1

Is there a web service out there that ranks the importance of words in an arbitrary string taking into account, for example, hot words or phrases that are currently trending?

For example, let's say a user types in:

"Charlie Sheen has a major drug abuse denial problem"

For this string, I'd want to the web service to indicate that the term "Charlie Sheen" is much more important (trend-wise) than the rest of the words in the string.

Example 2:

"The Chevy Volt is a complete and utter market failure"

For this string, I want to get back "Chevy Volt".

Armchair Bronco
  • 2,367
  • 4
  • 31
  • 44
  • I haven't seen one, but I was thinking about isolating the meaningful parts of a sentence earlier today and came across a great general answer on this thread: http://stackoverflow.com/questions/90580/word-frequency-algorithm-for-natural-language-processing – Brandon Frohbieter Mar 02 '11 at 23:38
  • Thanks for the link. That helps, especially if I end up having to do this myself. I've had pretty good results so far doing a brute force approach that simply throws out any word less than 3 characters in length as well as numbers and words with special characters or formatting. (I'm only interested in English words and phrases for the time being during proof-of-concept testing.) I've also found anecdotally that people *tend* to put important words closer to the beginning of a string, so I can account for that. – Armchair Bronco Mar 02 '11 at 23:48
  • But I'll need some kind of web service or search engine to determine whether a word or phrase is trending hot or not. – Armchair Bronco Mar 02 '11 at 23:50

1 Answers1

0

After posting this question on some other sites, I finally got the answer I was looking for.

And the winner is:

http://www.evri.com/

Armchair Bronco
  • 2,367
  • 4
  • 31
  • 44