README.md

In shariq-mohammed/stSpikeSlabEEG: Classification Models For Electroencephalography Data

stSpikeSlabEEG

R package for "Mohammed, S., Dey, D.K. and Zhang, Y., 2020. Classification of high-dimensional electroencephalography data with location selection using structured spike-and-slab prior. Statistical Analysis and Data Mining: The ASA Data Science Journal, pp.1-17. https://doi.org/10.1002/sam.11477"

Code to run the model in (Mohammed et.al, 2020) on the EEG data using the package stSpikeSlabEEG.

Contents

1. Load EEG data and divide it into training and test.
2. Build the model proposed in the manuscript with the training data and predict for the test data set.
3. Run a 5-fold cross-validation with the EEG data provided in the package.

# install the package (devtools package needed)
if(!require(devtools)) install.packages("devtools")
devtools::install_github('shariq-mohammed/stSpikeSlabEEG')

# Load the package
library(stSpikeSlabEEG)

Load EEG data and divide it into training and test

Load EEG Data

X = stSpikeSlabEEG::X_eeg

Load location indices for which distance matrix is available

ind64to57 = stSpikeSlabEEG::ind64to57_eeg

Update EEG data for locations above

X = X[,ind64to57,]

Load binary response

y = t(stSpikeSlabEEG::y_eeg)

Dimensions of EEG data

tau = dim(X)[1]
p = dim(X)[2]
n = dim(X)[3]

Scale data at subject level by its Frobenius norm

X_int = sapply(1:n, function(i) sqrt(sum(X[,,i]^2)))
X_sc = array(NA, dim = dim(X))
for(i in 1:n) X_sc[,,i] = X[,,i]/X_int[i]

Sample indices for training data

trn.ind = sample.int(n, size = 100)

Responses for training and testing

y.trn = y[trn.ind] # responses from training data
y.test = y[-trn.ind] # responses from test data

EEG data for training and testing

X.trn = aperm(X_sc[,,trn.ind], c(3,2,1)) # training EEG data
X.test = aperm(X_sc[,,-trn.ind], c(3,2,1)) # test EEG data

Build the model proposed in the manuscript with the training data and predict for the test data set

Fit the model

modelFit = strucBayes(y = y.trn, X = X.trn, dist.mat =  dist.mat,
                      Nmcmc = 1000, burnin = 100, thin = 10, # MCMC settings
                      X.test = X.test)

Estimates for the coefficients, local predicted probabilities.

modelFit$b.strucBayes
modelFit$p.strucBayes.pred

Response prediction for the test data set

modelFit$y.strucBayes.pred

Run a 5-fold cross-validation with the EEG data provided in the package

cvResults = crossVal(seed = 113, # seed for creating fold-splits
                     Nmcmc = 1000, burnin = 100, thin = 10)

shariq-mohammed/stSpikeSlabEEG documentation built on Aug. 2, 2020, 11:44 a.m.