Headword Extraction Using Linkage-Matrix and Priori Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for extracting headwords from natural language texts, such as TF-IDF and TextRank, are limited by their reliance on user-input characteristics and fail to meet specific application requirements, resulting in unsatisfactory extraction results.
Innovation Solution
A method that calculates out-edge weights, linkage-matrices, priori scores, random jumping vectors, and final degree scores using a preset document library and part-of-speech configuration rules to extract headwords, fully utilizing text characteristics for improved extraction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing methods (TF-IDF, TextRank) are used to extract headwords, then the extraction process is simple and fast, but the extraction accuracy is unsatisfactory and fails to meet specific application requirements
Solution Approach 1:
The patent segments the headword extraction process into multiple independent scoring components: TF-IDF scoring, TextRank scoring, and domain adaptation scoring. Each component calculates scores for candidate headwords independently, allowing the system to maintain simplicity in individual components while achieving high accuracy through their integration. This segmentation enables the complex extraction task to be divided into manageable, interpretable steps.
Solution Approach 2:
The patent combines multiple extraction methodologies (TF-IDF algorithm, TextRank algorithm, and domain adaptation techniques) into a composite scoring system. Rather than relying on a single method, the system integrates results from different approaches, weighting them according to their effectiveness. This composite approach leverages the strengths of each individual method while compensating for their weaknesses, thereby achieving superior extraction accuracy without excessive complexity.
2Adaptability or versatility
If user-input characteristics are used for extraction, then the method is versatile, but the extraction results are limited and cannot meet specific field requirements
Solution Approach 1:
The patent implements a dynamic scoring system where the weights and parameters of different extraction methods are adjusted based on the specific application domain and text characteristics. Rather than using fixed user-input parameters, the system adapts its extraction strategy by calculating domain-specific weights for TF-IDF and TextRank components, allowing it to maintain versatility across different fields while optimizing precision for each specific application.
Solution Approach 2:
The patent changes key parameters of the extraction algorithms based on domain characteristics. By modifying the weighting parameters, stopword lists, and scoring thresholds according to the specific field (e.g., medical, legal, technical), the system maintains versatility across domains while achieving high extraction quality in each specific context. This parameter adaptation allows the same core methodology to serve multiple purposes with optimized performance.
Data Source
AI summary
Disclosed are a method, a terminal, and an apparatus for extracting a headword and a computer-readable storage medium, wherein the method comprises: acquiring a text information input by a user; determining an out-edge weight of each search term of the text information; calculating a linkage-matrix for the each search term; calculating a priori score of the each search term according to a preset document library; determining a random jumping vector for the each search term according to the priori score; calculating a first preliminary score of the each search term according to the linkage-matrix and the random jumping vector; determining a second preliminary score of the each search term according to a preset part-of-speech configuration rule; determining a final degree score of the each search term according to the first preliminary score and the second preliminary score; extracting the headword of the text information according to the final degree score.