Hybrid Ranking System

The Hybrid Ranking System is the core decision-making layer of dBOOK’s Recommendation Engine. it bridges the gap between the user’s decoded intention (from the Intelligence Engine) and the available protocols in the registry while balancing fairness, safety, and profitability.

Key Features

• Intent Matching

  • Identifies all protocols that potentially match the user’s intent

  • It takes the user’s intent—say, bridging funds—and matches it with protocols that provide related or complementary functions (like staking or liquidity provision).

• Multi-Factor Ranking

  • Each candidate is scored based on the following criteria:

    • Relevance

    • Safety

    • User cost

    • Profitability

  • This ensures that the top-ranked protocols are both effective and efficient.

• Environmental Awareness

  • Considers factors like network congestion, current gas fees, and protocol risk factor, ensuring that recommendations fit the present conditions

• Dynamic Adjustment

  • The system learns and adapts in real time

  • Adjusts rankings based on user clicks (e.g., if users ignore Aave, its relevance score decays)

  • All feedback is analyzed within the current session context, and no historical user data is retained