OPTIMASI E-NOSE BERBASIS SENSOR GAS MQ DENGAN VARIASI TEMPERATUR UNTUK KLASIFIKASI KOPI

MUHAMMAD ROSYID SUSENO, . (2025) OPTIMASI E-NOSE BERBASIS SENSOR GAS MQ DENGAN VARIASI TEMPERATUR UNTUK KLASIFIKASI KOPI. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text COVER.pdf Download (781kB)
	Text BAB 1.pdf Download (274kB)
	Text BAB 2.pdf Restricted to Registered users only Download (777kB) | Request a copy
	Text BAB 3.pdf Restricted to Registered users only Download (868kB) | Request a copy
	Text BAB 4.pdf Restricted to Registered users only Download (2MB) | Request a copy
	Text BAB 5.pdf Restricted to Registered users only Download (261kB) | Request a copy
	Text Daftar Pustaka.pdf Download (266kB)
	Text Lampiran.pdf Restricted to Registered users only Download (293kB) | Request a copy

Abstract

Penelitian ini bertujuan untuk merancang dan mengoptimalkan sistem Electronic Nose (E-Nose) berbasis larik sensor gas MQ untuk mengklasifikasikan aroma bubuk kopi Arabika dan Robusta. Sistem E-Nose dirancang dengan ruang sampel berpemanas yang dikendalikan menggunakan algoritma PID, memungkinkan pengaturan temperatur yang stabil pada tiga titik suhu (35 °C, 40 °C, dan 45 °C). Data respons sensor diperoleh dari sembilan sensor MQ, lalu diekstraksi menjadi fitur statistik berupa nilai maksimum (max), rata-rata (mean), dan Area Under Curve (AUC), menghasilkan total 27 fitur per sampel. Fitur-fitur ini digunakan untuk melatih model klasifikasi Support Vector Machine (SVM) dengan kernel linear dan Radial Basis Function (RBF). Proses tuning hyperparameter dilakukan menggunakan Grid Search, sedangkan evaluasi performa model dilakukan dengan metode Leave-One-Out Cross Validation (LOOCV). Hasil evaluasi menunjukkan bahwa performa terbaik dicapai pada suhu 45 °C dengan kernel linear menghasilkan akurasi 97% dan AUC 1,00, serta kernel RBF mencapai akurasi 97,5% dan AUC 0,9925. Temuan ini membuktikan bahwa kombinasi sensor MQ dengan sistem pemanas dan pendekatan klasifikasi SVM dapat digunakan secara efektif untuk membedakan jenis kopi berdasarkan aroma, dengan akurasi tinggi dan konfigurasi perangkat yang relatif sederhana. *********************************************************** This research aims to design and optimize an Electronic Nose (E-Nose) system based on a MQ gas sensor array for classifying the aroma of Arabica and Robusta ground coffee. The E-Nose system was designed with a heated sample chamber controlled by a PID algorithm, allowing stable temperature settings at three levels (35 °C, 40 °C, and 45 °C). Sensor response data were collected from nine MQ sensors and then extracted into statistical features, including maximum value, mean, and Area Under the Curve (AUC), resulting in a total of 27 features per sample. These features were used to train a Support Vector Machine (SVM) classification model using linear and Radial Basis Function (RBF) kernels. Hyperparameter tuning was conducted using Grid Search, and model performance was evaluated using Leave-One-Out Cross Validation (LOOCV). Evaluation results showed that the best performance was achieved at 45 °C, where the linear kernel yielded 97% accuracy and an AUC of 1.00, while the RBF kernel achieved 97.5% accuracy and an AUC of 0.9925. These findings demonstrate that the combination of MQ sensors with a heating system and the SVM classification approach can be effectively used to distinguish coffee types based on aroma, with high accuracy and a relatively simple device configuration.

Item Type:	Thesis (Sarjana)
Additional Information:	1). Dr. rer. nat. Bambang Heru Iswanto, M.Si. ; 2). Haris Suhendar, M.Sc.
Subjects:	Sains > Sains, Ilmu Pengetahuan Alam Sains > Fisika Teknologi dan Ilmu Terapan > Teknologi (umum)
Divisions:	FMIPA > S1 Fisika
Depositing User:	Muhammad Rosyid Suseno .
Date Deposited:	14 Aug 2025 04:01
Last Modified:	14 Aug 2025 04:01
URI:	http://repository.unj.ac.id/id/eprint/60786

Actions (login required)

View Item