How can I reduce Stanford parser time of execution?

Question

Good Morning,

I am using Stanford part of speech for my application, and I need to parse many text files (can go to 20 files). However I noticed that stanford takes a lot of time (sometimes 30min or more), the problem is that Stanford automatically builds the syntax tree, even when I don't need it. Have anyone face this problem before? what can I do to reduce the parsing time of stanford?

Any help is welcome, Regards.

Answer 1

If you just want part of speech, not parse trees, you should run a (sequence model) part of speech tagger, such as the Stanford POS Tagger , which is orders of magnitude faster). See the Stanford Parser FAQ . (If you want fast tagging, use the left3words-wsj-0-18.tagger model of the Stanford POS Tagger.)