|
I'm tring to build free text search, and WA seems do most similar job. I'm tring to understand how to build similar parsing engine. Problem. We have user input with several chunks related to different input fields. We have database with data. We need to annotate query with most probable interpretation. One solution is to wind all possible parses of query(all possible boundaries and assigned label) then check probability by language models. What other approaches can I try? |
