A Feature Extraction Method for Brain Computer Interface Based on Multivariate Empirical Mode Decomposition / 生物医学工程学杂志

This paper presents a feature extraction method based on multivariate empirical mode decomposition (MEMD) combining with the power spectrum feature, and the method aims at the non-stationary electroencephalogram (EEG) or magnetoencephalogram (MEG) signal in brain-computer interface (BCI) system. Firstly, we utilized MEMD algorithm to decompose multichannel brain signals into a series of multiple intrinsic mode function (IMF), which was proximate stationary and with multi-scale. Then we extracted and reduced the power characteristic from each IMF to a lower dimensions using principal component analysis (PCA). Finally, we classified the motor imagery tasks by linear discriminant analysis classifier. The experimental verification showed that the correct recognition rates of the two-class and four-class tasks of the BCI competition III and competition IV reached 92.0% and 46.2%, respectively, which were superior to the winner of the BCI competition. The experimental proved that the proposed method was reasonably effective and stable and it would provide a new way for feature extraction.

Tensor Feature Extraction Using Multi-linear Principal Component Analysis for Brain Computer Interface / 生物医学工程学杂志

The brain computer interface (BCI) can be used to control external devices directly through electroencephalogram (EEG) information. A multi-linear principal component analysis (MPCA) framework was used for the limitations of tensor form of multichannel EEG signals processing based on traditional principal component analysis (PCA) and two-dimensional principal component analysis (2DPCA). Based on MPCA, we used the projection of tensor-matrix to achieve the goal of dimensionality reduction and features exaction. Then we used the Fisher linear classifier to classify the features. Furthermore, we used this novel method on the BCI competition II dataset 4 and BCI competition N dataset 3 in the experiment. The second-order tensor representation of time-space EEG data and the third-order tensor representation of time-space-frequency BEG data were used. The best results that were superior to those from other dimensionality reduction methods were obtained by much debugging on parameter P and testQ. For two-order tensor, the highest accuracy rates could be achieved as 81.0% and 40.1%, and for three-order tensor, the highest accuracy rates were 76.0% and 43.5%, respectively.

A New Method for Diagnosing Erythemato-squamous Diseases Based on Virtual Coding and Multinomial Logistic Regression Penalized via Elastic Net / 生物医学工程学杂志

Erythemato-squamous diseases are a general designation of six common skin diseases, of which the differential diagnosis is a difficult problem in dermatology. This paper presents a new method based on virtual coding for qualitative variables and multinomial logistic regression penalized via elastic net. Considering the attributes of variables, a virtual coding is applied and contributes to avoid the irrationality of calculating nominal values directly. Multinomial logistic regression model penalized via elastic net is thence used to fit the correlation between the features and classification of diseases. At last, parameter estimations can be attained through coordinate descent. This method reached accuracy rate of 98.34% +/- 0.0027% using 10-fold cross validation in the experiments. Our method attained equivalent accuracy rate compared to the results of other methods, but steps are simpler and stability is higher.

Group Lasso Penalized Classifier for Diagnosis of Diseases with Categorical Data / 生物医学工程学杂志

Six kinds of erythemato-squamous diseases have been common skin diseases, but the diagnosis of them has always been a problem. The quantitative data processing method is not suitable for erythemato-squamous data because they are categorical qualitative data. This paper proposed a new method based on group lasso penalized classification for the feature selection and classification for erythemato-squamous data with categorical qualitative data. The first categorical data of 33 dimensions were changed by the virtual code, and then 34th dimension age data were discretized and changed by the virtual code. Then the encoded data were grouped according to class group and variable group. Lastly Group Lasso penalized classification was executed. The classified accuracy of 10-fold cross validation was 98.88% ± 0.002 3%. Compared with those of other method in the literature, this new method is simpler, and better for effect and efficiency, and has stronger interpretability and stronger stability.

