Your own recomendation Engine with Neo4j AND Reco4PHP
summary
Claims that Reco4PHP is a standalone recommendation skeleton built atop Neo4j and that Neo4j can traverse millions of relationships per core per second. Demonstration loads a GitHub activity dataset, uses Cypher queries as discovery engines, assembles weighted scores, applies blacklists and post‑processors, and removes the author’s own repo from results. CSV bulk import supports the initial load (~4–5 million rows/sec reported). Outcome is a working real‑time recommender prototype, not a drop‑in product.
AI-written summary · about
chapters (14)
- 0:00Start
- 1:26Recommendations : Overview
- 2:24Main types of recommendations
- 8:23Technical Challenges
- 11:39Relationships have a type
- 11:59Relationships can have properties
- 15:04Cypher : the query language for graphs
- 15:15Declarative pattern matching language
- 25:12Performance
- 27:17Architecture
- 28:37Design decisions
- 30:04Discover
- 32:28Post process
- 33:19Assemble
related talks
38mRecommendation Engine with Neo4j, Piece of Cake
01h 11mBuilding a Recommendation Engine with Neo4j and Python
01h 38mBuilding a recommendation engine with Python and Neo4j
29mA recommendation engine in your PHP applications
51mBuilding a Meetup Recommendation Engine with Spring Data Neo4j
27mUsing Graph Theory to Build a Recommendation Engine in Node js
21mUnlocking Real-Time News Recommendations: The Power of Graph Relationships in Neo4j
38mBuilding Recommender System in PHP8