Relevance

  • Discovery and relevance is key to matching buyers and sellers.

Searching

  • For performance and scalability reasons, search is performed off-chain on cloud servers via REST API.

  • All user profiles are cached for performance, and indexed for fast searching. For the initial phase, we will use the good old-fashion PostgreSQL query.

Relevance Scoring

  • Relevance can be calculated per user given the following matching criteria:

    • Keywords
    • Skills
    • Interests
    • Availability

  • Relevance is scored as a positive floating number from 0.0 (least) to 1.0 (most) and is calculated based on following factors:

    • Match score of keywords
    • Match score of skills weighted by endorsements
    • Match score of intersts
    • Match score of availability
    • Reputation of seller

    • Trusted by buyer

    • Relevance Score = normalizedScore((KeywordWeight * KeywordMatchScore) + (SkillWeight * SkillMatchScore) + (InterestWeight * InterestMatchScore) + (AvailabilityWeight * AvailabilityMatchScore) + (ReputationWeight * ReputationScore) + (TrustWeight * TrustScore))