Knowledge Graph Trustworthy Search Method
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search engine evaluation methods lack comprehensive and reliable measures for assessing the trustworthiness of Web page content, relying heavily on external factors and failing to adequately consider the authenticity and reliability of entities and relations within the content.
Innovation Solution
A trustworthy search method based on a knowledge graph, which involves constructing a knowledge graph for selected Web page results, matching it with a reliable knowledge graph library, and computing a content support degree using fact, trust chain, and trust domain modes to provide a comprehensive evaluation of content trustworthiness, thereby sorting search engines for reliability and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If content trustworthiness is computed by external factors (links, click rates), then evaluation is simple and fast, but authenticity and reliability of content are not adequately measured
Solution Approach 1:
The patent introduces a knowledge graph as an intermediary between web page content and trustworthiness evaluation. The knowledge graph extracts entities, relations, and attributes from content, enabling semantic analysis and authenticity verification without relying solely on external factors like links or click rates.
Solution Approach 2:
The patent replaces traditional mechanical evaluation methods (counting links, tracking clicks) with semantic analysis and natural language processing technologies. This substitution enables deeper understanding of content authenticity through entity-relation analysis and semantic consistency verification.
2Measurement precision
If subjects in Web page content are extracted and analyzed using clustering and neural networks, then subject concentration analysis is improved, but trustworthiness of entities and relations is not measured
Solution Approach 1:
The patent segments content analysis into distinct components: entity extraction, relation extraction, and attribute extraction. Each component is processed separately through the knowledge graph, allowing targeted trustworthiness evaluation for entities and relations while maintaining subject concentration analysis.
Solution Approach 2:
The patent adds a new dimension of trustworthiness evaluation by incorporating entity reliability scores and relation reliability scores alongside subject concentration metrics. This multi-dimensional approach simultaneously addresses subject analysis precision and entity-relation trustworthiness.
3Reliability
If natural language processing and semantic inference are used to analyze entity relevance, then entity trustworthiness is considered, but comprehensive trustworthy evaluation is still lacking
Solution Approach 1:
The patent creates a universal knowledge graph framework that performs multiple functions: entity extraction, relation extraction, subject concentration analysis, and trustworthiness evaluation. This multi-functional system consolidates various evaluation tasks into a single integrated platform, managing complexity through unification.
Solution Approach 2:
The patent introduces specific parameters for trustworthiness evaluation, including entity reliability scores, relation reliability scores, and content support degrees. These quantifiable parameters transform qualitative trustworthiness assessment into measurable metrics, enabling comprehensive evaluation while maintaining system manageability.
Data Source
AI summary
The present invention relates to a trustworthy search method for a search engine based on a knowledge graph, which includes: acquiring a search keyword input by a user to construct a keyword pool; selecting a keyword according to a keyword selection policy, and respectively inputting the keyword to a search engine in sequence for searching to obtain a result returned by the search engine; constructing a knowledge graph of a Web page in sequence; selecting a specific knowledge mode; matching the knowledge graph with an existing semantics reliable knowledge graph library, and then computing a content support degree of each Web page according to a matching result; sorting the search engine under the same keyword by using a content support degree expectation of the Web page; and completing trustworthy search of the search engine based on the knowledge graph.