KLASIFIKASI TIGA KONDISI (IMAJINASI GERAKAN TANGAN KANAN DAN KIRI SERTA PENGUCAPAN KATA) BERBASIS DATA EEG MENGGUNAKAN METODE SUPPORT VECTOR MACHINE; THREE CONDITION CLASSIFICATION (IMAGERY LEFT AND RIGHT HAND MOVEMENT AND SPECH RANDOM) BASED ON EEG DATA USING SUPPORT VECTOR MACHINE METHOD

TONTOWI, IRVAN ALBAB, Agfianto Eko Putra

2016 | Skripsi | FMIPA

Advancements in biomedical signal processing tecniques Electroencephalography (EEG) signals, widely used in the diagnosis of brain diseases and in the application field was Brain Computer Interface (BCI). BCI uses signals from brain (EEG) as a input to control other devices such as a computer, wheel chair, etc. The aim of this work is to see how accurate classification 3 condition (right hand and left hand movement imagery and random generating word) using EEG data. 3 subject there were two seasion for each subject, and three possible types :right hand movement, left hand movement imagery, and random generating word. The EEG data was sampled at 512 Hz, bandpass filtered between 8 and 30 Hz. After the preprocessing of EED data, then data to decomposision level 5, and used mother wavelet daubechies 5 as mother wavelet. Signal calculate with power spectrum density method of feature exctraction. Then for classification using multiclass SVM and 4-fold cross validation for accuracy classification. Result for this work using wavelet transformation and multiclass SVM accuracy for each subject for 3 subject get average accuracy to 78.70% and for mixed subject accuracy amounted to 67.67%.

Kata Kunci : N