Synonym Tag Mining via Central Morpheme Templates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current synonym tag mining methods rely on sparse data resources, leading to lower coverage rates and reliability in obtaining synonym tags, particularly in vertical fields like movie searches, where user intentions are not accurately met due to insufficient recall of search results.
Innovation Solution
A method and apparatus that obtain synonym tags by identifying central morphemes and tag templates, using clusters of synonymy tag templates to expand queries, thereby improving the reliability of synonym tag retrieval without relying on click logs or bilingual parallel alignment data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If synonym tag mining depends on sparse mining resources such as click logs with paraphrase relationships or bilingual parallel alignment corpus, then the complexity of data requirements is reduced, but the coverage rate and reliability of synonym tags deteriorate
Solution Approach 1:
The patent introduces tag templates as an intermediary structure between the original sparse data and the target synonym tags. Tag templates serve as a mediating framework that organizes and generalizes mining resources, enabling the system to derive synonym tags even when direct paraphrase relationships or bilingual corpora are sparse. This intermediary structure allows for more reliable synonym tag generation by providing a systematic approach to data utilization.
Solution Approach 2:
The patent transforms the approach by changing the parameter of data utilization from direct use of sparse click logs or bilingual corpora to using structured tag templates with central morphemes. This parameter change enables the system to work effectively with limited data by organizing it into a template-based framework that can generalize across different contexts, thereby improving synonym tag reliability without requiring complex data resources.
2Ease of manufacture
If synonym tag mining uses sparse data resources, then the ease of data acquisition is improved, but the coverage rate of synonym tags deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-establishing tag templates with central morphemes before the actual synonym tag mining process. These pre-defined templates serve as a prepared framework that can be applied to various mining scenarios, allowing the system to achieve better coverage rates even when working with sparse data resources. The preliminary structure enables more effective utilization of limited data.
Solution Approach 2:
The patent changes the parameter of data structure from raw sparse data to organized tag templates, enabling easier data acquisition while maintaining high coverage rates. By transforming the data representation into a template-based format with central morphemes, the system achieves both ease of data acquisition and improved synonym tag coverage.
3Device complexity
If the user query does not match search results due to insufficient synonym mapping, then the simplicity of the search system is maintained, but the user experience and search recall deteriorate
Solution Approach 1:
The patent introduces tag templates as an intermediary mechanism between user queries and search results. These templates act as a bridge that expands and refines user queries by incorporating synonym relationships, thereby improving search recall without significantly increasing system complexity. The intermediary structure enables the system to understand and match diverse query formulations with relevant search results.
Solution Approach 2:
The patent applies universality by designing tag templates with central morphemes that can serve multiple functions: query expansion, synonym mapping, and search result enhancement. This multi-functional approach allows the same template structure to improve various aspects of the search system simultaneously, maintaining simplicity while enhancing user experience and search recall across different query types.
Data Source
AI summary
The present disclosure provides a synonym tag obtaining method and apparatus, a device, and a computer readable storage medium. By obtaining the central morpheme of the tag to be recognized and then obtaining the tag template of the tag to be recognized according to the tag to be recognized and the central morpheme of the tag to be recognized, to obtain the synonymy tag of the tag to be recognized according to the tag template, and the central morpheme of the tag to be recognized and at least one cluster of synonymy tag templates, it is possible to avoid the technical problem about reduction of the coverage rate of the synonymy tags caused by data sparsity of the mining resources and thereby improve the reliability of obtaining synonymy tags, without depending on mining resources such as a click log having a paraphrase relationship or data having bilingual parallel alignment corpus.