Abstract: This paper analyses various unsupervised automatic keyphrase extraction methods based on graphs as well as the impact of word embedding. Evaluation is made on three datasets. We show that there is no differences when using word embedding and when not using it.
Loading