Nearest Neighbour Searching
Posted: Mon Oct 17, 2005 9:46 am
Hi, I am a student and im currently working on a project with RSS news feeds. I set up a little crawler to visit a list of news feeds and grab the latest headlines. Thats been running every 10 minutes for the last 8 months or so now and I have a database of some 35,000 headlines. I've created a front end to view the headlines and search etc, but when the user is look reading the news article I want to be able to provide details of other related headlines.
Im just not exactly sure on how to implement it.
I've tried taking the headline and description of the article, removing common words ("a", "in" etc) and then using the top 3 words that occur most frequently as keywords to do bolean search on the headlines table.
is there a better way ?
Im just not exactly sure on how to implement it.
I've tried taking the headline and description of the article, removing common words ("a", "in" etc) and then using the top 3 words that occur most frequently as keywords to do bolean search on the headlines table.
is there a better way ?