Sistem Pendeteksi Kualitas Tahu Menggunakan Electronic Nose Dan Machine Learning

Benedict, Christopher (2022) Sistem Pendeteksi Kualitas Tahu Menggunakan Electronic Nose Dan Machine Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 05311840000024-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Tahu merupakan makanan sehat yang baik untuk dikonsumsi oleh masyarakat. Akan tetapi jika terdapat kandungan formalin dapat menimbulkan berbagai penyakit. Formalin dapat terletak di dalam sesuatu bahan makanan berguna untuk mengawetkan suatu makanan tersebut. Pembuatan electronic nose untuk mengetahui tingkatan formalin dalam sesuatu makanan bisa digunakan untuk mengetahui isi formalin yang terletak di dalam suatu makanan. Hasil Penelitian terdahulu oleh Thazin pada tahun 2019 dengan judul Formalin Adulteration Detection in Food Using E- menjelaskan cara E-nose based on Nanocomposite Gas Sensors untuk mengembangkan sistem electronic nose untuk mendeteksi formaldehida yang terdapat pada tahu dengan PCA. Proses dilakukan dengan mengambil data diambil dari tahu yang direndam dalam formalin selama 1 jam dengan 6 tingkatan yang berbeda [0%, 5%, 10%, 15%, 20%,30%]. Selanjutnya terdapat 6 jenis sensor MQ (MQ-2, MQ-3, MQ-7, MQ-8, MQ-135, MQ-138) untuk mendeteksi gas terhadap kadar formalin pada tahu. Data yang didapatkan akan dipre-proses untuk bagian sedot objek (P2). Lalu akan dievaluasi dengan 3 tipe machine learning model ialah Decision Tree, K- Nearest Neighbors (K- NN) serta Support Vector Machine (SVM). Evaluasi dari 3 model machine learning tersebut belajar dari informasi yang sudah dibersihkan yang setelah itu dilakukan komparasi untuk memastikan model terbaik didalam klasifikasi formalin pada tahu. Hasil dari pengujian adalah K- Nearest Neighbors merupakan model machine learning terbaik yang memiliki akurasi paling tinggi yaitu sebesar 97,0%.
====================================================================================================================================
Tofu is a healthy food that is good for consumption by the community. However, if there is formalin content, it can cause various diseases. Formalin can be located in a food ingredient that is useful for preserving a food. Making an electronic nose to determine the level of formalin in a food can be used to determine the content of formaldehyde in a food. The results of previous research by Thazin in 2019 with the title "Formalin Adulteration Detection in Food Using E-nose based on Nanocomposite Gas Sensors" explained how to develop an electronic nose system to detect formaldehyde contained in tofu with PCA. The process is carried out by taking data taken from tofu soaked in formalin for 1 hour with 6 different levels [0%, 5%, 10%, 15%, 20%, 30%]. Furthermore, there are 6 types of MQ sensors (MQ-2, MQ-3, MQ-7, MQ-8, MQ-135, MQ-138) to detect gas and formaldehyde levels in tofu. The data obtained will be pre-processed for the suction object section (P2). Then it will be evaluated with 3 types of machine learning models, namely Decision Tree, K-Nearest Neighbors (K-NN) and Support Vector Machine (SVM). The evaluation of the 3 machine learning models learned from the cleaned information which was then compared to ensure the best model in the formalin classification of tofu. The result of the test is that K-Nearest Neighbors is the best machine learning model that has the highest accuracy, which is 97.0%.

Item Type:	Thesis (Other)
Additional Information:	RSTI 629.89 Ben s-1 2022
Uncontrolled Keywords:	Tahu , Formalin, Electronic Nose, Machine Learning, Tofu , Formalin , Electronic Nose, Machine Learning
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7871.674 Detectors. Sensors
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	22 Apr 2026 07:14
Last Modified:	22 Apr 2026 07:14
URI:	http://repository.its.ac.id/id/eprint/132869

Actions (login required)

View Item