{"title":"IMPLEMENTASI KLASIFIKASI SENJATA TRADISIONAL JAWA BARAT MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK (CNN) DENGAN METODE TRANSFER LEARNING","authors":"Ryas Rafi Karim","doi":"10.23960/jitet.v12i2.4166","DOIUrl":null,"url":null,"abstract":"Studi ini membahas implementasi Convolutional Neural Network (CNN) dengan metode Transfer Learning untuk klasifikasi senjata tradisional Jawa Barat menggunakan EfficientNetB0. Dataset terbatas yang terdiri dari 754 gambar senjata tradisional dibagi menjadi train, validation, dan test set. Pengujian dilakukan dengan 75 epoch dan batch size 16. Hasil menunjukkan peningkatan konsisten dalam performa model, meskipun akurasi validasi tidak melampaui 0.9794. Evaluasi model mencapai akurasi 98,44%, dengan nilai loss rendah. Laporan klasifikasi menunjukkan kekurangan pada kelas gacok, sementara kelas lainnya seperti arit, bedog, dan kujang memiliki performa baik. Meskipun keterbatasan dataset, model ini berhasil dan dapat menjadi dasar untuk penelitian lebih lanjut dalam pengenalan senjata tradisional Jawa Barat.This study discusses the implementation of Convolutional Neural Network (CNN) with the Transfer Learning method for classifying traditional weapons in West Java using EfficientNetB0. A limited dataset consisting of 754 images of traditional weapons is divided into train, validation, and test sets. Testing was carried out with 75 epochs and batch size 16. Results showed consistent improvements in model performance, although validation accuracy did not exceed 0.9794. The model evaluation achieved an accuracy of 98.44%, with a low loss value. The classification report shows deficiencies in the gacok class, while other classes such as sickles, bedogs and cleavers have good performance. Despite the limitations of the dataset, this model is successful and can be a basis for further research in the introduction of West Javanese traditional weapons.","PeriodicalId":313205,"journal":{"name":"Jurnal Informatika dan Teknik Elektro Terapan","volume":"32 19","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Jurnal Informatika dan Teknik Elektro Terapan","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.23960/jitet.v12i2.4166","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Studi ini membahas implementasi Convolutional Neural Network (CNN) dengan metode Transfer Learning untuk klasifikasi senjata tradisional Jawa Barat menggunakan EfficientNetB0. Dataset terbatas yang terdiri dari 754 gambar senjata tradisional dibagi menjadi train, validation, dan test set. Pengujian dilakukan dengan 75 epoch dan batch size 16. Hasil menunjukkan peningkatan konsisten dalam performa model, meskipun akurasi validasi tidak melampaui 0.9794. Evaluasi model mencapai akurasi 98,44%, dengan nilai loss rendah. Laporan klasifikasi menunjukkan kekurangan pada kelas gacok, sementara kelas lainnya seperti arit, bedog, dan kujang memiliki performa baik. Meskipun keterbatasan dataset, model ini berhasil dan dapat menjadi dasar untuk penelitian lebih lanjut dalam pengenalan senjata tradisional Jawa Barat.This study discusses the implementation of Convolutional Neural Network (CNN) with the Transfer Learning method for classifying traditional weapons in West Java using EfficientNetB0. A limited dataset consisting of 754 images of traditional weapons is divided into train, validation, and test sets. Testing was carried out with 75 epochs and batch size 16. Results showed consistent improvements in model performance, although validation accuracy did not exceed 0.9794. The model evaluation achieved an accuracy of 98.44%, with a low loss value. The classification report shows deficiencies in the gacok class, while other classes such as sickles, bedogs and cleavers have good performance. Despite the limitations of the dataset, this model is successful and can be a basis for further research in the introduction of West Javanese traditional weapons.