This algorithm allows the user to use his own features. Thus, the featureSelection function selects the best set of the features among the features given by the user. It can be used to select features of any number of classes.
This function was designed to do different types of plots of EEG data. Graphs of the original data, of the spectrum, continuous wavelet transform and t-value scalogram of the signals can be plotted. The main idea is to help the user to find nice features to use in his final model.
svmEEG is used to train a support vector machine classifier of the features selected by the function FeatureEEG. Internally, this function uses the svm function available in the e1071 package. Thus, it is recommended to understand the svm function before using svmEEG.
Creates an object simulating ARIMA random variables. The created object contains data in the format required to use other methods of this package. Makes a simulation similar to EEG data to test the capabilities those methods. NOTE: The only purpose of the simulated data is to test the package features, it is not to be used to study properties of real EEG data!
This function was designed to do charts of EEG data. Basically, the data (for each channel, class and recording) is divided by windows and a statistical function (as the mean or variance) is applied for all signals in each window. Then, the sequence of values obtained for each window is plotted.
This function helps the user to produce an object of class Features. The purpose of this object is to be used in the function FeatureEEG as the parameter features. It is possible to construct this object "by hand" but it can be a difficult task. It takes some time to choose all the features, so it is recommended to save the object afterwards.
● Data Source:
CranContrib
● Keywords: features
● Alias: easyFeatures, print.Features, summary.Features
●
0 images
Select the best features to classify EEG data. This function receives as input a list of features defined by the user using the easyFeatures function. Then, the algorithm will use several statistical tests to search for the the best set of features in terms of classification. This kind of analysis is very useful to reduce the dimensionality of the data, producing much faster and accurate classifiers.
This package consists of a set of tools to classify electroencephalography (EEG) and to successfully reduce the feature space dimension. More specifically, this package contains functions to simulate data (randEEG), to train classifiers (svmEEG), to classify new data (classifyEEG) and to plot data (plotEEG and plotwindows). Nevertheless, what differentiates this package from others available in the community are the functions to automatically select the best features to use in the classification model (featureSelection and FeatureEEG).