Intelligent retrieval method for constructing GPS standard knowledge graph based on terms

Abstract

The invention discloses an intelligent retrieval method for constructing a GPS standard knowledge graph based on terms. The intelligent retrieval method comprises the following steps: step 1, constructing a GPS standard knowledge graph according to term library data; 2, text information extraction is carried out based on retrieval conditions input by a user, and word segmentation processing is carried out on the extracted text information; 3, word sense disambiguation is conducted on the basis of the text information obtained after word segmentation processing; 4, performing semantic recognition based on the text information subjected to word sense disambiguation to extract keywords; 5, term matching is carried out in the GPS standard knowledge graph based on the extracted keywords, and a retrieval result corresponding to a text matching result input by the user is searched in an associated mode.

Description

Technical Field

[0001] The present invention belongs to the technical field of search engines, and particularly relates to an intelligent retrieval method for constructing a GPS standard knowledge graph based on terms. Background Art

[0002] Geometric Product Specifications & Verification (GPS) is a geometric technology standard system established for all products in terms of geometric features, which standardizes a set of geometric technology standards from macroscopic geometric features to microscopic geometric features of products, covering the whole process of the product life cycle including product design, manufacturing, inspection, use, maintenance, and scrapping, and the application fields involve the entire industrial sector and even all sectors of the national economy.

[0003] Currently, the GPS standard system includes more than 150 current standards and more than 20 standards under development. With the wide application of GPS standards in product design, manufacturing, and inspection, etc., when users search for standards related to specific design requirements, they mainly rely on the retrieval method of keyword matching. However, this traditional retrieval method has obvious limitations. It is difficult to quickly and accurately obtain the most relevant standard information, often resulting in users spending a lot of time and energy screening suitable content from a large number of standards. Traditional retrieval technologies are okay for organizing and retrieving independent information, but for data with strong relevance and complex relationships, their search results often cannot meet the actual needs of users, cannot provide comprehensive and related information, and are even less able to prioritize according to the content that users are concerned about.

[0004] Therefore, it is necessary to provide an intelligent retrieval method that can quickly, accurately, and effectively search for the correct standard information. Summary of the Invention

[0005] The object of the present invention is to provide an intelligent retrieval method for constructing a GPS standard knowledge graph based on terms, so as to solve the problem that the existing retrieval technology can only organize and retrieve independent information, but for data with strong relevance and relatively complex relationships, the search results cannot meet the actual target requirements of users, and the existing technology cannot meet the needs of users for more comprehensive and related information.

[0006] To solve the above technical problems, the present invention proposes an intelligent retrieval method for constructing a GPS standard knowledge graph based on terms, and the method includes the following steps: Step 1: Construct a knowledge graph of GPS standards according to the data in the term library; Step 2: Extract text information based on the retrieval conditions input by the user, and perform word segmentation processing on the extracted text information; Step 3: Disambiguate the meanings of words based on the text information after word segmentation processing; Step 4: Based on the above text information after word meaning disambiguation, perform semantic recognition to extract keywords; Step 5: Based on the above extracted keywords, perform term matching in the GPS standard knowledge graph, and associate and search for retrieval results corresponding to the matching results of the user input text.

[0007] As a further preferred solution of an intelligent retrieval method for constructing a GPS standard knowledge graph based on terms of the present invention, the specific steps of step 1 are as follows: Step 1.1, Obtain core terms from GPS standard content and construct a term library; Step 1.2, For the GPS standard knowledge graph, 4 entity types are defined: standard, term, geometric feature, and link, and 6 relationship types: adoption relationship, substitution relationship, belonging relationship, upper-level relationship, lower-level relationship, and citation relationship; Step 1.3, Based on rules, the following entities and their attributes can be extracted from GPS standard data: (1) Standard: label, title, standard status (current, repealed), scope, implementation date, adoption method; (2) Term: name, definition; (3) Geometric features are divided into 9 categories: 1 dimension, 2 distance, 3 shape, 4 direction, 5 position, 6 runout, 7 profile surface structure, 8 area surface structure, 9 surface defect; (4) Links are divided into 7 categories: A symbols and notations, B element requirements, C element characteristics, D compliance and non-compliance, E measurement, F measuring equipment, G calibration; Step 1.4, Based on the extracted entities, identify and extract the relationships between entities: adoption relationship, substitution relationship, belonging relationship, upper-level relationship, lower-level relationship, parallel relationship, citation relationship; among them, the upper-level relationship and the lower-level relationship are determined by the order of links A to G; Step 1.5, Establish GPS knowledge point associations through the GPS concept framework, thereby forming a GPS standard knowledge graph.

