Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

RecSys Multi-Stack Ensemble for Job Recommendation, Pumpkin-Pie

803 views

Published on

Slides used @ RecSys Conference 2016 by Team Pumpkin-Pie

Published in: Data & Analytics
  • Be the first to comment

RecSys Multi-Stack Ensemble for Job Recommendation, Pumpkin-Pie

  1. 1. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Team Pumpkin-Pie T. Carpi M. Edemanti E. Kamberoski E. Sacchi R. Pagano P. Cremonesi M. Quadrana
  2. 2. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Outline • Collaborative Filtering • Content-Based • Interactions and Impressions • Multi-Stack Ensemble • Linear Ensemble • Evaluation Score Ensemble • Conclusions
  3. 3. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Collaborative-Filtering IDF value as a rate for the job
  4. 4. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Collaborative-Filtering User-based
  5. 5. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Collaborative-Filtering Item-based
  6. 6. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Content-Based Concept-based IDF value as a weight for the tag
  7. 7. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Content-Based Concept-based joint User-Item similarity
  8. 8. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Interactions & Impressions Click ReplyBookmark Already clicked jobs are likely to be clicked again
  9. 9. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Interactions & Impressions A B C D E F B C DA B D AC Dataset Filtering-step Reordering-step
  10. 10. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Multi-Stack Ensemble Past Interactions Ens CF Past Impressions UBCF 
 IntInt UBCF IntImp UBCF ImpImp UBCF ImpInt IBCF 
 IntInt IBCF ImpImp CBJUIS CBIS Baseline Ens 
 CF+CB Ens CB Final Ensemble
  11. 11. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Multi-Stack Ensemble Ens CF UBCF 
 IntInt UBCF IntImp UBCF ImpImp UBCF ImpInt IBCF 
 IntInt IBCF ImpImp CBJUIS CBIS Ens CB Stack Level 1
  12. 12. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Multi-Stack Ensemble Ens CF UBCF 
 IntInt UBCF IntImp UBCF ImpImp UBCF ImpInt IBCF 
 IntInt IBCF ImpImp CBJUIS CBIS Ens 
 CF+CB Ens CB Stack Level 2
  13. 13. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Multi-Stack Ensemble Past Interactions Ens CF Past Impressions UBCF 
 IntInt UBCF IntImp UBCF ImpImp UBCF ImpInt IBCF 
 IntInt IBCF ImpImp CBJUIS CBIS Baseline Ens 
 CF+CB Ens CB Final Ensemble Stack Level 3
  14. 14. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Linear Ensemble weight for algorithm a rank of item i in algorithm a decay for algorithm a
  15. 15. Politecnico di Milano Team Pumpkin-Pie09-15-2016 red 1.9990 pink 1.9980 yellow 1.9970 blue 1.9960 green 1.9985 orange 1.9970 purple 1.9955 red 1.9940 Algorithm A Weight 2 Decay 0.001 Algorithm B Weight 2 Decay 0.0015 Linear Ensemble 1 2 3 4 1 2 3 4
  16. 16. Politecnico di Milano Team Pumpkin-Pie09-15-2016 pink 1.9980 yellow 1.9970 blue 1.9960 green 1.9985 orange 1.9970 purple 1.9955 Algorithm A Weight 2 Decay 0.001 Algorithm B Weight 2 Decay 0.0015 Linear Ensemble red 3.9930
  17. 17. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Linear Ensemble Past Interactions Ens CF Past Impressions UBCF 
 IntInt UBCF IntImp UBCF ImpImp UBCF ImpInt IBCF 
 IntInt IBCF ImpImp CBJUIS CBIS Baseline Ens 
 CF+CB Ens CB Final Ensemble
  18. 18. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Evaluation-Score Ensemble leaderbord score # of elements
  19. 19. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Evaluation-Score Ensemble red 7.5660 pink 7.5660 purple 5.5660 orange 5.5660 yellow 9.4575 green 9.4575 blue 6.9575 grey 6.9575 Algorithm A l_a 200k n_a 1 Mln Weight 0.2 Algorithm B l_b 200k n_b 800k Weight 0.25
  20. 20. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Evaluation-Score Ensemble red 7.5660 pink 7.5660 purple 5.5660 orange 5.5660 yellow 9.4575 green 9.4575 blue 6.9575 grey 6.9575 Algorithm A l_a 200k n_a 1 Mln Weight 0.2 Algorithm B l_b 200k n_b 800k Weight 0.25
  21. 21. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Evaluation-Score Ensemble Past Interactions Ens CF Past Impressions UBCF 
 IntInt UBCF IntImp UBCF ImpImp UBCF ImpInt IBCF 
 IntInt IBCF ImpImp CBJUIS CBIS Baseline Ens 
 CF+CB Ens CB Final Ensemble
  22. 22. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Conclusions Competition promotes the algorithms that learn which are the best items among the ones recommended by the Xing platform
  23. 23. Politecnico di Milano Team Pumpkin-Pie09-15-2016 Conclusions Multi-Stack Ensemble 4th Place Team Pumpkin-Pie

×