Nasikhin, Yovi Ibnu (2025) Optimasi Model Convolutional Neural Network Dengan Hyperband Untuk Klasifikasi Tuberkulosis Pada Citra X-Ray Dada. Undergraduate thesis, Universitas Pembangunan Nasional "Veteran" Jawa Timur.
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Cover)
21081010128.-cover.pdf Download (763kB) |
|
Text (Bab1)
21081010128.-bab1.pdf Download (251kB) |
|
Text (Bab2)
21081010128.-bab2.pdf Restricted to Repository staff only until 22 January 2029. Download (721kB) |
|
Text (Bab3)
21081010128.-bab3.pdf Restricted to Repository staff only until 22 January 2029. Download (1MB) |
|
Text (Bab4)
21081010128.-bab4.pdf Restricted to Repository staff only until 22 January 2029. Download (1MB) |
|
Text (Bab5)
21081010128.-bab5.pdf Download (241kB) |
|
Text (Daftar Pustaka)
21081010128.-daftarpustaka.pdf Download (184kB) |
|
Text (Lampiran)
21081010128.-lampiran.pdf Restricted to Repository staff only until 22 January 2029. Download (300kB) |
Abstract
Tuberculosis is one of the infectious diseases that remains a global health problem with a high mortality rate. In the diagnostic process, chest X-ray images are often used. However, their interpretation is often influenced by differences in perception among healthcare professionals. Given these limitations, Convolutional Neural Network (CNN) can be a solution for detecting tuberculosis in chest X-ray images, thereby improving efficiency and accuracy. Although CNN has shown promising results in previous studies, significant challenges remain in selecting the appropriate hyperparameters to maximize model performance. Therefore, this study aims to optimize the CNN model using the Hyperband method, an efficient hyperparameter search technique that automatically identifies the best combination. This study will compare the performance of the baseline model with the Hyperband-optimized model and evaluate two dataset splitting scenarios: 80:10:10 and 70:10:20. The dataset used consists of 1173 Normal images and 1025 Tuberculosis images. The results show that the best model was obtained from optimization with a 70:10:20 data split, with a train accuracy of 85.33%, validation accuracy of 83.11%, and test accuracy of 88.43%. These values are higher than the baseline model, which only achieved a train accuracy of 76.75%, a validation accuracy of 79%, and a test accuracy of 83.67%. This improvement demonstrates that optimization using Hyperband can significantly enhance the performance of CNN models. The results of this study are expected to support the automation of tuberculosis diagnosis and contribute to efforts in the management and control of tuberculosis.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | Q Science > Q Science (General) T Technology > T Technology (General) |
||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Informatics | ||||||||||||
| Depositing User: | Yovi Ibnu Nasikhin | ||||||||||||
| Date Deposited: | 22 Jan 2026 07:28 | ||||||||||||
| Last Modified: | 22 Jan 2026 07:28 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/48969 |
Actions (login required)
 |
View Item |