際際滷

際際滷Share a Scribd company logo
Drug Design Recommendation commendo research & consulting GmbH
Project Idea / Main Targets R ecommender technology to optimize the drug   design process Speeding up development Highly customized drugs Minimization of adverse reactions Increasing probability to pass clinical tests Discover harmful drug combinations to minimize   side effects
Business Opportunity Market: Pharmaceutical Industry Long term success by   accelerating research in    drug design Reduce pharmaceutical    development costs >800 mio. $ for bringing one drug on the market 95% rejection rate Source: Burrill & Company
Data Relationship Accurate, customized recommender models can be applied Recommender analogy: User   Item Effect   Substance 0 0 1 1 1 0 0 0 1 1 0 1 0 0 1 1 0 1 0 0 1 1 0 Sub. 1 Sub. 3 Sub. 5 Sub. 4 Sub. 7 Sub. 8 Sub. 6 Sub. 2 Sub. 9 Sub. 11 Sub. 13 Sub. 12 Sub. 15 Sub. 14 Sub. 10 Effect 0 Effect 1 Effect 2 Effect 3
Holistic Model Integration of all data Considering: substances & effects  From all drug development stages Use of recommender algorithms   to    predict unknown effects of substances Identification of substances with desired  effects to develop customized drugs
The Team & the Netflix Prize Georg Preler, Andreas T旦scher, Michael Jahrer, Michael Schrotter Leading single team at Netflix Prize competition
Thank you for your attention commendo research & consulting GmbH

More Related Content

Commendo @ Strands $100K Call for Recommender Start-Ups - RecSys08

  • 1. Drug Design Recommendation commendo research & consulting GmbH
  • 2. Project Idea / Main Targets R ecommender technology to optimize the drug design process Speeding up development Highly customized drugs Minimization of adverse reactions Increasing probability to pass clinical tests Discover harmful drug combinations to minimize side effects
  • 3. Business Opportunity Market: Pharmaceutical Industry Long term success by accelerating research in drug design Reduce pharmaceutical development costs >800 mio. $ for bringing one drug on the market 95% rejection rate Source: Burrill & Company
  • 4. Data Relationship Accurate, customized recommender models can be applied Recommender analogy: User Item Effect Substance 0 0 1 1 1 0 0 0 1 1 0 1 0 0 1 1 0 1 0 0 1 1 0 Sub. 1 Sub. 3 Sub. 5 Sub. 4 Sub. 7 Sub. 8 Sub. 6 Sub. 2 Sub. 9 Sub. 11 Sub. 13 Sub. 12 Sub. 15 Sub. 14 Sub. 10 Effect 0 Effect 1 Effect 2 Effect 3
  • 5. Holistic Model Integration of all data Considering: substances & effects From all drug development stages Use of recommender algorithms to predict unknown effects of substances Identification of substances with desired effects to develop customized drugs
  • 6. The Team & the Netflix Prize Georg Preler, Andreas T旦scher, Michael Jahrer, Michael Schrotter Leading single team at Netflix Prize competition
  • 7. Thank you for your attention commendo research & consulting GmbH