Dwi Saputra, Dede Ikhsan (2023) IMPLEMENTASI ENSEMBLE LEARNING DENGAN ARSITEKTUR MOBILENETV2 DAN EFFICIENTNETB7 PADA KLASIFIKASI CITRA PENYAKIT KANKER KULIT. Undergraduate thesis, UPN Veteran Jawa Timur.
|
Text (Cover)
19081010150-cover.pdf Download (2MB) | Preview |
|
|
Text (Bab 1)
19081010150-bab1.pdf Download (144kB) | Preview |
|
![[img]](http://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Bab 2)
19081010150-bab2.pdf Restricted to Registered users only until 25 July 2025. Download (1MB) |
|
|
Text (Bab 3)
19081010150-bab3.pdf Restricted to Registered users only until 25 July 2025. Download (353kB) |
|
|
Text (Bab 4)
19081010150-bab4.pdf Restricted to Registered users only until 25 July 2025. Download (6MB) |
|
|
Text (Bab 5)
19081010150-bab5.pdf Download (9kB) | Preview |
|
|
Text (Daftar Pustaka)
19081010150-daftarpustaka.pdf Download (233kB) | Preview |
Abstract
In Indonesia, cancer is a serious health problem. Skin cancer ranks third after uterine cancer and breast cancer. Skin cancer is found 5.9 - 7.8% of all types of cancer per year. The most common skin cancer in Indonesia is Basal Cell Carcinoma (65.5%), followed by Melanoma (7.9%) Squamous Cell Carcinoma (23%) and other skin cancers. One of the efforts to make early prevention of skin cancer can be overcome by automation. The solution that can be applied to solve this problem is using image classification with deep learning methods that are applied to trained models. This technology can be applied and utilized as an intelligent image processing system, which is to detect the possibility of someone having skin cancer on someone's skin that has taken disease image samples based on their characteristics, shape, and color. In today's sophistication, deep learning technology is a hot topic of conversation and is increasingly being used because of the latest results obtained, such as image classification, object detection, and natural language processing. As the development of deep learning accelerates, deep learning technology is developed into a Convolutional Neural Network and creates a pre-trained model called transfer learning. In this study using pre-trained MobileNetV2 and EfficientNetB7 architectural models. Then to get more optimal results with this architecture, Ensemble Learning is used using a combination of several algorithms or models to get output with more optimal accuracy. Ensemble Stack Generalization is a type of Ensemble Learning used in this study by calculating the average weight value of each stored model. Based on this explanation, the authors classified skin cancer images using Ensemble Learning with MobileNetV2 and EfficientNetB7 architectures to determine the performance of Ensamble Learning with MobileNetV2 and EfficientNetB7 architectures in skin cancer image classification. The output produced in this study is the accuracy and performance model for skin cancer image classification. The highest val_accuracy results were obtained on the MobileNetV2 architecture through model training and model fine tuning of 97%. Then, the highest val_accuracy obtained on the EfficientNetB7 architecture through model training and fine tuning the model is 99%. Meanwhile, the highest val_accuracy resulting from the combined implementation of Ensemble Learning on MobileNetV2 and EfficientNetB7 is 97% The three results are obtained in a comparison of the distribution of training data, validation data, and test data of 80%:15%:5%.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | T Technology > T Technology (General) | ||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Informatics | ||||||||||||
| Depositing User: | Dede Ikhsan Dwi Saputra | ||||||||||||
| Date Deposited: | 25 Jul 2023 07:54 | ||||||||||||
| Last Modified: | 25 Jul 2023 07:54 | ||||||||||||
| URI: | http://repository.upnjatim.ac.id/id/eprint/16426 |
Actions (login required)
 |
View Item |