[0008] As a further preferred solution of an intelligent retrieval method for constructing a GPS standard knowledge graph based on terms of the present invention, step 2 is specifically: Extract text information based on the retrieval conditions input by the user, and perform word segmentation processing on the extracted text information, including: The word segmentation processing can adopt any one of the dynamic statistics method, the maximum matching method, the automatic word segmentation method, and the special rule correction method.

[0009] As a further preferred solution of the intelligent retrieval method for constructing a GPS standard knowledge graph based on terms in the present invention, step 3 is specifically as follows: Describe the meaning of words in the dictionary of the above GPS standard knowledge graph, and then obtain all the corresponding semantic conceptions of the text information after word segmentation and the semantic association relationships existing between the semantic conceptions. Take the semantic conceptions as vertices and the semantic association relationships as edges to construct a disambiguation graph. Use the PageRank algorithm to score the semantic conception nodes in the disambiguation graph to obtain the importance degree of each concept node, and select the semantic with the highest score as the disambiguation solution. Thus, the purpose of word sense disambiguation is achieved.

[0010] As a further preferred solution of the intelligent retrieval method for constructing a GPS standard knowledge graph based on terms in the present invention, step 4 includes: Extract keywords from the text information according to the user role (designer / detector), input behavior, idiomatic expressions, and in combination with the analysis of the extraction implementation stage, and recommend related words based on the term hierarchy.

[0011] As a further preferred solution of the intelligent retrieval method for constructing a GPS standard knowledge graph based on terms in the present invention, step 5 is specifically as follows: Based on the above extracted keywords, perform keyword matching in the legal knowledge graph, and associatively search for the retrieval results corresponding to the matching results of the user input text, including: Perform text fuzzy matching and text exact matching searches on the text to be detected in the knowledge graph through the segmented keywords, and obtain the text matching search results.

[0012] As a further preferred solution of the intelligent retrieval method for constructing a GPS standard knowledge graph based on terms in the present invention, step 5 includes: Classify and sort the retrieval results.

[0013] An intelligent retrieval method for constructing a GPS standard knowledge graph based on terms in the present invention has the following beneficial effects: The present invention is based on a knowledge graph that is separated from the traditional term library and uses a knowledge graph based on relationship integration. With the strong correlation between objects in the knowledge graph, a more accurate search hit rate can be obtained. By performing word sense disambiguation, semantic recognition on the case information input by the user, extracting keywords according to the user role (designer / detector), input behavior, idiomatic expressions, and in combination with the analysis of the extraction implementation stage, and recommending related words based on the term hierarchy, finally sorting the retrieval results according to keywords, geometric features, and links, and presenting the retrieval results to the user. Brief Description of the Drawings

[0014] Figure 1 It is a flowchart of the method of the present invention.

[0015] Figure 2 It is a flowchart of the knowledge graph construction of the present invention.

[0016] Figure 3 This is an example of the knowledge graph of the present invention. Detailed implementation manners

[0017] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings: As Figure 1 shown, in this embodiment, a method for constructing a knowledge graph for Product Geometric Technology Specification (GPS) standards specifically includes the following steps: Step 1: Construct a knowledge graph of GPS standards according to the data in the term library; Step 2: Extract text information based on the retrieval conditions input by the user, and perform word segmentation on the extracted text information; Step 3: Disambiguate the word meanings of the text information after word segmentation; Step 4: Semantically recognize and extract keywords from the text information after disambiguating the word meanings; Step 5: Perform term matching in the GPS standard knowledge graph based on the extracted keywords, and associatively search for retrieval results corresponding to the matching results of the user input text.

