Klasifikasi Rasa Manis Dan Asam Menggunakan Electroencephalography (EEG)

Astuti, Ulfi Widya (2023) Klasifikasi Rasa Manis Dan Asam Menggunakan Electroencephalography (EEG). Masters thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 6022211041_Master_Thesis.pdf] Text
6022211041_Master_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2025.

Download (4MB) | Request a copy

Abstract

Manusia memerlukan makanan dan minuman untuk nutrisi, dan rasa dari makanan dan minuman tersebut penting dalam proses konsumsinya. Metode konvensional untuk mengevaluasi rasa adalah melalui survei subjektif, tetapi pendekatan ini memiliki keterbatasan. Oleh karena itu, metode yang lebih objektif, Elektroensefalografi (EEG), digunakan untuk mengevaluasi persepsi rasa. Studi ini bertujuan untuk menguji persepsi rasa asam dan manis menggunakan EEG untuk merekam sinyal otak yang dihasilkan oleh stimulus rasa. Eksperimen melibatkan 32 partisipan, dan sinyal EEG direkam pada empat saluran: CP1, CP2, T3, dan T4. Sinyal EEG dibersihkan, diuraikan menjadi sub-band, dan dikenakan ekstraksi fitur menggunakan MATLAB, sedangkan Python digunakan untuk mengklasifikasikan rasa asam dan manis. Mean Absolute Value (MAV), Standard Deviation (STD), dan Power Spectral Density (PSD) adalah fitur yang diekstraksi dari sinyal EEG, dan empat algoritma digunakan dalam studi ini, yaitu K-Nearest Neighbor (K-NN), Support Vector Machine (SVM), Random Forest, dan Gate Recurrent Unit (GRU). Algoritma GRU memberikan hasil terbaik dengan akurasi 91,53% menggunakan tiga fitur yang disebutkan. Untuk setiap fitur, fitur PSD dan algoritma random forest memberikan akurasi tertinggi pada 91,46%.
================================================================================================================================
Humans require food and drink for nutrition, and the taste of these is critical in the process of consumption. The conventional method for evaluating taste is through subjective surveys, but this approach has limitations. Therefore, a more objective method, Electroencephalography (EEG), is used to evaluate taste perception. This study aims to examine the perception of sour and sweet tastes using EEG to record brain signals generated by taste stimuli. The experiment involved 32 participants, and EEG signals were recorded at four channels: CP1, CP2, T3, and T4. The EEG signals were cleaned, decomposed into sub-bands, and subjected to feature extraction using MATLAB, while Python was utilized to classify sour and sweet tastes. Mean Absolute Value (MAV), Standard Deviation (STD), and Power Spectral Density (PSD) were the features extracted from EEG signals, and four algorithms were employed in the study, namely K-Nearest Neighbor (K-NN), Support Vector Machine (SVM), Random Forest, and Gate Recurrent Unit (GRU). The GRU algorithm yielded the best results with an accuracy of 91.53% using the three mentioned features. For each feature, the PSD feature and random forest algorithm delivered the highest accuracy at 91.46 %.

Item Type: Thesis (Masters)
Uncontrolled Keywords: Manis dan Asam, Elektroensefalografi, MAV, STD, Power Spectral Density, K-NN, SVM, Random Forest, Algoritma GRU, Sweet and Sour, Electroencephalography, MAV,STD Power Spectral Density, K-NN, SVM, Random Forest, GRU algorithm.
Subjects: R Medicine > RC Internal medicine > RC0321 Neuroscience. Biological psychiatry. Neuropsychiatry
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User: Ulfi Widya Astuti
Date Deposited: 05 Jun 2023 05:57
Last Modified: 05 Jun 2023 05:57
URI: http://repository.its.ac.id/id/eprint/98036

Actions (login required)

View Item View Item