Cross-Platform User Identity Matching via Feature Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cross-platform data fusion is hindered by differences in user content across social networks, limiting big data analysis to a single platform and preventing full utilization of user identity matching, due to varying content themes, densities, and scales between platforms.
Innovation Solution
A method and device for cross-platform data matching that involves receiving a data matching request, obtaining group behavior data, performing behavior learning to generate feature distribution functions, calculating maximum entropy values, and determining matching users across different social networks to achieve user identity matching and integration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If user identity matching is performed across different social network platforms, then cross-platform data fusion capability is improved, but the difficulty of matching increases due to differences in content themes, densities, and scales between platforms
Solution Approach 1:
The patent transforms user behavior data into standardized feature vectors by changing the parameter representation from raw platform-specific content to normalized behavioral features. This allows users from different platforms with different content characteristics to be compared using a common feature space, resolving the matching difficulty while maintaining cross-platform adaptability
Solution Approach 2:
The patent introduces behavior feature vectors as an intermediary representation layer between raw user data from different platforms and the matching algorithm. This intermediary transforms diverse platform-specific content into a unified feature representation that facilitates cross-platform comparison while accounting for platform differences
2Measurement precision
If behavior learning is performed on group behavior data to generate feature distribution functions, then matching accuracy is improved, but the computational complexity and time consumption increase
Solution Approach 1:
The patent performs behavior learning and generates feature distribution functions in advance before the actual matching process. By pre-computing the feature distributions from group behavior data, the system avoids repeated computational operations during matching, thereby improving accuracy while reducing real-time computational time consumption
Solution Approach 2:
The patent divides the matching process into distinct stages: first learning behavior patterns from group data to create feature distributions, then using these pre-established distributions for matching. This segmentation allows complex behavior learning to be performed separately from the time-critical matching operation, balancing accuracy and efficiency
Data Source
AI summary
A method of matching cross-platform data, comprising: receiving a data matching request sent by a terminal; obtaining a group behavior data corresponding to the first user group in the first social network platform, and learning the group behavior data to obtain a group feature distribution function; obtaining associated users of the designated root node users and corresponding behavior data in the second social network platform; learning the behavior data of the root node users, and generating the group feature distribution function after matching the root node users; performing the behavior learning to the behavior data of the associated users; calculating a maximum entropy value of the group feature distribution function after matching the associated users, and determining the associated users corresponding to the largest maximum entropy value as the matching users of the first user group; and regarding the determined matching users as current root node users, determining a next matching user until the determined matching users meet a set quantity condition, and completing a group matching.