From: A graphical heuristic for reduction and partitioning of large datasets for scalable supervised training
word2vec input dimension
Training time (seconds)
Prediction accuracy (%)
LIBSVM
GSH (/speed-up)
GSH
500
100.88
12.06 /8.3
98.04
98.79
1000
201.33
21.64 /9.3
98.12
98.91