Klastering Emosi Berdasarkan Gelombang Otak Sinyal EEG Menggunakan Fuzzy C-Means Clustering

Fasich, Delvina Aulia (2017) Klastering Emosi Berdasarkan Gelombang Otak Sinyal EEG Menggunakan Fuzzy C-Means Clustering. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
2913100027_Undergraduate_Thesis.pdf - Published Version

Download (5MB) | Preview

Abstract

Berdasarkan ilmu psikologi, emosi berpengaruh besar terhadap kualitas dan kuantitas dari aktivitas individu. Keadaan emosi individu dapat dilihat secara nyata melalui ekpresi wajah maupun nada bicara. Selain melalui fitur wajah maupun fitur suara, identifikasi emosi juga bisa dilakukan melalui gelombang otak. Pada tugas akhir ini penulis menggunakan sinyal electroencephalogram sebagai input untuk melakukan klastering emosi. Sinyal electroencephalogram ini dipilih karena dapat merekam emosi sebenarnya dari individu. Dengan menggunakan pengukuran statistik pada domain waktu sinyal, fitur-fitur yang terdapat dalam sinyal EEG dijadikan acuan untuk melakukan klastering emosi. Fitur didapatkan dari delapan channel yaitu channel F8, T7, CP1, CP2, P7, FC2, F4 dan AF3. Emosi yang diolah dan dianalisis yaitu senang, sedih, puas, terkejut, terlindung, tidak peduli, marah dan takut berdasakan parameter valence, arousal dan dominance. Klastering sinyal dilakukan dengan menggunakan Fuzzy C-Means Clustering. Banyaknya cluster menunjukkan banyaknya emosi yang akan dikenali. Penelitian ini menghasilkan nilai output berupa sistem yang dapat mengelompokkan emosi. Nilai akurasi tertinggi didapatkan pada kondisi C=2 pada kombinasi channel F8, T7, CP1, CP2, P7, FC2 dan F4 dengan nilai akurasi rand index 60.31%. ================================================================================================== In psychology, emotions greatly a_ect the quality and quantity of individual activities. The emotional state of an individu can be seen through the real expression of the face and tone of speech. Other than using facial features as well as voice features, emotional recognition can also be done based on brain wave signal. In this final project author use electroencephalogram signal as input to do cluster some of human emotion. This electroencephalogram signal is chosen because it can record the actual emotions of the individu. Using statistical measurements in time domain signal, the features contained in EEG signals were used to cluster emotions. Features obtained from eight channels which are F8, T7, CP1, CP2, P7, FC2, F4 and AF3. Emotions that were processed and analyzed are happy, sad, satisfied, surprised, protected, unconcerned, angry and frightened based on parameters valence, arousal and dominance. The signal clustering was performed using Fuzzy C-Means Clustering. The number of clusters shows the number of emotions to be recognized. This research produces an output value of a system that can clustering emotions. The highest accuracy is obtained when C=2 in combination of channel F8, T7, CP1, CP2, P7, FC2 and F4 with 60.31% of rand index accuration value

Item Type: Thesis (Undergraduate)
Additional Information: RSKom 004.35 Fas k-1
Uncontrolled Keywords: Electroencephalogram, Emosi, Sinyal Domain Waktu, Klastering, Fuzzy C Means
Subjects: Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics)
T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.546 Computer algorithms
Divisions: Faculty of Industrial Technology > Multimedia and Network Engineering > (S1) Undergraduate Theses
Depositing User: Fasich Delvina Aulia
Date Deposited: 11 Jan 2018 07:08
Last Modified: 05 Mar 2019 06:19
URI: http://repository.its.ac.id/id/eprint/48198

Actions (login required)

View Item View Item