I have multiple raters extracting URLs from the internet based on search terms.
The core issue is that a URL amounts to a string wherein two raters might come to the same URL but one string is a subset of another. For example-
There is also some nesting going on with url's nested under search terms nested under search engines which further complicates things.
My core thought on tackling this issue is to match either define a match as two strings that are equivalent for a certain percentage of characters (say 90%). The other option is to create a nearest match algorithm.
The core goal is to determine how well they reliably find the same results when using a given search term.
Any help would be appreciated in how to statistically tackle this problem.