Scholar Positioning via Structured Affiliation Mining
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for precise positioning of scholars based on their scientific research achievements face challenges due to unstructured and diverse text information, leading to issues like information loss, ambiguity, and multi-language interference, resulting in low recall and accuracy rates, especially when extracting affiliation information.
Innovation Solution
A method and apparatus that extract key information from scholarly texts, construct structural information, mine implicit geographic information, and perform structural arrangements to achieve precise positioning by utilizing multiple map APIs and adaptive weight calculations to filter and determine reliable results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If map API is used to directly acquire mapping between geographic information and scholar's affiliation information, then positioning speed is improved, but recall rate and accuracy are low (50% recall, <80% accuracy)
Solution Approach 1:
The patent segments the affiliation information extraction process into multiple stages: initial map API lookup, candidate result generation, and verification against multiple knowledge bases. This segmentation allows the system to maintain speed while improving accuracy through progressive filtering and validation.
Solution Approach 2:
The patent introduces intermediary verification mechanisms including multiple knowledge bases and cross-validation processes between different map APIs. These intermediaries act as mediators to verify the accuracy of positioning results without significantly increasing processing time.
2Loss of information
If scholar's affiliation information is extracted from unstructured text, then information completeness is improved, but information loss and ambiguity increase
Solution Approach 1:
The patent implements feedback loops where extraction results are verified against multiple knowledge bases and used to refine subsequent extractions. The system continuously learns from verification results to improve extraction accuracy and reduce ambiguity in unstructured text processing.
Solution Approach 2:
The patent changes the parameters of information representation by transforming unstructured text into structured data with multiple validation attributes. This includes generating candidate affiliations with confidence scores and verifying against multiple knowledge bases to ensure reliability.
3Measurement precision
If multiple map APIs are used to improve positioning accuracy, then recall rate and accuracy are improved (91.72% recall, 98.34% accuracy), but system complexity increases
Solution Approach 1:
The patent merges multiple map APIs and knowledge bases into a unified positioning system with centralized coordination. The system combines results from multiple sources using a unified verification framework, reducing the complexity that would arise from managing separate independent systems.
Solution Approach 2:
The patent creates a universal positioning framework that can work with multiple map APIs and knowledge bases through a common interface. This multi-functional system handles different data formats and verification methods uniformly, reducing overall system complexity despite using multiple resources.
Data Source
AI summary
The present disclosure provides a method and an apparatus for precise positioning of a scholar based on mining of the scholar's scientific research achievement. The method includes: extracting text information in the scholar's scientific research achievement P to obtain key information, and constructing structural information; mining and constructing implicit information O with a geographic directivity in the scholar's scientific research achievement P according to the key information and the structural information; performing a structural arrangement on the structural information, and acquiring a final result R; and acquiring a mapping of A→R according to the final result R and the matrix U, acquiring and outputting the positioning information of the authors in the set A.