Klasifikasi Emosi Dari Data Sinyal EEG Menggunakan Independent Component Analysis (ICA), Wavelet Denoising Dan Multiple Discrimnant Analysis (MDA)

Syahdeini, Aldy (2015) Klasifikasi Emosi Dari Data Sinyal EEG Menggunakan Independent Component Analysis (ICA), Wavelet Denoising Dan Multiple Discrimnant Analysis (MDA). Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
5111100122-Undergraduate Thesis.pdf - Published Version

Download (1MB) | Preview

Abstract

Emosi ialah keadaan psikologi pada alam sadar manusia sebagai reaksi terhadap sebuah objek atau keadaan yang juga dikaitkan dengan suasana hati, temperamen, jati diri, watak dan motivasi. Emosi dapat diekspresikan dengan tindakan verbal melalui ucapan, maupun non-verbal seperti intonasi suara atau gerakan tubuh. Emosi sangat penting dalam meningkatkan kualitas kehidupan manusia. Dengan adanya sistem kecerdasan emosi pada komputer, diharapkan manusia dapat merasa lebih nyaman dalam berinteraksi dan bekerja dengan komputer. Emosi dapat dikenali dengan menggunakan sinyal electroencephalography (EEG). Pergerakan mata, kedipan mata, gerakan otot, sinyal jantung dapat membuat sinyal EEG menjadi tidak akurat sehingga diperlukan preprocessing data yang efektif. Dalam tugas akhir ini penulis melakukan klasifikasi emosi dari sinyal EEG menggunakan Multiple Discriminant Analysis (MDA), dimana sebelumnya dilakukan preprocessing menggunakan Independent Component Analysis (ICA) dan wavelet denoising. Berdasarkan hasil percobaan, klasifikasi emosi dengan preprocess menggunakan ICA dengan merata-ratakan 100 fitur menghasilkan akurasi sebesar 70%, sedangkan tanpa menggunakan ICA viii menghasilkan akurasi sebesar 65%. Hal ini membuktikan bahwa preprocess menggunakan ICA lebih efektif untuk mengurangi noise dan meningkatkan akurasi klasifikasi dibandingkan tanpa ICA. ================================================================================================ Emotion is a psychological state of the human subconscious as a reaction from an object or a situation which also associated with mood, temperament, identity, character and motivation. Emotion can be expressed by the verbal action through speech or body movement. Emotion is very important in improving the quality of human life. With the emotional intelligence in computer systems, it is expected that people can be more comfortable in interacting and working with computers. Emotion can be identified by using electroencehalography (EEG). Eye movements, blinking, muscle movement and heart signal can make the EEG signal becomes inaccurate, so it is necessary to preprocess the data using the effective method. In this thesis the author tries to classify emotions from EEG signals data using Multiple Discriminant Analysis (MDA) and to preprocess using Independent Component Analysis (ICA) and wavelet denoising. Based on the result of the trials, emotion classification using ICA and 100 fitures in average, the classification produces 70% of accuration. On other hand without using ICA, it produces 65% of accuracy. Therefore we can viii conclude that preprocessing signals using ICA are more effective for reducing the noise than without ICA.

Item Type: Thesis (Undergraduate)
Additional Information: RSIf 621.398 1 Sya k
Uncontrolled Keywords: Emosi, electroencephalography, klasifikasi, Independent Component Analysis, wavelet denoising, Multiple Discriminant Analysis
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing.
Divisions: Faculty of Information Technology > Informatics Engineering > (S1) Undergraduate Theses
Depositing User: Mr. Tondo Indra Nyata
Date Deposited: 16 May 2018 02:13
Last Modified: 16 May 2018 02:13
URI: http://repository.its.ac.id/id/eprint/51885

Actions (login required)

View Item View Item