A method and system for constructing industrial chain data knowledge graph based on artificial intelligence

Abstract

The present invention relates to the field of data processing technology, and specifically discloses a method and system for constructing an industrial chain data knowledge graph based on artificial intelligence, including: based on all industrial chain data texts and AI technology of each chemical, obtaining all standard extraction entities of the industrial chain data of each chemical; based on all standard extraction entities of the industrial chain data of each chemical, obtaining the original knowledge graph of the industrial chain data of each chemical; based on the original knowledge graph of the industrial chain data of each chemical, obtaining all trunk extraction entities of the industrial chain data of each chemical; based on all trunk extraction entities of the industrial chain data of each chemical, obtaining the final knowledge graph of the industrial chain data of each chemical. The present invention realizes the simplification of the knowledge graph and improves the readability of the knowledge graph.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for constructing an industrial chain data knowledge graph based on artificial intelligence. Background Art

[0002] At present, the chemical industry involves many links, from raw material procurement, production and processing, product circulation to final consumption, each link contains a large amount of information and data. These data are often scattered in various documents, reports, patents and other text materials, showing a high degree of unstructured and fragmented characteristics, and are difficult to use directly for analysis and decision-making. As an effective form of data organization, knowledge graphs can intuitively display complex data relationships in the form of graphs, which is easy to understand and analyze. In the field of chemical industry chain, building knowledge graphs can help enterprises and research institutions quickly grasp industry dynamics, identify key nodes, and discover potential opportunities and risks. However, after the original industry chain data is extracted and structured, the knowledge graph formed may contain a lot of redundant information and noise, which reduces the readability and practicality of the graph. Therefore, it is necessary to extract the most critical and representative information through further processing and optimization to form a refined knowledge graph representation.

[0003] However, the existing artificial intelligence-based industrial chain data knowledge graph construction method and system only extract entities and relationships from industrial chain data to construct systematic data with high accuracy and comprehensiveness, but do not consider how to simplify the knowledge graph and improve the readability of the graph. For example, the patent with the publication number of "CN117744769A" and the patent name of "knowledge graph construction method, device, electronic device and medium for industrial chain data" includes the following steps: extracting entities from industrial chain data to obtain entity results; based on the entity results, extracting relationships from the industrial chain data to obtain the associations corresponding to the entity results; and constructing the knowledge graph corresponding to the industrial chain data according to the entity results and the associations. This method extracts entities and relationships from industrial chain data to construct systematic data with high accuracy and comprehensiveness, that is, the knowledge graph corresponding to the industrial chain data, which helps users (such as enterprises and researchers) to manage and analyze industrial chain information more effectively, analyze industrial chain dynamics in real time, and comprehensively and accurately analyze potential business risks. However, this patent only extracts entities and relationships from industrial chain data to construct systematic data with high accuracy and comprehensiveness, but does not consider how to simplify the knowledge graph or improve the readability of the graph.

[0004] Therefore, the present invention proposes a method and system for constructing an industrial chain data knowledge graph based on artificial intelligence. Summary of the invention

[0005] The present invention provides an artificial intelligence-based industrial chain data knowledge graph construction method and system, which is used to obtain all industrial chain data texts of each chemical according to the industrial chain database, and then obtain all extraction entities and all extraction relationships in all industrial chain data texts of each chemical according to all industrial chain data texts of each chemical and AI technology, and obtain all standard extraction entities of industrial chain data of each chemical according to all extraction entities in all industrial chain data texts of each chemical, so as to facilitate the subsequent construction of the original knowledge graph of industrial chain data of the chemical, obtain the original knowledge graph of industrial chain data of each chemical according to all standard extraction entities of industrial chain data of each chemical and all extraction relationships in all industrial chain data texts of each chemical, and accurately obtain the original knowledge graph of industrial chain data of each chemical. The complete knowledge graph representation of the industrial chain data information obtains all predicted key extracted entities of the industrial chain data of each chemical based on the original knowledge graph of the industrial chain data of each chemical, accurately predicts all extracted entities that can reflect the key information in the industrial chain data of each chemical, and obtains all backbone extracted entities of the industrial chain data of each chemical based on all predicted key extracted entities of the industrial chain data of each chemical, accurately obtains all standard extracted entities that can serve as the final knowledge graph nodes of the industrial chain data of the chemical, and finally obtains the final knowledge graph of the industrial chain data of each chemical based on all backbone extracted entities of the industrial chain data of each chemical, accurately obtains the refined knowledge graph representation of the industrial chain data information of each chemical, and improves the readability of the knowledge graph.

