Penerapan Algoritma Convolutional Neural Network dan Random Forest Untuk Klasifikasi Huruf Aksara Jawa

Purwanto, Novan Windi Eko (2024) Penerapan Algoritma Convolutional Neural Network dan Random Forest Untuk Klasifikasi Huruf Aksara Jawa. S1 thesis, Universitas PGRI Madiun.

[img] Text
Halaman Depan.pdf

Download (525kB)
[img] Text
Abstrak.pdf

Download (67kB)
[img] Text
BAB I.pdf

Download (107kB)
[img] Text
BAB II.pdf

Download (436kB)
[img] Text
BAB III.pdf
Restricted to Repository staff only

Download (123kB)
[img] Text
BAB IV.pdf
Restricted to Repository staff only

Download (1MB)
[img] Text
BAB V.pdf

Download (51kB)
[img] Text
Daftar Pustaka.pdf

Download (188kB)
[img] Text
Lampiran.pdf

Download (1MB)

Abstract

Penelitian ini bertujuan untuk mengaplikasikan Algoritma Convolutional Neural Network (CNN) dan Random Forest (RF) pada sistem klasifikasi huruf aksara jawa. Sistem ini dibangun melalui sebuah aplikasi berbasis website yang menggunakan bahasa pemrograman Phyton, HTML, CSS, serta framework Flask. Selain itu, penelitian ini menggunakan metode pengembangan perangkat lunak Agile Development. Metode ini memungkinkan tim pengembang untuk merespon perubahan kebutuhan dengan cepat dan fleksibel. Metode ini terdiri dari beberapa siklus pengembangan, yang masing-masing terdiri dari perencanaan, analisis, desain, implementasi, dan evaluasi. Evaluasi dilakukan dengan membandingkan tingkat akurasi, precision, recall, dan fi-score antara model CNN dan RF. Hasil penelitian menunjukan bahwa model CNN memiliki tingkat akurasi yang sangat tinggi, mencapai 99,4% pada tes akurasi, namun pada classification report sebesar 0,07% dengan nilai precision, recall, dan fi-score mencapai 07%. Sementara itu, model RF memiliki tingkat akurasi yang lebih rendah, yaitu 42,4% dengan nilai precision, recall, fi-score sekitar 41%. Secara keseluruhan, penelitian ini berhasil menunjukan bahwa penerapan algoritma CNN pada sistem klasifikasi huruf aksara jawa melalui aplikasi berbasis website memberikan hasil yang sangat mengesankan dalam hal akurasi. Pendekatan Agile Development juga membantu dalam menghadapi perubahan dan tantangan dalam proses pengembangan perangkat lunak. Hasil penelitian ini dapat memberikan kontribusi penting dalam pengembangan lebih lanjut dalam bidang pengenalan karakter dan pengenalan tulisan tangan untuk huruf-huruf aksara jawa. Kata kunci: Convolutional Neural Network, Random Forest, Huruf Aksara Jawa This research aims to apply the Convolutional Neural Network (CNN) and Random Forest (RF) algorithms to the Javanese script letter classification system. This system was built through a website-based application that uses the Python programming language, HTML, CSS, and the Flask framework. In addition, this research uses the Agile Development software development method. This method allows the development team to respond to changing needs quickly and flexibly. This method consists of several development cycles, each of which consists of planning, analysis, design, implementation and evaluation. Evaluation is carried out by comparing the level of accuracy, precision, recall and fi-score between the CNN and RF models. The research results show that the CNN model has a very high level of accuracy, reaching 99.4% in the accuracy test, but in the classification report it is 0.07% with precision, recall and fi-score values reaching 07%. Meanwhile, the RF model has a lower level of accuracy, namely 42.4% with precision, recall, fi-score values of around 41%. Overall, this research succeeded in showing that the application of the CNN algorithm to the Javanese script classification system through a website-based application gave very impressive results in terms of accuracy. The Agile Development approach also helps in dealing with changes and challenges in the software development process. The results of this research can provide an important contribution to further development in the field of character recognition and handwriting recognition for Javanese script letters. Keywords: Convolutional Neural Network, Random Forest, Javanese Script

Item Type: Thesis/Skripsi/Tugas Akhir (S1)
Kata Kunci: Convolutional Neural Network; Random Forest; Huruf Aksara Jawa
Subjects: L Education > LB Theory and practice of education
L Education > LB Theory and practice of education > LB1603 Secondary Education. High schools
T Technology > T Technology (General)
Divisions: Fakultas Teknik > Teknik Informatika
Depositing User: EKO WINDI NOVAN
Date Deposited: 22 Aug 2024 03:21
Last Modified: 22 Aug 2024 03:21
URI: http://eprint.unipma.ac.id/id/eprint/1721

Actions (login required)

View Item View Item