Fig. 4
From: Graph-based exploration and clustering analysis of semantic spaces
Confusion matrices for (a) standard classification and (b) classification when words included in a maximum clique around word ‘excelent’ from the respective word embedding are treated as words with the same meaning. Accuracy scores of a classifier on a training and test instances are equal to (a) 0.9483, 0.9433 and (b) 0.9521, 0.9461 respectively