Right now the model takes a brute force approach of calculating similarity between an SoW with all the standard abstracts. This takes a some time. We need to reduce this time.
Ideas:
We can do some filtering using an inverted index and number of keyword matches to reduce the candidate size of standards we want to compare the SoW with.
Right now the model takes a brute force approach of calculating similarity between an SoW with all the standard abstracts. This takes a some time. We need to reduce this time.