Implementasi Deep Learning Pada Klasifikasi Jenis Beras Impor

Gunawan, Kevin (2023) Implementasi Deep Learning Pada Klasifikasi Jenis Beras Impor. S1 Teknik Informatika thesis, STMIK Widya Cipta Dharma.

	Text 1843095-S1-Jurnal.pdf Restricted to Repository staff only Download (481kB) | Request a copy
	Text 1843095-S1-Teknik Informatika.pdf Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Kevin Gunawan 2023, Implementasi Deep Learning Pada Klasifikasi Jenis Beras Impor. Skripsi Program Studi Teknik Informatika, Sekolah Tinggi Manajemen Informatika dan Komputer Widya Cipta Dharma Samarinda, Pembimbing Utama : Eka Arriyanti, S.Pd., M.Kom., I.G., Pembimbing Pendamping: Wahyuni, S.kom., M.Kom. Kata Kunci : Data Mining, Deep Learning, Transfer Learning, Klasifikasi, Arsitektur VGG-16, Arsitektur ResNet50, Perbandingan. Penelitian ini dilakukan untuk membuat Implementasi Deep Learning Pada Klasifikasi Jenis Beras Impor. Dengan fokus untuk mencari nilai akurasi terbaik diantara kedua arsitektur tersebut. Beras Impor adalah jenis beras yang tidak diproduksi oleh dalam negri, sehingga harus mengambil dari negara lain. Arsitektur VGG-16 dan ResNet50 sendiri merupakan pengembangan arsitektur dalam algoritma Convolutional Neural Network yang dalam pengaplikasian nya jika digunakan maka temasuk kedalam Transfer Learning, karena melatih model dengan arsitektur yang telah ada. Dalam pengembangan model, diperlukan sebuah metode Transfer Learning untuk dapat menghasilkan akurasi model. Peneliti menggunakan metode General Data Mining Steps dalam pembuatan model, dengan tahapan-tahapan perbandingan Defining Data Domain, Collecting The Data, Pre Processing The Data, Estimating The Model dan Interpreting The Result.. Yang kemudian dilakukan perbandingan terhadap nilai akurasi, serta mengevaluasi model dengan confusion matrix. Dataset yang digunakan yaitu dataset yang berasal dari kaggle.com. Hasil penelitian berupa arsitektur VGG-16 memiliki akurasi sebesar 99.08%, dan arsitektur ResNet50 memiliki akurasi sebesar 95.24%. Kedua arsitektur tersebut memiliki nilai yang sangat baik. Namun pada penelitian ini arsitektur VGG-16 dinilai memiliki nilai akurasi yang lebih baik ketimbang arsitektur ResNet50 dalam proses klasifikasi jenis beras impor. ======================================================================================================================== Kevin Gunawan 2023, Implementation of Deep Learning in the Classification of Types of Imported Rice. Thesis Informatics Engineering Study Program, Widya Cipta Dharma Samarinda College of Informatics and Computer Management, Main Supervisor: Eka Arriyanti, S.Pd., M.Kom., I.G., Associate Advisor: Wahyuni, S.kom., M.Kom. Keywords: Deep Learning, Transfer Learning, Classification, VGG-16 Architecture, ResNet50 Architecture, Comparison. This research was conducted to create a Deep Learning Implementation on Classification of Imported Rice Types. With a focus on finding the best accuracy value between the two architectures. Imported rice is a type of rice that is not produced domestically, so it must be imported from other countries. The VGG-16 and ResNet50 architectures themselves are architectural developments in the Convolutional Neural Network algorithm, which in their application, if used, are included in Transfer Learning, because they train models with existing architectures. In model development, a Transfer Learning method is needed to produce model accuracy. The researcher used the General Data Mining Steps method in making the model, with the comparison stages of Defining the Data Domain, Collecting The Data, Pre Processing The Data, Estimating The Model and Interpreting The Result. Then a comparison was made to the accuracy value, and evaluating the model with confusion matrix. The dataset used is a dataset originating from kaggle.com. The results of the research are that the VGG-16 architecture has an accuracy of 99.08%, and the ResNet50 architecture has an accuracy of 95.24%. Both architectures have very good value. However, in this study the VGG-16 architecture was considered to have a better accuracy value than the ResNet50 architecture in the process of classifying imported rice types.

Item Type:	Thesis (S1 Teknik Informatika)
Additional Information:	Pembimbing 1 : Eka Arriyanti, S.Pd., M.Kom., I.G Pembimbing 2 : Wahyuni, S.Kom., M.Kom
Uncontrolled Keywords:	Deep Learning, Transfer Learning, Classification, VGG-16 Architecture, ResNet50 Architecture, Comparison.
Subjects:	Q Science > QA Mathematics Q Science > QA Mathematics > QA76 Computer software
Divisions:	Teknik Informatika
Depositing User:	Gunawan Kevin
Date Deposited:	25 Aug 2023 03:15
Last Modified:	25 Aug 2023 03:15
URI:	http://repository.wicida.ac.id/id/eprint/5052

Actions (login required)

View Item