Adding Keywords to the Postings

If we look at the postings in the Erlang mailing list, you’ll see they don’t have any keywords. But even if they did, there is a more fundamental problem. Two different people might read the same document and disagree as to what keywords should be used to describe the document. So, if I do a keyword search, it won’t work if the keywords that the author of the document chose were different from the keywords I chose to search the document with.

Sherlock computes a keyword vector for each posting in the mailing list. We can ask Sherlock what the derived keyword vector was for posting 946 in 2009.

 
1>​ sherlock:get_keyword_vector("2009", 946).
 
[{"protocols",0.6983839995773734},
 
{"json",0.44660371850799946}, ...

Get Programming Erlang, 2nd Edition now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.