[0006] The present invention provides a method for constructing an industrial chain data knowledge graph based on artificial intelligence, comprising:

[0007] S1: Based on the industrial chain database, all industrial chain data texts of each chemical are obtained. Based on all industrial chain data texts of each chemical and AI technology, all extraction entities and all extraction relationships in all industrial chain data texts of each chemical are obtained. Based on all extraction entities in all industrial chain data texts of each chemical, all standard extraction entities of industrial chain data of each chemical are obtained.

[0008] S2: Based on all standard extraction entities of the industrial chain data of each chemical and all extraction relations in the text of all industrial chain data of each chemical, the original knowledge graph of the industrial chain data of each chemical is obtained;

[0009] S3: Based on the original knowledge graph of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are obtained, and based on all predicted key extraction entities of the industrial chain data of each chemical, all backbone extraction entities of the industrial chain data of each chemical are obtained;

[0010] S4: Based on all the backbone extraction entities of the industrial chain data of each chemical, the final knowledge graph of the industrial chain data of each chemical is obtained.

[0011] Preferably, the method for constructing the industrial chain data knowledge graph based on artificial intelligence, S1: based on the industrial chain database, all industrial chain data texts of each chemical are obtained, based on all industrial chain data texts of each chemical and AI technology, all extraction entities and all extraction relationships in all industrial chain data texts of each chemical are obtained, and based on all extraction entities in all industrial chain data texts of each chemical, all standard extraction entities of the industrial chain data of each chemical are obtained, including:

[0012] Obtain all industrial chain data texts of each chemical from the industrial chain database, and based on each industrial chain data text of each chemical and AI technology, obtain all extracted entities and all extracted relationships in each industrial chain data text of each chemical;

[0013] Taking any two different extracted entities from all extracted entities in all industrial chain data texts of each chemical as a set of input groups of the corresponding chemical, and obtaining all group input groups of each chemical;

[0014] Based on all group input groups of each chemical and the preset entity determination model, all standard extraction entities of the industrial chain data of each chemical are obtained.

[0015] Preferably, the method for constructing the knowledge graph of industrial chain data based on artificial intelligence obtains all standard extraction entities of the industrial chain data of each chemical based on all group input groups and a preset entity judgment model for each chemical, including:

[0016] Based on each input group of each chemical and a preset entity determination model, obtaining entity determination results for each input group of each chemical;

[0017] If the entity determination results of each input group of each chemical are the same, the two extracted entities in the corresponding input group of the corresponding chemical are determined to be mutually converted entities;

[0018] Summarize all the extracted entities that are mutually converted entities in all the group input groups of each chemical as a group of conversion entity groups for each chemical, and obtain all the group conversion entity groups for each chemical;

[0019] The extracted entity with the least number of characters in each conversion entity group of each chemical is regarded as the preferred entity of the corresponding conversion entity group of the corresponding chemical, and the preferred entity of each conversion entity group of each chemical is regarded as the standard extracted entity of the industrial chain data of the corresponding chemical.

[0020] Preferably, the method for constructing the knowledge graph of industrial chain data based on artificial intelligence, S2: based on all standard extraction entities of the industrial chain data of each chemical and all extraction relations in the text of all industrial chain data of each chemical, the original knowledge graph of the industrial chain data of each chemical is obtained, including:

[0021] Based on all standard extraction entities of the industrial chain data of each chemical, all extraction relationships in all industrial chain data texts of each chemical are simplified to obtain all simplified extraction relationships;

[0022] All standard extracted entities of the industrial chain data of each chemical are taken as nodes, and all simplified extracted relationships are taken as relationship edges to obtain the original knowledge graph of the industrial chain data of each chemical.