[0018] In an embodiment of the present invention, a knowledge graph of GPS standards is constructed according to the data in the term library, as Figure 2 shown, including: Step 1.1, obtain core terms from the GPS standard content and construct a term library; Step 1.2, define 4 entity types for the GPS standard knowledge graph: standard, term, geometric feature, and link, and 6 relationship types: adoption relationship, substitution relationship, belonging relationship, upper-level relationship, lower-level relationship, and citation relationship; Step 1.3, based on rules, the following entities and their attributes can be extracted from the GPS standard data: (1) Standard: label, title, standard status (current, abolished), scope, implementation date, adoption method; (2) Term: name, definition; (3) Geometric features are divided into 9 categories: 1 dimension, 2 distance, 3 shape, 4 direction, 5 position, 6 runout, 7 profile surface texture, 8 area surface texture, 9 surface defect; (4) Links are divided into 7 categories: A symbols and notations, B element requirements, C element features, D conformity and non-conformity, E measurement, F measuring equipment, G calibration; Step 1.4, based on the extracted entities, identify and extract the relationships between the entities: adoption relationship, substitution relationship, belonging relationship, upper-level relationship, lower-level relationship, parallel relationship, citation relationship; among them, the upper-level relationship and the lower-level relationship are determined by the order of links A to G; Step 1.5, establish the association of GPS knowledge points through the GPS conceptual framework, thereby forming the GPS standard knowledge graph.

[0019] Taking roundness as an example, an example of the constructed GPS knowledge graph is shown as Figure 3 shown. In the knowledge graph, circular terms are used to identify term entities, oblongs are used to identify standard entities, squares are used to identify the attributes of standard entities, ellipses are used to identify geometric features, and hexagons are used to identify links. In the knowledge graph of GPS standards, entities are mainly standards, and the considered attributes are as follows: label, title, standard status (current, repealed), scope, implementation date, and method of adopting international standards.

[0020] Entity nodes of a certain GPS standard can be constructed through entity and attribute information. After constructing the nodes, further construct the relationships between standards and standards, and between terms and terms. Entity relationships include the relationship of adopting international standards, substitution relationship, belonging relationship, upper-level relationship, lower-level relationship, parallel relationship, and citation relationship.

[0021] In an embodiment of the present invention, text information extraction is performed based on the retrieval conditions input by the user, and the extracted text information is subjected to word segmentation processing, including: any one of the dynamic statistics method, the maximum matching method, the automatic word segmentation method, and the special rule correction method can be used for word segmentation processing.

[0022] In an embodiment of the present invention, word sense disambiguation is performed based on the text information after word segmentation processing. Specifically: describe the meaning of words in the dictionary of the above GPS standard knowledge graph, and then obtain all the corresponding word sense concepts and the semantic association relationships existing between the word sense concepts for the text information after word segmentation processing. Use the word sense concepts as vertices and the semantic association relationships as edges to construct a disambiguation graph. Use the PageRank algorithm to score the word sense concept nodes in the disambiguation graph, obtain the importance of each concept node, and select the word sense with the highest score as the disambiguation solution. Thus, the purpose of word sense disambiguation is achieved.

[0023] In an embodiment of the present invention, semantic recognition is performed based on the above text information after word sense disambiguation to extract keywords, including: extracting keywords from the text information according to the user role (designer / detector), input behavior, habitual expressions, and combining with the analysis of the extraction implementation stage, and recommending related words based on the term hierarchy.

[0024] In an embodiment of the present invention, keyword matching is performed in the GPS standard knowledge graph based on the above extracted keywords, and the retrieval results corresponding to the matching results of the user input text are searched for associatively, including: performing text fuzzy matching and text exact matching searches on the text to be detected in the knowledge graph through the segmented keywords, and obtaining the text matching search results. In addition, the retrieval results can be classified and sorted.

[0025] An intelligent retrieval method for constructing a GPS standard knowledge graph based on terms of the present invention has the following beneficial effects: The present invention is based on a term library that is separated from the traditional one. By using a knowledge graph based on relationship integration and relying on the strong correlation between objects in the knowledge graph, a more accurate search hit rate can be obtained. By performing word sense disambiguation and semantic recognition on the case information input by the user, extracting keywords according to the user role (designer / detector), input behavior, idiomatic expressions, and combining with the analysis of the implementation stage, recommending related terms based on the term hierarchy, and finally sorting the retrieval results according to the keywords, geometric features, and link pairs, and presenting the retrieval results to the user. The method of the present invention solves the problem that the existing retrieval technology can only organize and retrieve independent information, and for data with strong correlation and complex relationships, the search results cannot meet the actual target requirements of users, and the existing technology cannot meet the need for more and more comprehensive information with associated relationships required by users.

