Liking cljdoc? Tell your friends :D

clj-similar.core


approximate-jaccard-indexclj

(approximate-jaccard-index minhash sig1 sig2)
source

build-treeclj

(build-tree points size)
source

index-setclj

(index-set v->idx s)
source

index-setsclj

(index-sets v->idx coll)

Given a mapping of values to indexes and a collection of sets, returns a new collection of sets where each set has have their values replaced by indexes

Given a mapping of values to indexes and a collection of sets,
returns a new collection of sets where each set has have their
values replaced by indexes
sourceraw docstring

jaccard-indexclj

(jaccard-index s1 s2)
source

nearestclj

(nearest similar s)
(nearest similar
         s
         n
         &
         {:keys [threshold exact?] :or {threshold 0.0 exact? true}})

Given a similar data structure and a target set s, finds the nearest matching set. Optionally takes a parameter n for the n nearest sets. Pass threshold as an optional arguments to filter elements with a jaccard index below the threshold. By default this is calculated exactly, pass exact? false to use the approximate MinHash jaccard index instead. Returning sets have distance metrics and vector associated as metadata.

Given a `similar` data structure and a target set `s`, finds the
nearest matching set. Optionally takes a parameter `n` for the `n`
nearest sets. Pass `threshold` as an optional arguments to filter
elements with a jaccard index below the threshold. By default this
is calculated exactly, pass `exact? false` to
use the approximate MinHash jaccard index instead. Returning sets have distance
metrics and vector associated as metadata.
sourceraw docstring

pointsclj

(points hash-fn indexed-sets)
source

Similarclj

source

similarclj

(similar coll)
(similar coll error)

Constructs a new similar set from a collection of sets. Can be used for lookup of nearest sets using nearest. Optionally takes a target similarity estimate error (default 0.1).

Constructs a new similar set from a collection of sets.
Can be used for lookup of nearest sets using `nearest`.
Optionally takes a target similarity estimate error (default 0.1).
sourceraw docstring

similarityclj

(similarity exact? minhash sig1 sig2 s1 s2)
source

value-indexclj

(value-index v)
source

cljdoc is a website building & hosting documentation for Clojure/Script libraries

× close