Features Interaction Lasso for Liver Disease Classification / 生物医学工程学杂志

To solve the complex interaction problems of hepatitis disease classification, we proposed a lasso method (least absolute shrinkage and selection operator method) with feature interaction. First, lasso penalized function and hierarchical convex constraint were added to the interactive model which is newly defined. Then the model was solved with the convex optimal method combining Karush-Kuhn-Tucker (KKT) condition with generalized gradient descent. Finally, the sparse solution of the main effect features and interactive features were derived, and the classification model was implemented. The experiments were performed on two liver data sets and proved that features interaction contributed to the classification of liver diseases. The experimental results showed that the feature interaction lasso method was of strong explanatory ability, and its effectiveness and efficiency were superior to those of lasso, of all pair-wise lasso, support vector machine (SVM) method, K nearest neighbor (KNN) method, linear discriminant analysis (LDA) classification method, etc.

A novel method of multi-channel feature extraction combining multivariate autoregression and multiple-linear principal component analysis / 生物医学工程学杂志

Brain-computer interface (BCI) systems identify brain signals through extracting features from them. In view of the limitations of the autoregressive model feature extraction method and the traditional principal component analysis to deal with the multichannel signals, this paper presents a multichannel feature extraction method that multivariate autoregressive (MVAR) model combined with the multiple-linear principal component analysis (MPCA), and used for magnetoencephalography (MEG) signals and electroencephalograph (EEG) signals recognition. Firstly, we calculated the MVAR model coefficient matrix of the MEG/EEG signals using this method, and then reduced the dimensions to a lower one, using MPCA. Finally, we recognized brain signals by Bayes Classifier. The key innovation we introduced in our investigation showed that we extended the traditional single-channel feature extraction method to the case of multi-channel one. We then carried out the experiments using the data groups of IV-III and IV - I. The experimental results proved that the method proposed in this paper was feasible.

Application of semi-supervised sparse representation classifier based on help training in EEG classification / 生物医学工程学杂志

Electroencephalogram (EEG) classification for brain-computer interface (BCI) is a new way of realizing human-computer interreaction. In this paper the application of semi-supervised sparse representation classifier algorithms based on help training to EEG classification for BCI is reported. Firstly, the correlation information of the unlabeled data is obtained by sparse representation classifier and some data with high correlation selected. Secondly, the boundary information of the selected data is produced by discriminative classifier, which is the Fisher linear classifier. The final unlabeled data with high confidence are selected by a criterion containing the information of distance and direction. We applied this novel method to the three benchmark datasets, which were BCI I, BCI II_IV and USPS. The classification rate were 97%, 82% and 84.7%, respectively. Moreover the fastest arithmetic rate was just about 0. 2 s. The classification rate and efficiency results of the novel method are both better than those of S3VM and SVM, proving that the proposed method is effective.

Research of controlling of smart home system based on P300 brain-computer interface / 生物医学工程学杂志

Using electroencephalogram (EEG) signal to control external devices has always been the research focus in the field of brain-computer interface (BCI). This is especially significant for those disabilities who have lost capacity of movements. In this paper, the P300-based BCI and the microcontroller-based wireless radio frequency (RF) technology are utilized to design a smart home control system, which can be used to control household appliances, lighting system, and security devices directly. Experiment results showed that the system was simple, reliable and easy to be populirised.

Study on the method of feature extraction for brain-computer interface using discriminative common vector / 生物医学工程学杂志

Discriminative common vector (DCV) is an effective method that was proposed for the small sample size problems of face recognition. There is the same problem in brain-computer interface (BCI). Using directly the linear discriminative analysis (LDA) could result in errors because of the singularity of the within-class matrix of data. In our studies, we used the DCV method from the common vector theory in the within-class scatter matrix of data of all classes, and then applied eigenvalue decomposition to the common vectors to obtain the final projected vectors. Then we used kernel discriminative common vector (KDCV) with different kernel. Three data sets that include BCI Competition I data set, Competition II data set IV, and a data set collected by ourselves were used in the experiments. The experiment results of 93%, 77% and 97% showed that this feature extraction method could be used well in the classification of imagine data in BCI.

