User Recommendation System Using Bayesian Matching Success Rate

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional user recommendation methods for social networking applications are complex and inefficient, requiring intricate six-degree relationship calculations and interest graph model establishment, which hampers recommendation performance and efficiency.

Innovation Solution

A user recommendation method and system that calculates a matching success rate for each user based on social networking quality data using a Bayesian method, simplifying the data model and calculation process, and recommending users with the highest matching success rate.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional user recommendation methods using six degree spatial theory or interest graph models are used, then user recommendation functionality is provided, but the implementation complexity increases and recommendation efficiency decreases

Engineering Contradiction:

Improverecommendation performanceVSAvoidimplementation complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent changes the fundamental parameters of the recommendation system by switching from complex graph-based models (six degree spatial theory, interest graph) to a simplified data model using user social networking quality data. This parameter change maintains recommendation functionality while dramatically reducing implementation complexity and improving efficiency

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent extracts only the essential elements needed for recommendation (user social networking quality data) and discards the complex computational frameworks (six degree relationships, interest graph models). This extraction approach retains core recommendation functionality while eliminating unnecessary complexity

Inventive Principle:

Principle #2Taking out (Extraction)

2Adaptability or versatility

If conventional user recommendation methods using six degree spatial theory are used, then user recommendation functionality is provided, but the calculation process becomes complex and time-consuming

Engineering Contradiction:

Improverecommendation functionalityVSAvoidcalculation time

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent uses simple, easily computable user social networking quality data instead of complex, time-consuming six degree relationship calculations. This approach treats the recommendation calculation as a simple, fast operation that can be performed efficiently without requiring elaborate computational processes

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

Data Source

PatentUS9519684B2User recommendation method and a user recommendation system using the same

Publication Date: 2016.12.13 TENCENT TECHNOLOGY (SHENZHEN) CO LTD

AI summary

A user recommendation method for supporting a social networking application includes receiving a user recommendation triggering command from a user at a mobile terminal; generating a recommended candidate user list based on the user recommendation triggering command; reading user social networking quality data, and calculating a matching success rate for each user in the recommended candidate user list based on the user social networking quality data; and selecting at least one user with a highest matching success rate from the recommended candidate user list for recommendation. By implementing the user recommendation method, recommendation performance and recommendation efficiency in the social networking application are improved. In addition, a user recommendation system implemented with the user recommendation method is also provided.