Magfira, Dike Bayu (2019) Pengklasifikasian Aroma Kopi Arabika Luwak Dan Non Luwak Dengan Elektronik Nose. Masters thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
09211750053009-Master_Thesis.pdf Download (5MB) | Preview |
Abstract
Beberapa penelitian Electronic Nose (E-nose) tentang klasifikasi kopi telah dilakukan. E-nose menggunakan sensor gas untuk mendeteksi aroma kopi dan menghasilkan sinyal. Kemudian, sinyal diklasifikasikan menggunakan algoritma pembelajaran mesin. Dalam penelitian ini, E-nose menggunakan 5 sensor gas untuk mengklasifikasikan kopi luwak dan non-luwak, dan algoritma pembelajaran mesin yang digunakan adalah SVM, KNN dan Decision Tree. Varian kopi yang digunakan kopi Arabika dengan jenis Kopi luwak dan non-luwak berasal dari Aceh, Arjuno Malang, Bengkulu. Dalam penelitian ini juga dilakukan pencampuran kopi luwak: non-luwak dengan persentase 100: 0, 90:10, 10:90, 80:20, 20:80, 75:25, 25:75, 50:50. Akurasi klasifikasi dari kopi luwak Aceh (LA) dengan kopi non-luwak Aceh (NLA): 90% (SVM), 100% (KNN), 100% (Decision Tree). Akurasi klasifikasi dari kopi luwak Arjuno (LAR) dengan non-luwak Arjuno (NLAR): 100% (SVM, KNN, Decision Tree). Akurasi klasifikasi kopi luwak Bengkulu (LB) dengan non-luwak Bengkulu (NLB): 45% (SVM), 100% (KNN, Decision Tree). Akurasi klasifikasi kopi mixture (luwak aceh dengan non luwak aceh): 90% (SVM), 93,75% (KNN), 95% (Decision Tree). Tingkat akurasi yang dihasilkan dipengaruhi suhu ruangan, proses saat pengambilan data, keadaan dan usia kopi, dan jumlah atribut kelas yang digunakan.
==================================================================================================================================
Several Electronic Nose (E-nose) studies of coffee classification have been conducted. E-nose uses a gas sensor to detect the aroma of coffee and produce a signal. Then, the signal is classified using a machine learning algorithm. In this study, E-nose used 5 gas sensors to classify civet and non-civet coffee, and the machine learning algorithm used was SVM, KNN and Decision Tree. The coffee variant used is Arabica coffee with the type of civet coffee and non-civet originating from Aceh, Arjuno Malang, Bengkulu. In this study also mixed civet: non-civet coffee with a percentage of 100: 0, 90:10, 10:90, 80:20, 20:80, 75:25, 25:75, 50:50. Classification accuracy of civet Aceh (LA) coffee with Aceh non-civet coffee (NLA): 90% (SVM), 100% (KNN), 100% (Decision Tree). Classification accuracy of civet Arjuno (LAR) coffee with non-civet Arjuno (NLAR): 100% (SVM, KNN, Decision Tree). Accuracy of Bengkulu civet (LB) coffee classification with Bengkulu non-civet (NLB): 45% (SVM), 100% (KNN, Decision Tree). Coffee mixture classification accuracy (civet Aceh with non civet Aceh): 90% (SVM), 93.75% (KNN), 95% (Decision Tree). The level of accuracy produced is influenced by room temperature, process when retrieving data, state and age of coffee, and the number of class attributes used.
| Item Type: | Thesis (Masters) |
|---|---|
| Additional Information: | RTMT 005.1 Mag p-1 |
| Uncontrolled Keywords: | E-nose, SVM, KNN, Decision Tree, Luwak Kopi, Non Luwak Kopi |
| Subjects: | Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) |
| Divisions: | Faculty of Business and Management Technology > Management Technology > 61101-(S2) Master Thesis |
| Depositing User: | Dike Bayu Magfira |
| Date Deposited: | 26 Mar 2025 02:51 |
| Last Modified: | 26 Mar 2025 02:51 |
| URI: | http://repository.its.ac.id/id/eprint/67062 |
Actions (login required)
 |
View Item |