Question

I know how to compare two texts and get all the single words that appear in both. But how can I match expressions/phrases?

For example:
1. “This is the computer maker Apple”
2. “Apple is a California based great computer maker”

Now 🙂

1. Apple is clearly present in both.

2. computer and maker are present in both. I could check at this point if they are a group of words(one follows the other one).

But for the speed of processing, isn’t there a way to match “computer maker” and not each one and then check if present as a group.

Keep in mind that the example given is trivial and just for the purpose of exemplifying, in practice more complicated sentences/texts may be presented.

Answer 1

You could parse both strings and split on whitespace to get token arrays A1 and A2. Then, simply check every contiguous subsequence in A1 for a matching one in A2. This looks like O(n^4) to me, which is better than getting all the single matches and looking for combinations… which is not polynomial.

  1. the cat is on the roof
  2. a man is on the stage

  A1 = [the, cat, is, on, the, roof]
  A2 = [a, man, is, on, the, stage]

  [the]: no match
  [cat]: no match
  [is]: match
  [is, on]: match
  [is, on, the]: match
  [is, on, the, roof]: no match
  [on]: match
  [on, the]: match
  [on, the, roof]: no match
  [the]: match
  [the, roof]: no match
  [roof]: no match
  -end-

Recursion seems like an elegant way to implement something like this. If you need something more efficient, I’m sure that there is a smarter way to do it than this.