Consistency of Concentration and Relaxation Classification in EEG Signal

• Main Authors: Yu-Cheng Wu, 吳育呈
• Other Authors: Shing-Chern You
• Format: Others
• Language: en_US
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/e5h24y

碩士 === 國立臺北科技大學 === 資訊工程系 === 106 === The human brain is the central organ of the human nervous system. It is so mysterious that even in modern technology nowadays people still cant wholly decode the brain. However, we can record the electrical activity of the brain through Electroencephalography (EEG). It is known that we can record the power spectrum data of the brain by EEG. However, it is hard to tell the emotion of the subject from these data since theres no such a rule to distinguish the representation of EEG data. The brainwaves are changing all the time. A person whose brainwaves may not be identical even if doing the same thing at the same time. The purpose of this study is to verify whether brainwaves have the consistency or not, we applied different machine learning algorithm and analyzing the consistency of brainwaves at different times on different days. This thesis is organized into four parts. The first section, EEG and Database, establishes the datasets that recorded on our own and the recording method. In the second section, Data Preprocessing, the data from EEG is a high dimensionality data that would cause overhead and affects the accuracy of machine learning if we dont apply dimensionality reduction or feature extraction algorithms. The third section, Introduction to Machine Learning, specifics explained the machine learning algorithms that applied in this study, including LDA, ANN, backpropagation, KNN, SVM, and autoencoder. The final section, Observations and Results in Experiment, describes the observations and results in the experiment.