Design and implementation of controlling smart car systems using P300 brain-computer interface / 生物医学工程学杂志

Using human electroencephalogram (EEG) to control external devices in order to achieve a variety of functions has been focus of the field of brain-computer interface (BCI) research. P300 is experiments which stimulate the eye to produce EEG by using letters flashing, and then identify the corresponding letters. In this paper, some improvements based on the P300 experiments were made??. Firstly, the matrix of flashing letters were modified into words which represent a certain sense. Secondly, the BCI2000 procedures were added with the corresponding source code. Thirdly, the smart car systems were designed using the radiofrequency signal. Finally it was realized that the evoked potentials were used to control the state of the smart car.

Research on operating characteristics of multiclass receiver in machine learning / 生物医学工程学杂志

The Receiver Operating Characteristic (ROC) curve has become a standard tool for the analysis and comparison of binary classifiers when the costs of misclassification are unknown. In fact ROC curve has replaced the correct rate or error rate. Extending this to the multiclass case has recently become a growing topic of interest. The conceptions and application of ROC curve are expounded. The history and some algorithms of the multi-class ROC surface are given in detail in this paper. Finally research trends of the multi-class ROC surface are approximating computing and visualization.

Research on the methods for multi-class kernel CSP-based feature extraction / 生物医学工程学杂志

To relax the presumption of strictly linear patterns in the common spatial patterns (CSP), we studied the kernel CSP (KCSP). A new multi-class KCSP (MKCSP) approach was proposed in this paper, which combines the kernel approach with multi-class CSP technique. In this approach, we used kernel spatial patterns for each class against all others, and extracted signal components specific to one condition from EEG data sets of multiple conditions. Then we performed classification using the Logistic linear classifier. Brain computer interface (BCI) competition III_3a was used in the experiment. Through the experiment, it can be proved that this approach could decompose the raw EEG singles into spatial patterns extracted from multi-class of single trial EEG, and could obtain good classification results.

An overview of feature selection algorithm in bioinformatics / 生物医学工程学杂志

Feature selection (FS) techniques have become an important tool in bioinformatics field. The core algorithm of it is to select the hidden significant data with low-dimension from high-dimensional data space, and thus to analyse the basic built-in rule of the data. The data of bioinformatics fields are always with high-dimension and small samples, so the research of FS algorithm in the bioinformatics fields has great foreground. In this article, we make the interested reader aware of the possibilities of feature selection, provide basic properties of feature selection techniques, and discuss their uses in the sequence analysis, microarray analysis, mass spectra analysis etc. Finally, the current problems and the prospects of feature selection algorithm in the application of bioinformatics is also discussed.

Extraction method of the visual graphical feature from biomedical data / 生物医学工程学杂志

The vector space transformations such as principal component analysis (PCA), linear discriminant analysis (LDA), independent component analysis (ICA) or the kernel-based methods may be applied on the extracted feature from the field, which could improve the classification performance. A barycentre graphical feature extraction method of the star plot was proposed in the present study based on the graphical representation of multi-dimensional data. The feature order question of the graphical representation methods affecting the star plot was investigated and the feature order method was proposed based on the improved genetic algorithm (GA). For some biomedical datasets, such as breast cancer and diabetes, the obtained classification error of barycentre graphical feature of star plot in the GA based optimal feature order is very promising compared to the previously reported classification methods, and is superior to that of traditional feature extraction method.

Research on magnetoencephalography-brain computer interface based on the PCA and LDA data reduction / 生物医学工程学杂志

The magnetoencephalography (MEG) can be used as a control signal for brain computer interface (BCI). The BCI also includes the pattern information of the direction of hand movement. In the MEG signal classification, the feature extraction based on signal processing and linear classification is usually used. But the recognition rate has been difficult to improve. In the present paper, a principal component analysis (PCA) and linear discriminant analysis (LDA) method has been proposed for the feature extraction, and the non-linear nearest neighbor classification is introduced for the classifier. The confusion matrix is analyzed based on the results. The experimental results show that the PCA + LDA method is effective in the analysis of multi-channel MEG signals, improves the recognition rate to the extent of the average recognition rate 55.7%, which is better than the recognition rate 46.9% in the BCI competition IV.