian ini bertujuan mengembangkan sistem cerdas berbasis Convolutional Neural Network (CNN) untuk klasifikasi Alzheimer melalui citra MRI otak. Dataset MRI diperoleh dari Kaggle dan terdiri dari empat kelas: Mild Demented, Moderate Demented, Non Demented, dan Very Mild Demented, dengan total 6400 citra MRI. Untuk mengatasi ketidakseimbangan kelas, diterapkan metode Synthetic Minority Oversampling Technique (SMOTE) yang menyeimbangkan setiap kelas menjadi 10264 citra. Dua model dikembangkan: Baseline CNN dan Transfer Learning menggunakan arsitektur ResNet-18. Evaluasi menunjukkan bahwa Baseline CNN dengan SMOTE mencapai akurasi pengujian 98%, F1-score 96%, dan Cohen’s Kappa 0.9638, sedangkan ResNet-18 memperoleh akurasi 96%, F1-score 96%, dan Kappa 0.9285. Optimasi hyperparameter dilakukan menggunakan Optuna dan Grid Search, dengan Optuna menghasilkan akurasi validasi tertinggi 100%, lebih unggul dibandingkan Grid Search yang mencapai 95,63%. Model terbaik kemudian diintegrasikan ke dalam aplikasi mobile berbasis Machine Learning Ops, memungkinkan diagnosis awal yang cepat, interaktif, dan efisien. Hasil penelitian menunjukkan bahwa pendekatan ini efektif dalam membantu deteksi dini Alzheimer dan berpotensi dikembangkan lebih lanjut dalam sistem kesehatan digital di Indonesia.
=======================================================================================================================================
Alzheimer disease is a form of dementia that causes a progressive decline in memory and cognitive function. In Indonesia, low public awareness—where 86.3% of the population does not understand the term Alzheimer—hinders early ddiagnosis and increases the burden on caregivers. This study aims to develop an intelligent system based on Convolutional Neural Network (CNN) for Alzheimer classification using brain MRI images. The MRI dataset was obtained from Kaggle and consists of four classes: Mild Demented, Moderate Demented, Non Demented, and Very Mild Demented, with a total of 6,400 MRI images. To address class imbalance, the Synthetic Minority Oversampling Technique (SMOTE) was applied, balancing each class to a total of 10,264 images. Two models were developed: a Baseline CNN and a Transfer Learning model using the ResNet-18 architecture. Evaluation showed that the Baseline CNN with SMOTE achieved a test accuracy of 98%, F1-score of 96%, and Cohen’s Kappa of 0.9638, while ResNet-18 achieved 96% accuracy, 96% F1-score, and a Kappa score of 0.9285. Hyperparameter optimization was conducted using both Optuna and Grid Search, with Optuna yielding the highest validation accuracy of 100%, outperforming Grid Search which reached 95.63%. The best-performing model was then integrated into a mobile application using a Machine Learning Ops framework, enabling rapid, interactive, and efficient early diagnosis. The findings indicate that this approach is effective in supporting early detection of Alzheimer’s and holds significant potential for further development in Indonesia’s digital health systems.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	alzheimer, convolutional neural network, citra mri, transfer learning, apliakasi mobile, aplikasi kesehatan.
Subjects:	R Medicine > R Medicine (General) > R858 Deep Learning
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	I Gusti Agung Jaya Hiswara
Date Deposited:	28 Jul 2025 09:44
Last Modified:	28 Jul 2025 09:54
URI:	http://repository.its.ac.id/id/eprint/122276

Actions (login required)

View Item