Search for collections on Undip Repository

SISTEM KLASIFIKASI UNTUK KELAS SISWA BARU TUNAGRAHITA MENGGUNAKAN KOMBINASI ALGORITMA C4.5 DAN PARTICLE SWARM OPTIMIZATION (PSO)

NOVA, Sausan Hidayah and Warsito, Budi and Widodo, Aris Puji (2023) SISTEM KLASIFIKASI UNTUK KELAS SISWA BARU TUNAGRAHITA MENGGUNAKAN KOMBINASI ALGORITMA C4.5 DAN PARTICLE SWARM OPTIMIZATION (PSO). Masters thesis, UNIVERSITAS DIPONEGORO.

[img] Text
Cover.pdf

Download (637kB)
[img] Text
BAB I.pdf

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

Download (320kB)
[img] Text
BAB III.pdf
Restricted to Repository staff only

Download (418kB)
[img] Text
BAB IV.pdf
Restricted to Repository staff only

Download (582kB)
[img] Text
BAB V.pdf

Download (86kB)
[img] Text
Daftar Pustaka.pdf

Download (167kB)
[img] Text
Lampiran.pdf
Restricted to Repository staff only

Download (1MB)

Abstract

Setiap manusia berhak mendapatkan pendidikan, termasuk anak berkebutuhan khusus, seperti anak tunagrahita. Proses penerimaan siswa baru tunagrahita di SLB Negeri 1 Pelaihari dilaksanakan dengan asesmen untuk menempatkan kelas siswa baru tunagrahita berdasarkan klasifikasi tunagrahita ringan dan tunagrahita sedang. Oleh karena karakteristik anak tunagrahita bermacam-macam dan banyak, perlu adanya suatu metode klasifikasi untuk membantu sekolah agar dapat menerima siswa baru tunagrahita sesuai karakteristik yang telah ditentukan. Penelitian ini menerapkan dua model algoritma machine learning, yaitu C4.5, serta kombinasi C4.5 dan PSO. C4.5 merupakan salah satu algoritma klasifikasi terbaik, namun keberadaan data yang tidak relevan dan terlalu banyak atribut membuat model sulit dibaca. Oleh karena itu, PSO sebagai algoritma seleksi fitur dipilih untuk mengatasi kelemahan C4.5 Penelitian ini bertujuan untuk menerapkan kombinasi C4.5 dan PSO untuk meningkatkan hasil akurasi dari klasifikasi, serta menganalisis seleksi fitur dari PSO untuk mengklasifikasi kelas siswa baru tunagrahita. Hasil penelitian menunjukkan bahwa model kombinasi C4.5 dan PSO mampu menyeleksi atribut dari 20 atribut menjadi 6 atribut dengan melakukan percobaan parameter menggunakan populasi 20 dan iterasi 40, sehingga memperoleh tingkat akurasi sebesar 92%, precision kelas C sebesar 100%, precision kelas C1 sebesar 84,61%, recall kelas C sebesar 85,71%, dan recall kelas C1 sebesar 100%.
Kata kunci: sistem klasifikasi, C4.5, PSO, machine learning, tunagrahita

Every human being has the right to education, including children with special needs, such as mental retardation children. The process of admitting new students with mental retardation at SLB Negeri 1 Pelaihari is carried out with an assessment to place new classes of mental retardation students based on the classification of mild mental retardation and moderate mental retardation. Because the characteristics of mental retardation children vary and are many, it is necessary to have a classification method to help schools accept new mental retardation students according to predetermined characteristics. This study applies two machine learning algorithm models, namely C4.5, as well as a combination of C4.5 and PSO. C4.5 is one of the best classification algorithms, but the presence of irrelevant data and too many attributes makes the model difficult to read. Therefore, PSO as a feature selection algorithm was chosen to overcome C4.5's weaknesses. This study aims to apply a combination of C4.5 and PSO to increase the accuracy of classification results, as well as analyze feature selection from PSO to classify a class of newly mental retardation students. The results showed that the C4.5 and PSO combination model was able to select attributes from 20 attributes to 6 attributes by experimenting with parameters using 20 populations and 40 iterations, resulting in an accuracy rate of 92%, a precision value of class C of 100%, a precision value of class C1 of 84.61%, a recall value of class C of 85.71%, and a recall value of class C1 of 100%.
Keywords: classification system, C4.5, PSO, machine learning, mental retardation

Item Type: Thesis (Masters)
Uncontrolled Keywords: sistem klasifikasi, C4.5, PSO, machine learning, tunagrahita
Subjects: Sciences and Mathemathic
Divisions: Postgraduate Program > Master Program in Information System
Depositing User: ekana listianawati
Date Deposited: 23 Nov 2023 07:19
Last Modified: 23 Nov 2023 07:19
URI: https://eprints2.undip.ac.id/id/eprint/18266

Actions (login required)

View Item View Item