Friend Recommendation Blacklist Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current friend recommendation methods often recommend users who have been rejected multiple times, leading to inefficiency and user resentment, as they fail to adapt to user preferences effectively.
Innovation Solution
A friend recommendation method and apparatus that creates a recommendation blacklist to filter out unwanted friends and adjusts familiarity scores based on recommendation time, ensuring only relevant friends are suggested, thereby enhancing user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the system recommends objects with high familiarity scores repeatedly, then the initial recommendation coverage is improved, but the recommendation efficiency deteriorates when users consistently reject the same objects
Solution Approach 1:
The patent extracts problematic recommended objects that users consistently reject and places them into a blacklist. This separates the harmful recommendation loop from the main recommendation process, preventing wasted recommendations on uninterested users while maintaining high familiarity scoring for other potential friends.
Solution Approach 2:
The patent dynamically adjusts the recommendation process by updating the blacklist based on user feedback. The system transitions from static repeated recommendations to dynamic adaptive recommendations, where the recommendation pool changes based on user acceptance or rejection patterns.
2Quantity of substance
If the system continues to recommend rejected objects, then the recommendation pool remains large, but user experience deteriorates due to repeated unwanted suggestions
Solution Approach 1:
The patent converts the harmful effect of user rejections into a beneficial filtering mechanism. Each rejection is used to update the blacklist, transforming negative user feedback into positive recommendation quality improvement by eliminating unwanted suggestions.
Solution Approach 2:
The patent implements a feedback loop where user acceptance or rejection of recommended objects is recorded and used to update the blacklist. This closed-loop system continuously learns from user behavior to improve recommendation quality and reduce user resentment.
3Adaptability or versatility
If the system recommends all high familiarity objects, then comprehensive coverage is achieved, but the time to find suitable friends increases
Solution Approach 1:
The patent performs preliminary filtering by maintaining a blacklist of rejected objects before the actual recommendation process. This pre-filtering eliminates unwanted candidates in advance, so users are not presented with rejected objects, saving time and improving the efficiency of finding suitable friends.
Data Source
AI summary
Disclosed is a friend recommendation method, comprising: creating a recommendation backlist for a user; deleting a recommended object included in the recommendation backlist of the user from a recommended object list; and recommending remaining recommended objects in the recommended object list to the user. The present application further discloses a server and a storage medium, so as to recommend friends based on the requirements of a user and recommend objects that the user is really interested in. Therefore, the accuracy and efficiency of friend recommendation are improved, friends can be recommended effectively and the user experience is enhanced.