Alfarizy, Muhammad Rifal (2021) Klasifikasi Status Mikrosatelit Pada Sel Kanker Gastrointestinal Menggunakan Algoritma Convolutional Neural Networks. Undergraduate (S1) thesis, Universitas Muhammadiyah Malang.

Preview	Text PENDAHULUAN.pdf Download (1MB) | Preview
Preview	Text Bab 1.pdf Download (158kB) | Preview
Preview	Text Bab 2.pdf Download (526kB) | Preview
	Text Bab 3.pdf Restricted to Registered users only Download (372kB) | Request a copy
	Text Bab 4.pdf Restricted to Registered users only Download (1MB) | Request a copy
	Text Bab 5.pdf Restricted to Registered users only Download (209kB) | Request a copy
	Text Poster.pdf Restricted to Registered users only Download (273kB) | Request a copy

Abstract

Microsatelite is a unique type of recurrent genomic circuit that is spread throughout the genome and exhibits a high level of allele polymorphism.The microsatelite state turns unstable due to a mismatch repair deoxyribonucleic acid system error.Identification of microsatelite status can be done using deep learning method, Convolutional Neural Networks (CNN).There have been several studies using CNN in this case, but there are problems with model stability and accuracy resulting in less than the maximum that may be caused by overfitting.Based on the problem, the study proposed CNN method by proposing modification of VGG19 model then adding augmentation and dropout after pooling layer (APL).The data used is an image of colorectal and gastric cancer cells.In colorectal cancer cells, the proposed model obtained an accuracy of 76.2%, the addition of augmentation gained an accuracy of 77.1%, then gave the APL dropout lowering accuracy to 76.3%.While in gastric cancer cells testing was conducted using a proposed model with augmentation and the addition of callbacks obtained accuracy of 76%.The proposed model is still overfitting, but augmentation can slightly address overfitting and improve accuracy.While the APL dropout lowers accuracy but provides a very stable development of accuracy charts and no indication of overfitting.

Item Type:	Thesis (Undergraduate (S1))
Student ID:	201710370311219
Thesis Advisors:	Agus Eko Minarno (0729118203), Yufis Azhar (0728088701)
Keywords:	Microsatellite, Gastrointestinal, Modified VGG19, Dropout APL, CNN
Subjects:	Q Science > Q Science (General) Q Science > QA Mathematics > QA75 Electronic computers. Computer science T Technology > T Technology (General)
Divisions:	Faculty of Engineering > Department of Informatics (55201)
Depositing User:	201710370311219 mrifala29
Date Deposited:	24 Sep 2021 03:04
Last Modified:	24 Sep 2021 03:04
URI :	http://eprints.umm.ac.id/id/eprint/79325

Actions (login required)

View Item