I'm curious to know how the relevance ranking is determined after a text search. I assumed the most relevant would be the most amount of times the word appears in a document, but that doesn't seem to be the case.
Question
Question
How is the relevance ranking determined after a text search?
asked on May 23, 2018
•
Show version history
0
0
Answer
SELECTED ANSWER
replied on May 23, 2018
For the last several releases Laserfiche has used a variation of the BM25 ranking function. There is a Wikipedia article on BM25 that appears to be accurate where you can learn more. Ranking documents by the number of times a search term appears would bias results in favor of longer documents. Probabilistic methods like BM25 attempt to remove such biases to assign more relevant documents a higher rank; this is why Laserfiche uses BM25.
Old versions of Laserfiche (before 8.0) do rank by the number of occurrences of the search terms. Users can still sort search results by the number of occurrences by the hit count column to sort by in the search results view in the Laserfiche client.
2
0
Replies
You are not allowed to reply in this post.
You are not allowed to follow up in this post.