Keyword Extraction Using Word Association Topology
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current keyword extraction methods are subjective, inaccurate, and time-consuming, often relying on statistical features or labeled corpora, and struggle to ensure semantic association with the text, especially in articles with multiple topics, and are limited in their ability to generalize semantics beyond the text content.
Innovation Solution
A keyword extraction method that involves receiving an original document, extracting candidate words, determining association degrees, and using a word association topology to select keywords, which allows for the selection of keywords that may not be present in the document but are semantically relevant, thereby improving accuracy and reducing calculation complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If statistical features or labeled corpora are used for keyword extraction, then the extraction process can be automated, but the accuracy and semantic association with the text deteriorate
Solution Approach 1:
The patent introduces an association degree calculation mechanism as an intermediary between automated processing and semantic accuracy. By calculating association degrees between candidate keywords and document semantics using vector space models, the system bridges the gap between automation and precision, selecting keywords that both automate the process and maintain high semantic relevance
Solution Approach 2:
The patent transforms the keyword extraction problem by changing parameters from simple frequency counts to association degree measurements. By using cosine similarity and vector space representations, the system converts traditional statistical features into semantic-based parameters, improving accuracy while maintaining automation
2Device complexity
If traditional keyword extraction methods are used, then the process is simpler, but the time consumption increases and productivity decreases
Solution Approach 1:
The patent segments the keyword extraction process into distinct stages: candidate keyword generation, association degree calculation, and final selection. This segmentation allows each stage to be optimized independently, improving overall productivity while maintaining manageable complexity through modular processing
Solution Approach 2:
The patent performs preliminary actions by pre-calculating document vectors and candidate keyword vectors before the actual extraction. This preliminary vectorization and clustering work enables faster association degree calculations during the extraction phase, reducing overall time consumption while maintaining simplicity
3Speed
If high-frequency words are prioritized for keyword extraction, then the extraction process is faster, but the semantic generalization ability deteriorates
Solution Approach 1:
The patent fundamentally changes the selection parameter from word frequency to association degree with document semantics. By using cosine similarity between candidate keywords and document vectors, the system prioritizes semantically relevant words over high-frequency words, improving semantic generalization while maintaining extraction speed through efficient vector calculations
Solution Approach 2:
The patent transitions from a one-dimensional frequency-based selection to a multi-dimensional vector space approach. By representing documents and keywords as vectors in semantic space and calculating associations across multiple dimensions, the system achieves better semantic generalization without sacrificing speed
4Measurement precision
If word association topology is introduced to improve semantic relevance, then the accuracy of keyword extraction improves, but the device complexity increases
Solution Approach 1:
The patent uses vector space models and association degree calculations as intermediaries between the word association topology and the final keyword selection. This intermediary layer simplifies the complex topology into manageable similarity scores, improving accuracy while controlling system complexity through mathematical abstraction
Solution Approach 2:
The patent transforms the complex word association topology into a simpler parameter - the association degree score. By converting topological relationships into quantitative similarity measurements, the system maintains high accuracy in keyword selection while reducing the apparent complexity of the extraction system
Data Source
AI summary
A keyword extraction method includes: extracting candidate words from an original document to form a first word set; acquiring a first association degree between each first word thereof and the original document, and determining a second word set according to the first association degree; for each second word in the second word set, inquiring, in a word association topology, at least one node word satisfying a condition of association with the second word and forming a third word set, the word association topology indicating an association relation among multiple node words in a predetermined field; and determining a union set of the second and third word sets, acquiring a second association degree between each candidate keyword in the union set and the original document, and selecting, according to the second association degree, at least one candidate keyword from the union set, to form a keyword set of the original document.