Link Prediction Model for Wireless Network and POI Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying the association relationship between a point of interest (POI) and a wireless network suffer from poor identification accuracy and low identification rates due to issues such as random SSID settings and linguistic differences.
Innovation Solution
A method and apparatus for data matching that involves acquiring target locations for target wireless networks, calculating relationship matching feature data between these networks and POIs, inputting this data into a link prediction model to determine corresponding POIs and wireless networks, and filtering out non-target wireless networks based on type discrimination information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to identify association relationship between POI and wireless network, then the process is simple, but the identification accuracy and rate are poor
Solution Approach 1:
The patent transforms the identification problem by changing parameters: instead of direct matching, it calculates relationship matching feature data including cosine similarity of name vectors, distance features, and type features. These transformed parameters are then input to a link prediction model to achieve accurate identification while maintaining reasonable complexity
Solution Approach 2:
The patent introduces a link prediction model as an intermediary between the raw data and the identification result. This intermediary processes the relationship matching feature data and outputs the association relationship, effectively resolving the contradiction between simple process and accurate identification
2Measurement precision
If comprehensive matching features are calculated to improve identification accuracy, then the accuracy improves, but the calculation complexity increases
Solution Approach 1:
The patent segments the comprehensive matching process into distinct feature calculations: cosine similarity of name vectors, distance features between POI and wireless network, and type features. Each segment is calculated independently and then combined, making the overall complex process more manageable and efficient
Solution Approach 2:
The patent calculates multiple types of feature data (name similarity, distance, type) which may seem excessive, but these partial features collectively provide comprehensive information for accurate identification. The link prediction model then processes these features to achieve the desired accuracy
Data Source
AI summary
A method for data matching, a readable medium and an electronic device are provided. The method includes: acquiring a target location corresponding to each target wireless network of a plurality of target wireless networks, and at least one point of interest to be matched within a preset range of the target location; calculating relationship matching feature data between each target wireless network and the at least one point of interest to be matched that corresponds to the target wireless network; inputting the relationship matching feature data into a preset link prediction model to determine a target point of interest corresponding to the target wireless network; and determining at least one of the target wireless networks corresponding to the target point of interest according to the relationship matching feature data and the first identification information of each target wireless network.