Please use this identifier to cite or link to this item:
https://repositori.uma.ac.id/handle/123456789/26033
Title: | Analisis Klasifikasi Tahu yang Mengandung Formalin Menggunakan Model Arsitektur Googlenet |
Other Titles: | Classification Analysis of Tofu Containing Formalin Using the Googlenet Architectural Model |
Authors: | Fadhillah, Ainun |
metadata.dc.contributor.advisor: | Susilawati |
Keywords: | klasifikasi;tahu;formalin;cnn;googlenet;classification;tofu;formaldehyde |
Issue Date: | Aug-2024 |
Publisher: | UNIVERSITAS MEDAN AREA |
Series/Report no.: | NPM;198160034 |
Abstract: | Tahu merupakan produk pangan yang menggunakan bahan dasar kacang kedelai yang diendapkan. Proses pembuatan tahu biasanya membutuhkan modal sedikit yang membuat para pelaku usaha melakukan kecurangan dengan mencampurkan zat berbahaya salah satunya adalah formalin ke dalam tahu. Hal ini berdampak pada Kesehatan. Apabila mengkonsumsi formalin pada dosis rendah dapat menyebabkan gangguan pencernaan disertai dengan muntah-muntah, timbulnya depresi, dan peredaran darah tidak lancar. Oleh sebab itu, diperlukan pendekatan digital agar dapat membantu untuk mengenali tahu tersebut mengandung formalin atau tidak. Analisis yang dilakukan dalam penelitian ini menggunakan model arsitektur GoogLeNet dari CNN. Terdapat 8 model skenario yang diuji (training) dan diperoleh performa model terbaik pada model menggunakan hyperparameter dengan jumlah epoch 30, batch size 32, optimizer SGD, learning rate 0.001 mendapat akurasi 99% pada proses training. Setelah diuji menggunakan data testing dan dievaluasi menggunakan confusion matrix dan classification report diperoleh nilai accuracy sebesar 99%, precision 100%, recall 99%, dan f1-score 99%. Tofu is a food product made from precipitated soybeans. The tofu-making process generally requires low capital, which leads some business practitioners to commit fraud by adding harmful substances, such as formaldehyde, to tofu. This has a detrimental effect on health. Consuming formaldehyde at low doses can cause digestive issues accompanied by vomiting, depression, and poor blood circulation. Therefore, a digital approach is needed to help identify whether tofu contains formaldehyde. The analysis conducted in this research used the GoogleNet architecture model from CNN. Eight model scenarios were tested (training), and the best model performance was achieved using hyperparameters with 30 epochs, a batch size of 32, the SGD optimizer, and a learning rate of 0.001, achieving 99% accuracy in the training process. After testing with test data and evaluating using a confusion matrix and classification report, accuracy was 99%, precision was 100%, recall was 99%, and F1-score was 99%. |
Description: | 70 Halaman |
URI: | https://repositori.uma.ac.id/handle/123456789/26033 |
Appears in Collections: | SP - Informatic Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
198160034 - Ainun Fadhillah - Fulltext.pdf | Cover, Abstract, Chapter I, II, III, V, Bibliography | 2.16 MB | Adobe PDF | View/Open |
198160034 - Ainun Fadhillah - Chapter IV.pdf Restricted Access | Chapter IV | 681.98 kB | Adobe PDF | View/Open Request a copy |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.