Dissertation > Excellent graduate degree dissertation topics show
Feature Extraction and Recognition of EEG Based on the AR Model
Author: ZouQing
Tutor: TangJingTian
School: Central South University
Course: Biomedical Engineering
Keywords: BCI EEG AAR algorithm MVAAR algorithm Tasks Classification
CLC: R318
Type: Master's thesis
Year: 2008
Downloads: 405
Quote: 9
Read: Download Dissertation
Abstract
|
Brain - machine interfaces become a hot research topic in recent years, brain science , rehabilitation engineering , automatic control , the military field and biomedical engineering field because of its broad application value and prospect . The processing of the EEG is the emphasis and difficulty in the course of the study . Desynchronization EEG events / phase synchronization phenomenon as characteristic information , in-depth discussion of the adaptive algorithm based on AR model (AAR) and multi - variable parameters the the AAR model algorithm (MVAAR) EEG characteristics Extraction application . Describes a variety of model coefficients estimation method , using the Kalman filter method and fast QR decomposition respectively the AAR, MVAAR model coefficient estimates , appeared to maximize the characteristic information in the EEG signal . Linear analysis , task identification , classification and leave-one- three classifier based on the Mahalanobis distance . The introduction of mutual information , and kappa value , the concept of the value of the area in the ROC under the curve for the performance evaluation of the classification results . From the experimental results , MVAAR algorithm than AAR algorithms achieve a higher classification accuracy . AAR model to describe the non- stationary random characteristics of the EEG signal MVAAR algorithm to identify law subjectivity smaller low order generally select consistent high data simulation , multi- lead data input have a stronger versatility. Traditional linear classification , based on the Mahalanobis distance secondary classification , leave-one classification to achieve good results , but also have their own advantages and disadvantages . LDA and MDA algorithms are only determined by the mean and covariance of the data , when the larger of the two types of covariance matrix difference LDA method will exhibit a large deviation MDA approach will exhibit better results. Leave-one principle is simple , easy to implement , but when a large experimental data , the amount of computation and computing time will be the question we must consider . Different object because a different individual differences and test feedback period , its use classification of the same set of algorithms effect .
|
Related Dissertations
- A Study on the Coma Scales and Electroencephalogram for the Prediction of the Comatose Patient,R741.044
- Research of the Mental Fatigue Classification Based on HHT and Kernel Method,TN911.7
- EEG -based emotion recognition,TP391.4
- Neural electrophysiological dress Contact Comfort Research,Q42
- Machine learning method based automatic classification of EEG,TP181
- Research and design of the television remote control system based on brain-computer interface technology,TP872
- Study on Assessment Method of Mental Fatigue Based on Brain-Computer Interface,R318
- Intraoperative Electrocorticography Monitoring of Focal Excision of Intractable Epilepsy by Microsurgical Treatment,R742.1
- Evaluation of Sample Entropy and Bispectral Index on Monitoring Sedation Depth during Target-controlled Infusion of Propofol Anesthesia,R614
- EEG based on compressed sensing compressive sampling,TN911.7
- The Mental Arithmetic Study of EEG by Factor Analysis,R741.044
- EEG Analysis by Using High-Performance Computation,R318.0
- EEG Early-warning Indicators of Exploratory Study on Patients with Mild Cognitive Impairment or Alzheimer’ Disease,R749.16
- Quantitative Eeg Power Spectral Characteristics of Mild Cognitive Impairment,R749.1
- The EEG Experimental Study on the Function of Simple Self-relaxation-hypnosis in Releasing Learning Fatigue,B842.3
- The EEG Signal Collecting and Processing System of White Rat Based in PSoC and LabVIEW,TP274.2
- The Control Method Research for Prosthetic Hand Based on EEG,TP391.4
- Research on Analysis & Recognition Method for Multi-motor ERD/ERS Signal in BCI System,TN911.6
- Study on Extraction of α Rhythm in EEG Based on Wavelet Package and ICA,TN911.7
- Research and Design of the Sistem That Getting and Handling Mice Brain Signal,TH789
- Primary Study on the Rehabilitation Mechanism of Internet Addicction Disorder Based on ET Detecting Technology,R749.99
CLC: > Medicine, health > Basic Medical > Medical science in general > Biomedical Engineering
© 2012 www.DissertationTopic.Net Mobile
|