A Brain Wave Clustering Method for Emotion Classification

• Main Authors: Ting-Mei Li, 李亭玫
• Other Authors: Han-Chieh Chao
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/853kp5

碩士 === 國立宜蘭大學 === 資訊工程學系碩士班 === 105 === Emotional analysis is a very important topic in life. Human life belongs to group life. Emotional communication between people and people is a very important factor. Through emotional analysis can understand people's many situation. Emotional analysis can also provide much help to human social communication. There are many ways to study emotional analysis. Most of these methods use questionnaire analysis, semantic analysis, facial expression analysis ... and so on. These methods have many problems and shortcomings. Everyone has different background, personality, living conditions, and personal subjective influences. These differences can lead to inaccurate and unreliable analysis. In order to solve these problems it was suggested that the use of brain waves to do emotional classification. There are many studies confirmed. Through brainwave data analysis can be resolved, and for the analysis of human emotions. The brainwave data analysis method is the most direct and reliable. Brain wave data belongs to a kind of human biological message. Biological messages usually have emotional characteristics. Human emotional characteristics can be extracted from biological messages. There are many studies using brain waves to do emotional analysis. Most of these studies use classified algorithms to classify "positive" and "negative" emotions. But human emotions are very complicated. Human emotional development, in the baby period only a few basic emotions. The type of emotion increases with age. Because the degree of basic emotional response to the different, for the identification of emotions are gradually different. Because the degree of emotion is different, the combination of emotions becomes complicated. Emotional combination of diversification means an increase in the type of emotion. The above mentioned each person's personality subjective and differences in cultural environment. Because these differences only to do classification is not enough in the emotional analysis. In order to get more accurate emotions. Brain waves of data should also pay attention to the subtle. Any part of the data has a sense of emotional significance. So in addition to the use of classification algorithm for brain wave data to do emotional classification, should be coupled with the clustering algorithm. In order to solve the problem of emotional combination of diversification. In this study, we use the clustering algorithm to improve the accuracy of classification algorithm. Also designed the emotional model to establish the method. Using similarity method to find out good emotional information. Use this method to build an emotional model. In addition to achieving the purpose of improving accuracy. We will use the clustering algorithm to establish each person exclusive brainwave model. When different subjects receive the same stimulus, will produce a different reaction. So we use the clustering algorithm customization. Establish the exclusive brainwave model for the respondent's response to human events. The brain wave instrument used in this study can be applied to any place. Not limited to medical places can be used. The brain wave instrument can be used in any place. The system with this study can immediately understand the emotions of the subjects. This study can be applied in many ways, such as: engineering education, teaching assessment, to help language disorders express emotions ... and so on.