Totally automated keyword extraction
Tayfun Pay · 2016
We develop and analyze an unsupervised and domain-independent method for extracting keywords from single documents. Our approach differs from the previous ones in the way of identifying candidate keywords, pruning the list of candidate keywords with several filtering heuristics and selecting keywords from the list of candidate keywords according to dynamic threshold functions. We were able to obtain better precision and recall on the same data set with our method than the previously studied ones.