Is fuzzy query in elasticsearch related to fuzzy logic? -
as title states, in elasticsearch's fuzzy-query related fuzzy logic?
for example, given string, fuzzy query fuzziness
of 2
return indexed strings have levenshtein distance of 2. how system decide answers return if there multiple matches?
is there fuzzy system behind it? 1 has triangular functions (for instance) , can expressed in this:
1| b | /\ /\ = fuzzy set 1 | / \/ \ b = fuzzy set 2 | / /\ \ 0|/ / \ \ ------------ b c d
i more theoretical answer tackles in fuzzy queries fuzzy
?
fuzzy string matching in elasticsearch way of saying "approximate string matching". not implemented using fuzzy logic.
lucene (the library underpinning elasticsearch , solr), implements "fuzzy" (approximate) search edit distance of 2 using finite state transducer representing union of possible transitions (including edits , deletes edit distance 1 or 2) between characters in every indexed term.
it's efficient data structure storing , tracing world of existing terms meet criteria entered. here's pic good article these.
(shows levenshtein automaton representing word "food" 2 edits.)
Comments
Post a Comment