How can I find out similar texts / strings given a string?

Question

In sites like Quora or Stack Overflow, on a particular question, they are able to show questions that are possibly related. A quick look at them would reveal that they are merely looking for questions with similar text content. Is there any standard technique to find out such similar texts from a DB table where all texts are stored?

For example if you go to this question - How to remove Application icon from Action Bar in Android? it shows the following question as related - Remove application icon and title from Honeycomb action bar .

If I have column questionText, where the questions texts are stored, in a table questions, how will I find out such related strings?

It is called a "full text index". – Lasse V. Karlsen Feb 18 '15 at 16:08 — Lasse V. Karlsen, Feb 18 '15 at 16:08
Other used Sphinx (with MySQL) for example. – ares777 Feb 18 '15 at 16:10 — ares777, Feb 18 '15 at 16:10

score 0 · Accepted Answer · answered Feb 18 '15 at 16:12

0

You'll need to extract keyword tokens from your documents and them find correlation between them. You can use a FOSS tool like Apache Lucene to do that for you.

They can rely on tags also.

answered Feb 18 '15 at 16:12

lnrdo

396
1
13

How can I find out similar texts / strings given a string?

1 Answers1