Please use this identifier to cite or link to this item: https://repositori.uma.ac.id/handle/123456789/18998
Title: Klasifikasi Jenis Hiou Simalungun Sumatera Utara Menggunakan Algoritma Convolutional Neural Network
Other Titles: Classification of North Sumatra Simalungun Sharks Using the Convolutional Neural Network Algorithm
Authors: Girsang, Nardianti Dewi
Keywords: hiou;image;convolutional neural network
Issue Date: 27-Sep-2022
Publisher: Universitas Medan Area
Series/Report no.: NPM;178160001
Abstract: Indonesia terkenal akan keragaman seni dan budaya. Salah satunya yaitu di daerah Sumatera Utara yang merupakan daerah yang dimana memiliki budaya dan multi etnik. Sampai saat ini, mereka masih tetap menjaga dan memelihara berbagai unsur budaya yang telah diwarisi oleh nenek moyangnya. Salah satunya yaitu masyarakat suku simalungun. Masyarakat suku simalungun ini, sangat menghormati, menghargai dan menjunjung tinggi adat istiadatnya, hal ini disebabkan karena tercerminnya kepribadian yang mengandung norma dan nilai-nilai yang perlu dimiliki oleh setiap masyarakatnya. Di wilayah Simalungun, kain hasil tenunan disebut hiou. Terdapat banyak jenis hiou di Sumatera Utara yang sering digunakan dalam acara adat. Setiap jenis hiou ini memiliki makna, fungsi dan penggunaan yang berbeda-beda. Hampir semua motif hiou mirip antara satu dengan yang lainnya. Banyaknya motif hiou ini, membuat masyarakat sulit dalam mengidentifikasinya. Untuk mengatasi masalah tersebut, maka perlu adanya sistem klasifikasi yang dapat mengenali motif hiou simalungun. Pada penelitian ini, Algoritma CNN digunakan sebagai metode klasifikasi untuk menentukan tingkat akurasi. Model CNN ini menerapkan ukuran gambar 150x150, batch size 128, nilai epoch 10, dan menggunakan 7 optimizer. Data citra yang digunakan untuk proses trainingnya sebanyak 900 gambar, yang dimana jumlah gambar perkelasnya 100. Data training ini dibagi lagi menjadi dua yaitu training dan validasi, dengan jumlah sebanyak 720 data training dan 180 data validasi. Dari hasil penelitian ini, disimpulkan bahwa optimizer Adam menghasilkan tingkat akurasi tertinggi. Dimana tingkat akurasi training 96,11%, nilai loss 0.127 dan nilai rata-rata precision 97%, recall 96%, dan f1-score 96%. Indonesia is famous for its diversity of arts and culture. One of them is in the area of North Sumatra which is an area which has a multi-ethnic culture. Until now, they still maintain and maintain various cultural elements that have been inherited by their ancestors. One of them is the simalungun tribal community. This simalungun tribal community really respects, appreciates and upholds their customs, this is due to the reflection of a personality that contains norms and values that every society needs to have. In the Simalungun area, the woven fabric is called hiou. There are many types of hiou in North Sumatra which are often used in traditional events. Each type of hiou has a different meaning, function and use. Almost all hiou motifs are similar to one another. The many hiou motifs make it difficult for people to identify them. To overcome this problem, it is necessary to have a classification system that can recognize the hibiscus motif in simalungun. In this study, the CNN algorithm is used as a classification method to determine the level of accuracy. This CNN model applies an image size of 150x150, a batch size of 128, an epoch value of 10, and uses 7 optimizers. The image data used for the training process is 900 images, of which the number of images per class is 100. This training data is further divided into two, namely training and validation, with a total of 720 training data and 180 validation data. From the results of this study, it is concluded that the Adam optimizer produces the highest level of accuracy. Where the level of training accuracy is 96.11%, the loss value is 0.127 and the average value is 97% precision, 96% recall, and 96% f1-score.
Description: 88 Halaman
URI: https://repositori.uma.ac.id/handle/123456789/18998
Appears in Collections:SP - Informatic Engineering

Files in This Item:
File Description SizeFormat 
178160001 - Nardianti Dewi Girsang - Fulltext.pdfCover, Abstract, Chapter I, II, III, V, Bibliography2.07 MBAdobe PDFView/Open
178160001 - Nardianti Dewi Girsang - Chapter IV.pdf
  Restricted Access
Chapter IV1.15 MBAdobe PDFView/Open Request a copy


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.