The Office of Technology Licensing was established in 1970 to transfer technologies developed at Stanford. Find out more about OTL's history, mission, staff, and statistics.

Login to TechFinder » 

Mathematical model and algorithm for computing personalized trust/compatibility scores between two individuals based on their profiles and social networks.


Stanford Reference:

14-030


Abstract


Stanford researchers have developed and tested a novel algorithm, which computes a personalized trust-score between two online users to increase trust during online transactions. This trust score is based on users’ social network, social activity (likes, mentions, posts, tweets) and social profiles. The two users do not have to be linked directly to compute a score. Moreover, the researchers have also developed an algorithm that determines the parameters of the transaction (transaction fee, commission, warranty, other data that the buyer/seller might be required to provide) based on this score. It can be used for personalized ranking of online reviews, personalized recommendations of products/services and personalized matching/scoring of buyers and sellers in an online marketplace.

Youtube demonstration of teapot - application of this platform

Stage of Research:
  • Algorithm used for roommate recommendation for Stanford graduate housing for fall 2014. (https://teapot-roommates.stanford.edu -- accessible only to Stanford affiliates).
  • Reduced to practice- Launched teapot A Trust-as-a-Service platform for online marketplaces, sharing economy services, and social applications.

  • Applications


    • Personalized ranking of reviews of hotels/tour packages (e.g. on travel websites), products sold by online retailers, and restaurants
    • Personalized recommendation of products/services to users
    • Personalized reputation/ranking of buyers/sellers to each other in online marketplaces, and sharing economy services (e.g. roommate matching, subletting)

    Advantages


    • Provides personalized, quantitative score to increase trust between two online users
    • More complex and comprehensive method - Existing methods only use direct friends/connections, or at most two hop connections in the social network. This method generalizes to any distance between two users in the social network
    • Determines transaction parameters as a function of the trust score
    • Aims to improve the online transaction experience

    Publications



    Related Web Links



    Innovators & Portfolio


    • Pranav Dandekar   
    • Ashish Goel   

    Date Released

     6/12/2014
     

    Licensing Contact


    Evan Elder, Licensing Associate
    650-725-9558 (Mobile)
    Login to Request Information

    [-] Map/Timeline

    99-050 Recall
    02-009 Distributed Internet Reverberation for Acoustic Collaboration (DIRAC)
    02-044 Real-Time Optimized Reconstruction Algorithm for Adaptive Imaging

    more technologies

    Related Keywords


    trust   social network   personalized trust score   e-commerce   algorithm   computational algorithm   internet advertising   internet search   internet: information sharing   internet: search tools   online   online advertising   online assessment   online auction   ranking algorithm   social commerce   social simulation   social software   social software commerce   social technology   software: social interaction   software: social simulation   online games   online payment   
     

       

      

    Also of Interest...
    99-050 Recall
    02-009 Distributed Internet Reverberation for Acoustic Collaboration (DIRAC)
    02-044 Real-Time Optimized Reconstruction Algorithm for Adaptive Imaging

    Recently Viewed...
    S14-030 Mathematical model and algorithm for computing personalized trust/compatibility scores between two individuals based on their profiles and social networks.