Skip to content
This repository was archived by the owner on Apr 22, 2020. It is now read-only.

Conversation

@tomasonjo
Copy link
Collaborator

I have tested our new pageRank formula from [princeton course] and check comparison to apoc.(http://www.cs.princeton.edu/~chazelle/courses/BIB/pagerank.htm)... know that the results from princeton course do not have specified accurate number of iterations, so we cannot be exact precise. I am using 20 iterations in this examples.

Example 1
name | algos | apoc | princeton
A | 1.443 | 1.42 | 1.49
B | 0.760 | 0.75 | 0.78
C | 1.530 | 1.514 | 1.58
D | 0.15 | 0.15 | 0.15

Example 3
name | algos | apoc | princeton
Home | 3.232 | 3.215 | 3.35
Product | 1.059 | 1.054 | 1.1
Links | 1.059 | 1.054 | 1.1
About | 1.059 | 1.054 | 1.1
Site A | 0.328 | 0.327 | 0.34
Site B | 0.328 | 0.327 | 0.34
Site C | 0.328 | 0.327 | 0.34
Site D | 0.328 | 0.327 | 0.34

Example 5
name | algos | apoc | princeton
Home | 1.854 | 0.666 | 1.92
Product | 0.676 | 0.666 | 0.69
More | 0.676 | 0.666 | 0.69
About | 0.676 | 0.666 | 0.69

Princeton results are close but not exact, because I do not know exact iteration number, but looks like we are ok. There is a slight difference between apoc pagerank and algos pagerank using the same 20 iterations.

@jexp jexp merged commit 77cf0f0 into neo4j-contrib:3.1 Jun 12, 2017
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants