KLASIFIKASI SPESIES BERDASARKAN DNA BARCODE SEQUENCE MENGGUNAKAN RANDOM FERNS

M Ammar Fadhlur Rahman, - (2022) KLASIFIKASI SPESIES BERDASARKAN DNA BARCODE SEQUENCE MENGGUNAKAN RANDOM FERNS. S1 thesis, Universitas Pendidikan Indonesia.

[img] Text
S_KOM_1507506_Title.pdf

Download (676kB)
[img] Text
S_KOM_1507506_Chapter1.pdf

Download (123kB)
[img] Text
S_KOM_1507506_Chapter2.pdf
Restricted to Staf Perpustakaan

Download (1MB) | Request a copy
[img] Text
S_KOM_1507506_Chapter3.pdf

Download (137kB)
[img] Text
S_KOM_1507506_Chapter4.pdf
Restricted to Staf Perpustakaan

Download (713kB) | Request a copy
[img] Text
S_KOM_1507506_Chapter5.pdf

Download (51kB)
[img] Text
S_KOM_1507506_Appendix.pdf
Restricted to Staf Perpustakaan

Download (363kB) | Request a copy
Official URL: http://repository.upi.edu/

Abstract

Machine learning telah diterapkan dalam berbagai domain, termasuk bioinformatika. Salah satu persoalan bioinformatika yang dapat diselesaikan dengan pendekatan machine learning adalah klasifikasi spesies. Penelitian ini berupaya mengklasifikasikan spesies ke dalam famili berdasarkan sekuens DNA barcode menggunakan pendekatan supervised learning dengan algoritma Random Ferns. Digunakan model komputasi dengan 13 tahapan, termasuk pengunduhan data, rangkaian praproses data, model training, prediksi, dan evaluasi. Gen ribulose-1,5-bisphosphate carboxylase-oxygenase large sub-unit (rbcL) yang merupakan salah satu lokus DNA barcode untuk tanaman, digunakan untuk merepresentasikan spesies dalam famili Amarilis dan Lili. Berdasarkan hasil eksperimen dengan 1.245 sekuens DNA training dan 220 sekuens testing menunjukkan bahwa Random Ferns dapat digunakan untuk mengklasifikasikan spesies ke dalam famili yang sesuai secara cepat dan akurat. Tercapai tingkat akurasi yang konsisten hingga 99,09% dalam waktu training selama 180ms dengan hanya menggunakan memori sebanyak 14,5MB. Perbandingan dengan algoritma Random Forest yang menjadi state-of-the-art menunjukkan Random Ferns dapat mencapai tingkat akurasi yang sepadan secara lebih efisien. Machine learning has been applied in various domains, including bioinformatics. One of the bioinformatics problems that can be solved by using a machine learning approach is species classification. This study attempts to classify species into families based on their DNA barcode sequences using supervised learning approach, i.e., the Random Ferns algorithm. A computational model consisting of 13 steps was proposed, including data retrieval, a series of data preprocessing, model training, prediction, and evaluation. The ribulose-1,5-bisphosphate carboxylase-oxygenase large sub-unit (rbcL) gene that has been selected as one of the DNA barcode loci for plants is used to represent species in the Amaryllidaceae and Liliaceae families. By using 1,245 DNA sequences for training and 220 sequences as testing data, the experiment results show that Random Ferns can be used to classify species sequences quickly and accurately into appropriate families based on their DNA barcode sequences. The trained model could achieve persistent accuracy result as high as 99,09% within 180ms of training time and using only 14,5 MB of memory. A comparison against the state-of-the-art Random Forest algorithm showed Random Ferns was able to achieve the same level of accuracy more efficiently.

Item Type: Thesis (S1)
Additional Information: ID SINTA Dosen Pembimbing: Lala Septem Riza: 5975668 Herbert Siregar: 5991008
Uncontrolled Keywords: machine learning, random ferns, klasifikasi supervised, klasifikasi spesies, DNA barcode, gen rbcL, analisa data, bioinformatika
Subjects: L Education > L Education (General)
Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Q Science > QH Natural history > QH301 Biology
Q Science > QH Natural history > QH426 Genetics
Divisions: Fakultas Pendidikan Matematika dan Ilmu Pengetahuan Alam > Program Studi Ilmu Komputer
Depositing User: M Ammar Fadhlur Rahman
Date Deposited: 15 Sep 2022 08:15
Last Modified: 15 Sep 2022 08:15
URI: http://repository.upi.edu/id/eprint/80431

Actions (login required)

View Item View Item