Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Recommendation Engine for wide transactions

Recommendation Engine for wide transactions

E15a196892a14ce427ab468a470de640?s=128

harjinder-hari

June 09, 2017
Tweet

Transcript

  1. Rec Sys - wide transactions Harjinder Mistry Red Hat |

    @hmistry
  2. Agenda 1. RecSys - 2 min primer 2. Problem -

    Definition 3. Challenges in Standard Approaches 4. Our approach & architecture
  3. RecSys examples

  4. Basic terminologies user-item matrix explicit vs implicit feedback — user-user

    — user-item — item-item image source
  5. Frequent Pa!ern mining Applications — Customer Analysis — Brick-and-mortar retail

    — Handling cold-start situation — Retrieval
  6. Frequent Pa!ern mining Algorithms — apriori — FP Growth

  7. openshi!.io

  8. Helping developers become more efficient recommendations on packages recommendations on

    the stack
  9. Input data Projects/stacks - from code repositories — Java (pom.xml)

    — Node.js (packages.json) — Python (requirements.txt)
  10. spark, elastic cloud compute.... cool - let's rock

  11. developers are amazing - but, of course

  12. Wide transactions - challenges — existing methods didn't work —

    time to train was huge — memory issues
  13. As a self-serve platform, turnaround time as important as accuracy

  14. Matrix Factorization is fast image source

  15. Let's use matrix factorization (ALS) to generate frequent pa!erns

  16. Step 1: Train ALS model

  17. Step 2: Generate initial seed: random candidate set

  18. Step 3: Find recommended product(package)

  19. Step 4: Add it to the frequent pa!ern list and

    continue
  20. None
  21. Why not deep learning?

  22. Code, Slides and Contact ____ Code will be open-sourced soon!

    Harjinder Mistry email:hmistry@redhat.com