Recommender System Using User-Mediated Signal Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional recommender systems primarily send recommendations directly to users, which may not be as effective as personal recommendations from trusted contacts, and advertisers face challenges in reaching users amidst a crowded advertising space.
Innovation Solution
A recommender system that generates electronic suggestion signals for users to recommend items to their contacts based on the contacts' profiles, using a like-degree determiner to assess potential interest and send signals only if the similarity exceeds a certain threshold, allowing for semi-automatic personal recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If recommendations are sent directly to users by the system, then the system can provide recommendations efficiently, but the recommendations are less effective because users trust personal recommendations from contacts more than machine-generated recommendations
Solution Approach 1:
The patent introduces users as intermediaries who forward recommendations to their contacts. Instead of the system directly recommending items to all users, the system generates recommendations and leverages existing social relationships by having users act as mediators to distribute these recommendations within their social networks, thereby increasing trust and effectiveness
Solution Approach 2:
The patent segments the recommendation distribution process into two parts: the system generates recommendations based on algorithms, and then users personally forward selected recommendations to their contacts. This segmentation allows the system to maintain algorithmic efficiency while incorporating personal trust factors through user-mediated forwarding
2Productivity
If the system sends suggestions to all users for all items, then all users receive potential recommendations, but users are overwhelmed with unwanted information reducing engagement
Solution Approach 1:
The patent applies local quality by customizing recommendations based on individual user profiles, contact relationships, and historical behavior patterns. Each user receives a tailored set of suggestions relevant to their specific context and interests, rather than a generic list applicable to all users
Solution Approach 2:
The patent uses partial action by selectively generating and forwarding only those recommendations that meet specific criteria (user interest, contact relevance, item suitability) rather than flooding users with all possible recommendations. The system filters and prioritizes recommendations to provide a manageable subset
3Measurement precision
If the system tracks and processes detailed user feedback and contact relationships, then recommendation accuracy improves, but the data processing complexity and computational resources increase
Solution Approach 1:
The patent implements preliminary action by pre-processing and storing user profile data, contact relationship information, and item characteristics in structured formats before recommendation generation. This advance preparation enables faster and more accurate matching during the actual recommendation process without requiring complex real-time computations
Data Source
AI summary
The present invention is related to a recommender system (100), a computer-implemented recommending method, a corresponding computer readable medium and a corresponding computer program. A recommender system (100) is configured to send an electronic suggestion signal (186) to a respective user (199) of a user database (140) in dependence of a first like-degree (164) of a specific item, which has been determined based on items that the respective user has already recommended to his contacts. The electronic suggestion signal (186) suggests the respective user (199) to recommend the specific item to one or more of his contacts (198). In this way, personal based recommendations (188) are stimulated.