One vs One Cross Domain

AbstractThe goal of this paper is to use all available Polish language data sets to seek the best possible performance in supervised sentiment analysis of short texts. We use text …...