How to use Jaccard similarity algorithm in neo4j to find the similar nodes

Neo4j Graph Platform Procedures & APOC
9

Hi Michael,

In context to the Jaccard Algorithm usage -
My graph looks like this -
api_mining_jaccard_q1

[1]
Here when i run this query :

MATCH (s:claimIntimationRequestHeader)-[:Request]-(claimIntimationRequestBody)
WITH {item:id(s), categories: collect(id(claimIntimationRequestBody))} as userData
RETURN userData

I get this output -


userData
{
  "item": 1929,
  "categories": [
    1928
  ]
}

So using it like this -

WITH[{
  item: 1929,
  categories: [
    1928
  ]
}] as data 
CALL algo.similarity.jaccard.stream(data, {similarityCutoff:0.1}) YIELD item1, item2, count1, count2, intersection, similarity
RETURN item1, item2, count1, count2, intersection, similarity LIMIT 10

i get this output -

(no changes, no records)

So i guess because i have only one Item in the userData.

And when i do its count -


MATCH (s:claimIntimationRequestHeader)-[:Request]-(claimIntimationRequestBody) 
WITH {item:id(s), categories: collect(id(claimIntimationRequestBody))} as userData 
WITH collect(userData) as data 
CALL algo.similarity.jaccard.stream(data) YIELD item1, item2, count1, count2, intersection, similarity
RETURN count(*)

i get this output -

count(*)
0

[2]
When i try this cypher query :

MATCH (s:claimIntimationRequestBody)-[:Parameter]-(requestID) 
WITH {item:id(s), categories: collect(id(requestID))} as userData
RETURN userData

Output -

userData
{
  "item": 0,
  "categories": [
    1794,
    1793,
    1792,
    1791,
    1790,
    1789,
    1788,
    1787,
    1786,
    1785,
    1784,
    1783,
    1782,
    1781,
    1780,
    1779,
    1778
  ]
}

And then when i utilize this output with this query -

WITH[{
  item: 0,
  categories: [
    1794,
    1793,
    1792,
    1791,
    1790,
    1789,
    1788,
    1787,
    1786,
    1785,
    1784,
    1783,
    1782,
    1781,
    1780,
    1779,
    1778
  ]
}] as data 
CALL algo.similarity.jaccard.stream(data, {similarityCutoff:0.1}) YIELD item1, item2, count1, count2, intersection, similarity
RETURN item1, item2, count1, count2, intersection, similarity LIMIT 10

Output -

(no changes, no records)

Please help in creating a proper Jaccard algorithm query.