Permission-Based Feature Selection for Secure Data Recommendation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning models for data recommendation face inefficiencies and accuracy issues due to incomplete feature data availability and user reluctance to share certain data, leading to unsatisfactory performance.
Innovation Solution
A method and apparatus that obtain first feature data, specify a permission type for data usage, generate second feature data based on this type, and determine matching data items from a dataset, ensuring data security and flexibility in data recommendation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If comprehensive feature data is used in machine learning models for data recommendation, then recommendation accuracy is improved, but data security risks and user privacy concerns increase
Solution Approach 1:
The patent segments the feature data into multiple parts based on user permission types. Instead of using all comprehensive feature data, the system selectively uses only the permitted portions of feature data corresponding to the user's permission level, thereby balancing recommendation accuracy with data security and privacy concerns
Solution Approach 2:
The patent applies different quality levels of feature data to different users based on their permission types. Each user receives recommendation services tailored to their specific permission scope, allowing high accuracy for users with broader permissions while maintaining security for users with restricted permissions
2Measurement precision
If more feature data is collected to improve recommendation performance, then model accuracy improves, but user reluctance to share data increases
Solution Approach 1:
The patent introduces a dynamic permission type mechanism that allows users to control the scope of their data sharing. The system adapts to user preferences by adjusting which feature data is collected and processed based on real-time permission settings, making users more willing to share data selectively rather than comprehensively
Solution Approach 2:
The patent changes the parameter of data collection from comprehensive to selective based on permission types. By adjusting which feature data is collected according to user permission settings, the system maintains model accuracy while reducing user burden and increasing data sharing willingness
3Adaptability or versatility
If feature data processing is made more flexible to accommodate user permissions, then system adaptability improves, but processing complexity increases
Solution Approach 1:
The patent performs preliminary classification of feature data into different permission types before the recommendation process. By pre-organizing feature data according to permission categories, the system simplifies the processing complexity while maintaining high flexibility in adapting to different user permission settings
Data Source
AI summary
A method, an apparatus, a device and a medium for recommending data are provided. In a method, first feature data of an object is obtained. A permission type for using the first feature data is obtained, the permission type specifying a portion of the first feature data allowed to be used in data recommendation. The first feature data is updated based on the permission type to generate second feature data; and based on the second feature data, a group of data items matching the second feature data is determined from a data set including a plurality of data items.