標題: 高斯混合模型與線性判別分析—應用於辨識不同狀態下之腦波
Gaussian Mixture Model and Linear Discriminant Analysis with Application to EEG Recognition
作者: 余承叡
王秀瑛
Yu, Cheng-Rui
Wang, Hsiuying
統計學研究所
關鍵字: 腦波;腦電圖;組合範式;高斯混合模型;線性判別分析;brainwave;EEG;montage;Gaussian mixture model;linear discriminant analysis
公開日期: 2016
摘要: 由於現今科技進步,可以在短時間內記錄大量的資料,但如何從這些龐大的數據中找出重要和可利用的資訊,為許多研究者討論的議題。在醫學界,腦電圖(Electroencephalography, EEG)屬於非侵入式檢查,每秒可以記錄256或512個資料點,因此腦電圖資料量很大,本論文發展一套統計方法來辨識不同狀態下的腦波。在本論文中,利用高斯混合模型進行特徵萃取,再透過線性判別分析進行分類,進一步使用不同的電極排序之組合範式,針對不同觀察角度做分析,並與傳統上使用的最大概似方法做比較。最後由蒐集的三十四筆腦波資料中,可以發現,本論文提出的方法比傳統的方法更有效的辨識出想像平靜與運動狀態下的腦波。
With the rapid development of technology, the amount of data produced is enormous and increasing. How to obtain useful information from huge data, such as Electroencephalography data, becomes a challenge problem. In the medical application, an electroencephalograph is a useful test to help diagnose epilepsy or other diseases, which can record 256 or 512 sample points in a second. In order to analyze these large data, in this study, we develop a method using the Gaussian mixture model and the linear discriminant analysis to efficiently recognize EEG under difference conditions. In addition, we compare this method with a traditional method using a collected data with 34 samples. The result shows that the proposed method is more efficiently to recognize the collected EEG data than the traditional method.
URI: http://etd.lib.nctu.edu.tw/cdrfb3/record/nctu/#GT070352604
http://hdl.handle.net/11536/143405
顯示於類別:畢業論文