August 29, 2009 at 12:25 pm
OK, this is going to be along post and not sure it will solve your problem...
What you are trying to achieve is divided into 2 categories:
1. Finding similarity between your indexed words and the sought text
2. Using Full Text Search
Well, the two just don't go together...
FTS doesn't allow you to iterate or enumerate all the keywords it indexed, so you'll need a third party tool. The one I'm using is DTSEARCH and it does a great job compared to FTS, but that's for another post. You might try LOGPARSER and use its -i:TEXTWORD option to get a list of all the unique words in your text.
After obtaining the keyword list, you will need to apply some sort of similarity algorithm. The most common for this kind for task is Levenshtein distance. It even has its SQL port somewhere on the web.
Last but not least, you might wanna try ASPELL open source speller but that will not promise any relation between the suggested corrections and your actual content.
One more thing: This is my specialty so if you need further help, links or answers - just ask.
Yours
Viewing post 1 (of 2 total)
You must be logged in to reply to this topic. Login to reply