Text this: Speech prediction of a listener via EEG-based classification through subject-independent phase dissimilarity model