DETEKSI KONTAMINAN PADA FORMULASI CAMPURAN METARHIZIUM SP DAN BAHAN ORGANIK DENGAN ELECTRONIC NOSE MENGGUNAKAN MACHINE LEARNING

INDRIANI LUTFIYYATUNNISA, . (2024) DETEKSI KONTAMINAN PADA FORMULASI CAMPURAN METARHIZIUM SP DAN BAHAN ORGANIK DENGAN ELECTRONIC NOSE MENGGUNAKAN MACHINE LEARNING. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text 1. COVER_compressed.pdf Download (735kB)
	Text 2. BAB I.pdf Download (286kB)
	Text 3. BAB II.pdf Restricted to Repository staff only Download (717kB) | Request a copy
	Text 4. BAB III_compressed.pdf Restricted to Repository staff only Download (799kB) | Request a copy
	Text 5. BAB IV_compressed.pdf Restricted to Repository staff only Download (747kB) | Request a copy
	Text 6. BAB V.pdf Restricted to Repository staff only Download (123kB) | Request a copy
	Text 7. DAFTAR PUSTAKA.pdf Download (270kB)
	Text 8. LAMPIRAN DAN DAFTAR RIWAYAT HIDUP.pdf Restricted to Repository staff only Download (1MB) | Request a copy

Abstract

Penggunaan campuran agensia hayati dan limbah bahan organik dapat dijadikan alternatif dari pupuk kimia sebagai produk bioinsektisida yang ramah lingkungan. Namun, keberadaan organisme agensia hayati dalam bahan organik dan rentan kontaminasi pada komposisi tertentu sulit dideteksi secara dini. Pada penelitian ini Electronic nose (E-nose) digunakan untuk deteksi kontaminan pada formulasi campuran Metarhizium anisopliae dan bahan organik berupa limbah ampas tebu berdasarkan pola aromanya. E-nose yang digunakan versi II dengan jumlah sensor sebanyak 16 sensor tipe Metal Oxide Semiconductor (MOS). Fitur diekstrak dari data respon e-nose yang diambil dari 50 sampel campuran jamur agensia hayati dengan bahan organik yang terdiri dari empat variasi perlakuan. Penelitian dilakukan dengan empat sampel dengan perlakuan bahan organik saja, bahan organik dengan M. anisopliae, bahan organik dengan M. anisopliae dan Aspergilus niger serta bahan organik dengan M. anisopliae dan Trichoderma harzianum. Total terdapat 50 sampel dengan lima kali pengulangan dengan waktu pengambilan data yang berbeda. Untuk analisis komponen utama data menggunakan Principal Component Analysis (PCA) dengan fitur statistik deskriptif nilai maksimum, minimum, median, rata-rata, dan standar deviasi. Model Machine learning yang digunakan adalah Support Vector Machine (SVM) dengan tiga kernel yang berbeda dan Random Forest (RF). Akurasi dari variasi data ekstraksi fitur dan model ini untuk data uji dan data tes masing-masing menghasilkan 100% dan 100%. Kata-kata kunci: agensia hayati, electronic nose, ekstraksi fitur, machine learning ***** The use of a mixture of biological agents and organic matter waste can be used as an alternative to chemical fertilizers as an environmentally friendly bioinsecticide product. However, the presence of biological agent organisms in organic materials and susceptible to contamination in certain compositions is difficult to detect early. In this study, Electronic nose (E-nose) was used for contaminant detection in a mixed formulation of Metarhizium anisopliae and organic material in the form of bagasse waste based on its aroma pattern. The E-nose used is version II with a total of 16 Metal Oxide Semiconductor (MOS) type sensors. Features were extracted from e-nose response data taken from 50 samples of a mixture of biological agent fungi with organic materials consisting of four treatment variations. The research was conducted with four samples treated with organic matter alone, organic matter with M. anisopliae, organic matter with M. anisopliae and Aspergilus niger and organic matter with M. anisopliae and Trichoderma harzianum. There were a total of 50 samples with five repetitions with different data collection times. For data principal component analysis using Principal Component Analysis (PCA) with descriptive statistical features of maximum, minimum, median, mean, and standard deviation values. Machine learning models used are Support Vector Machine (SVM) with three different kernels and Random Forest (RF). The accuracy of these extraction feature and model variations for test data and test data resulted in 100% and 100% respectively. Keywords: biological agent, electronic nose, feature extraction, machine learning

Item Type:	Thesis (Sarjana)
Additional Information:	1). Dr. rer. nat Bambang Heru Iswanto, M.Si. ; 2). Agustin Sri Mulyatni, S.P., M.P.
Subjects:	Sains > Sains, Ilmu Pengetahuan Alam Sains > Matematika > Software, Sistem Informasi Komputer Sains > Fisika Sains > Botani Sains > Mikro Biologi > Ekologi
Divisions:	FMIPA > S1 Fisika
Depositing User:	Users 24710 not found.
Date Deposited:	16 Aug 2024 04:58
Last Modified:	16 Aug 2024 04:58
URI:	http://repository.unj.ac.id/id/eprint/49856

Actions (login required)

View Item