標題: 應用有限混合常態模型與支持向量機分析想像情境下的腦波訊號
Recognizing EEG Data Using the Finite Mixture Normal Model and Support Vector Machine
作者: 呂品慧
王秀瑛
Lu, Pin-Hui
Wang, Hsiu-ying
統計學研究所
關鍵字: 支持向量機;有限混合常態模型;腦波辨識;Support Vector Machine;Finite Mixture Normal Model;EEG
公開日期: 2016
摘要: 大腦是人類各種機能的訊號中心,掌管身體各部位及器官之運作,因此腦科學一直以來都是醫學上重要的研究領域,而近年來更是廣泛運用到許多不同領域,例如透過偵測腦波訊號辨識疾病、操控機械或神經義肢…等,因此在本論文中,我們嘗試錄製健康受試者的腦波資料並進行分類。我們將腦波資料透過有限混合常態模型並利用EM演算法所估計參數,以平均數及標準差來表示一連串的腦波資料後,將這些參數作為支持向量機的分類依據,分別使用不同的參數組合並透過Grid Search以及 leave-one-out cross validation計算分類正確率,找出較適合作為本研究之腦波分類的參數。
In this study, we aim to recognize Electroencephalography (EEG) data under different imaging conditions. We recorded 17 healthy persons’ EEG. The data are fitted by the finite mixture normal model, and the parameters of the mixture models are estimated by the EM algorithm. We propose using the support vector machine method to classify the EEG data, and adopt the leave-one-out cross validation to compute the accuracy of classification. The accuracy of the classification based on the support vector machine is higher than that based on the maximum likelihood approach.
URI: http://etd.lib.nctu.edu.tw/cdrfb3/record/nctu/#GT070352611
http://hdl.handle.net/11536/143337
顯示於類別:畢業論文