[0026] The above specific implementation cases are used to explain the present invention. They are only the preferred embodiments of the present invention and do not limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and scope of the protection of the claims of the present invention fall within the protection scope of the present invention.

Claims

1. An intelligent retrieval method for constructing a GPS standard knowledge graph based on terms, characterized in that: The method includes the following steps: Step 1: Construct a knowledge graph of GPS standards based on the data in the term library; Step 2: Extract text information based on the retrieval conditions input by the user, and perform word segmentation on the extracted text information; Step 3: Disambiguate the word senses of the text information after word segmentation; Step 4: Based on the above, perform semantic recognition on the text information with disambiguated word senses to extract keywords; Step 5: Based on the above extracted keywords, perform term matching in the GPS standard knowledge graph, and associatively search for retrieval results corresponding to the matching results of the user input text.

2. The intelligent retrieval method according to claim 1, characterized in that: The specific steps of Step 1 are as follows: Step 1.1, obtain core terms from the GPS standard content and construct a term library; Step 1.2, for the GPS standard knowledge graph, 4 entity types are defined: standard, term, geometric feature, and link, and 6 relationship types: adopting standard relationship, substitution relationship, belonging relationship, upper-level relationship, lower-level relationship, citation relationship; Step 1.3, based on rules, the following entities and their attributes can be extracted from the GPS standard data: (1) Standard: label, title, standard status (current, abolished), scope, implementation date, method of adopting standard; (2) Term: name, definition; (3) Geometric features are divided into 9 categories: 1 dimension, 2 distance, 3 shape, 4 direction, 5 position, 6 runout, 7 profile surface texture, 8 area surface texture, 9 surface defect; (4) Links are divided into 7 categories: A symbols and notations, B element requirements, C element features, D compliance and non-compliance, E measurement, F measuring equipment, G calibration; Step 1.4, based on the extracted entities, identify and extract the relationships between entities: adopting standard relationship, substitution relationship, belonging relationship, upper-level relationship, lower-level relationship, parallel relationship, citation relationship; among them, the upper-level relationship and the lower-level relationship are determined by the order of links A to G; Step 1.5, establish GPS knowledge point associations through the GPS concept framework, thereby forming a GPS standard knowledge graph; As a further preferred solution of the intelligent retrieval method for constructing a GPS standard knowledge graph based on terms of the present invention, Step 2 is specifically: extract text information based on the retrieval conditions input by the user, and perform word segmentation on the extracted text information, including: the word segmentation process can adopt any one of the dynamic statistics method, the maximum matching method, the automatic word segmentation method, and the special rule correction method.

3. The intelligent retrieval method according to claim 1, characterized in that: The specific steps of Step 3 are: describe the meaning of words in the dictionary of the above GPS standard knowledge graph, then obtain all corresponding word sense concepts and the semantic association relationships existing between the word sense concepts for the text information after word segmentation, use the word sense concepts as vertices and the semantic association relationships as edges to construct a disambiguation graph. Use the PageRank algorithm to score the word sense concept nodes in the disambiguation graph, obtain the importance of each concept node, and select the word sense with the highest score as the disambiguation solution. Thus, the purpose of word sense disambiguation is achieved.

4. The intelligent retrieval method according to claim 1, wherein: Step 4 includes: extract keywords for the text information according to the user role (designer / detector), input behavior, idiomatic expressions, and in combination with the analysis of the extraction implementation stage, and recommend associated words based on the term hierarchy.

5. The intelligent retrieval method according to claim 1, characterized in that: Specifically, step 5 is to perform keyword matching in the legal knowledge graph based on the above-extracted keywords, and associatively search for retrieval results corresponding to the matching results of the user input text, including: performing text fuzzy matching and text exact matching searches on the text to be detected in the knowledge graph through the segmented keywords, and obtaining text matching search results.

6. The intelligent retrieval method according to any one of claims 1-5, characterized in that: Step 5 includes: classifying and sorting the retrieval results.

Images