Prosiding Seminar Nasional Ilmu Komputer Universitas Semarang 2021
Optimasi Hyperparameter Pada Convolutional Neural Network Menggunakan Hyperband Pada Klasifikasi Penyakit Daun Tomat
Dublin Core
Title
Prosiding Seminar Nasional Ilmu Komputer Universitas Semarang 2021
Optimasi Hyperparameter Pada Convolutional Neural Network Menggunakan Hyperband Pada Klasifikasi Penyakit Daun Tomat
Optimasi Hyperparameter Pada Convolutional Neural Network Menggunakan Hyperband Pada Klasifikasi Penyakit Daun Tomat
Subject
convolutional neural network, hyperband, deep learning, hyperparameter, image classification.
Description
Computer vision merupakan salah domain dari machine learning yang memungkinkan komputer dan sistem mengambil tindakan atau membuat rekomendasi berdasarkan informasi yang diperoleh dari citra digital, video, dan input visual lainnya. Dewasa ini, algoritma deep learning mulai diterapkan pada aplikasi computer vision, salah satunya yaitu Convolutional Neural Network (CNN). Convolutional Neural Network (CNN) merupakan salah satu algoritma terbaik untuk memahami konten citra dan telah menunjukkan performa yang baik dalam segmentasi, klasifikasi, deteksi, dan tugas lainnya yang berkaitan dengan citra. Performa Convolutional Neural Network (CNN) memberikan peluang dan memperluas penelitian ke bidang pertanian, salah satunya mengidentifikasi penyakit tanaman tomat. Salah satu faktor yang mempengaruhi performa Convolutional Neural Network (CNN) adalah pemilihan hyperparameter yang sesuai. Pemilihan hyperparameter yang sesuai membutuhkan waktu dan pengalaman yang cukup untuk memperoleh hasil yang maksimal. Sehingga tujuan pada penelitian ini adalah mencari nilai hyperparameter dari Convolutional Neural Network (CNN) agar memperoleh hasil yang lebih baik. Penelitian ini melakukan klasifikasi citra penyakit daun tomat berdasarkan dataset yang berjumlah 11000 citra dengan 10 kelas. Metode yang digunakan adalah Convolutional Neural Network (CNN) dan Hyperband sebagai optimasi hyperparameter. Harapannya dengan dilakukan optimasi hyperparameter pada arsitektur Convolutional Neural Network (CNN) dapat memberikan hasil prediksi yang optimal.
Computer vision is a domain of machine learning that allows computers and systems to take action or make recommendations based on information obtained from digital images, videos, and other visual inputs. Today, deep learning algorithms are starting to be applied to computer vision applications, one of which is the Convolutional Neural Network (CNN). Convolutional Neural Network (CNN) is one of the best algorithms for understanding image content and has shown good performance in segmentation, classification, detection, and other tasks related to images. Convolutional Neural Network (CNN) performance provides opportunities and expands research into agriculture, one of which is identifying tomato plant diseases. One of the factors that affect the performance of Convolutional Neural Network (CNN) is the selection of appropriate hyperparameters. Selection of the appropriate hyperparameters requires sufficient time and experience to obtain maximum results. So the purpose of this study is to find the hyperparameter value of the Convolutional Neural Network (CNN) in order to obtain better results. This study performs image classification of tomato leaf disease based on a dataset of 11000 images with 10 classes. The method used is Convolutional Neural Network (CNN) and Hyperband as hyperparameter optimization. It is hoped that by optimizing hyperparameters on the Convolutional Neural Network (CNN) architecture, it can provide optimal prediction results.
Computer vision is a domain of machine learning that allows computers and systems to take action or make recommendations based on information obtained from digital images, videos, and other visual inputs. Today, deep learning algorithms are starting to be applied to computer vision applications, one of which is the Convolutional Neural Network (CNN). Convolutional Neural Network (CNN) is one of the best algorithms for understanding image content and has shown good performance in segmentation, classification, detection, and other tasks related to images. Convolutional Neural Network (CNN) performance provides opportunities and expands research into agriculture, one of which is identifying tomato plant diseases. One of the factors that affect the performance of Convolutional Neural Network (CNN) is the selection of appropriate hyperparameters. Selection of the appropriate hyperparameters requires sufficient time and experience to obtain maximum results. So the purpose of this study is to find the hyperparameter value of the Convolutional Neural Network (CNN) in order to obtain better results. This study performs image classification of tomato leaf disease based on a dataset of 11000 images with 10 classes. The method used is Convolutional Neural Network (CNN) and Hyperband as hyperparameter optimization. It is hoped that by optimizing hyperparameters on the Convolutional Neural Network (CNN) architecture, it can provide optimal prediction results.
Creator
Ardiansyah Kamal Alkaff, Budi Prasetiyo
Publisher
Universitas Semarang
Date
13 Oktober 2021
Contributor
Sri Wahyuni
Rights
ISSN: 2614-1205
Format
PDF
Language
Indonesian
Type
Text
Coverage
Prosiding Seminar Nasional Ilmu Komputer Universitas Semarang 2021
Files
Citation
Ardiansyah Kamal Alkaff, Budi Prasetiyo, “Prosiding Seminar Nasional Ilmu Komputer Universitas Semarang 2021
Optimasi Hyperparameter Pada Convolutional Neural Network Menggunakan Hyperband Pada Klasifikasi Penyakit Daun Tomat,” Repository Horizon University Indonesia, accessed April 4, 2025, https://repository.horizon.ac.id/items/show/3486.
Optimasi Hyperparameter Pada Convolutional Neural Network Menggunakan Hyperband Pada Klasifikasi Penyakit Daun Tomat,” Repository Horizon University Indonesia, accessed April 4, 2025, https://repository.horizon.ac.id/items/show/3486.