http://snipplr.com/language/mysql/tags/ranking Recent snippets posted on Snipplr.com en-us Wed, 19 Jun 2013 21:33:29 GMT http://snipplr.com/view/51513/very-simple-search-retrieval-query-ranking-with-tfidf-weight/ This assumes two tables: `search_results` and `keyword_counts`. `search_results` is an inverted index with, at minimum, fields for the keyword indexed, its occurrences, and the indexed item's id in the table it is in (e.g. `item_id` in `search_results` could be `45`, referencing `id = 45` in `documents`) in the table it's in. `keyword_counts` has a row for each indexed item with the total number of unique words in it and its id in the table the actual item is in. Released into the public domain. This may have bugs in it, but it's worked so far for me.

]]> Mon, 04 Apr 2011 20:33:17 GMT http://snipplr.com/view/51513/very-simple-search-retrieval-query-ranking-with-tfidf-weight/ http://snipplr.com/view/22559/bayesian-ranking/ Use Bayesian values to get a "more fair" ranking than average value based ranking. Example with a "games" table having some ratings in table "ratings"

]]> Thu, 05 Nov 2009 05:07:51 GMT http://snipplr.com/view/22559/bayesian-ranking/