[eluser]Pascal Kriete[/eluser]
I'll try to explain.
With each entry the user also enters tags - simple keywords that describe the situation. So for a sample article, let's use "The Extinct Short-snouted Aardvark: A Case Study", you might have "short-snouted", "aardvark", and "starved". Then you can query your database for articles that contain one or more of the same tags.
The biggest problem I see here is that news are only relevant for about a week. So just drawing on old articles doesn't present relevant information. Instead you might try to categorize news. For example the New York Times has "World", "US", "Region", "Technology", "Science", "Arts", etc. They just display recent articles in the same category, as those are more relevant.