[0023] Preferably, the method for constructing the knowledge graph of industrial chain data based on artificial intelligence, S3: based on the original knowledge graph of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are obtained, and based on all predicted key extraction entities of the industrial chain data of each chemical, all trunk extraction entities of the industrial chain data of each chemical are obtained, including:

[0024] The number of other standard extraction entities connected to each standard extraction entity in the original knowledge graph of the industrial chain data of each chemical is regarded as the number of connections of the corresponding standard extraction entity of the industrial chain data of the corresponding chemical;

[0025] Based on each standard extraction entity of the industrial chain data of each chemical and a preset entity type scoring model, a type scoring value of a corresponding standard extraction entity of the industrial chain data of the corresponding chemical is obtained;

[0026] The standard extraction entity corresponding to each chemical in the original knowledge graph of the industrial chain data of the corresponding chemical is used as the core extraction entity of the industrial chain data of the corresponding chemical, and the minimum number of intermediate entities between each standard extraction entity of the industrial chain data of each chemical and the core extraction entity of the industrial chain data of the corresponding chemical is used as the minimum number of intermediate entities corresponding to the standard extraction entity of the industrial chain data of the corresponding chemical;

[0027] Based on the number of connections, type score and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are obtained;

[0028] Based on all predicted key extraction entities of the industrial chain data of each chemical, all backbone extraction entities of the industrial chain data of each chemical are obtained.

[0029] Preferably, the method for constructing the knowledge graph of industrial chain data based on artificial intelligence obtains all predicted key extraction entities of the industrial chain data of each chemical based on the number of connections, type score and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical, including:

[0030] Based on the number of connections, type score and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical, the entity evaluation value of each standard extraction entity of the industrial chain data of the corresponding chemical is obtained, which is:

[0031]

[0032] Among them, δ is the entity evaluation value of the current calculation standard extraction entity of the industrial chain data of the current calculation chemical, b is the number of connections of the current calculation standard extraction entity of the industrial chain data of the current calculation chemical, g is the type score value of the current calculation standard extraction entity of the industrial chain data of the current calculation chemical, v is the minimum middle number of the current calculation standard extraction entity of the industrial chain data of the current calculation chemical, ln is the natural logarithm, and the value of the natural constant e is 2.718;

[0033] Based on the entity evaluation values ​​of all standard extracted entities of the industrial chain data of each chemical, all predicted key extracted entities of the industrial chain data of each chemical are obtained.

[0034] Preferably, the method for constructing the knowledge graph of industrial chain data based on artificial intelligence obtains all predicted key extracted entities of the industrial chain data of each chemical based on the entity evaluation values ​​of all standard extracted entities of the industrial chain data of each chemical, including:

[0035] If the entity evaluation value of each standard extraction entity of the industrial chain data of each chemical is greater than a preset threshold, the corresponding standard extraction entity is regarded as a predicted key extraction entity of the industrial chain data of the corresponding chemical.

[0036] Preferably, the method for constructing the knowledge graph of industrial chain data based on artificial intelligence obtains all the main extraction entities of the industrial chain data of each chemical based on all the predicted key extraction entities of the industrial chain data of each chemical, including:

[0037] The standard extraction entities on the shortest intermediate paths of all predicted key extraction entities in the original knowledge graph of each chemical's industrial chain data are regarded as the backbone extraction entities of the chemical's industrial chain data.

[0038] Preferably, the method for constructing a knowledge graph of industrial chain data based on artificial intelligence, S4: extracting entities based on all the backbones of the industrial chain data of each chemical to obtain a final knowledge graph of the industrial chain data of each chemical, including:

[0039] In the original knowledge graph of the industrial chain data of each chemical, all standard extraction entities except the backbone extraction entities are deleted to obtain a new knowledge graph, and the new knowledge graph is used as the final knowledge graph of the industrial chain data of the corresponding chemical.

[0040] The present invention provides an artificial intelligence-based industrial chain data knowledge graph construction system, which is used to execute any one of the artificial intelligence-based industrial chain data knowledge graph construction methods in embodiments 1 to 9, including:

[0041] An extraction module is used to obtain all industrial chain data texts of each chemical based on the industrial chain database, obtain all extraction entities and all extraction relationships within all industrial chain data texts of each chemical based on all industrial chain data texts of each chemical and AI technology, and obtain all standard extraction entities of industrial chain data of each chemical based on all extraction entities within all industrial chain data texts of each chemical;

[0042] The first building module is used to obtain the original knowledge graph of the industrial chain data of each chemical based on all standard extraction entities of the industrial chain data of each chemical and all extraction relationships in the text of all industrial chain data of each chemical;

[0043] A calculation module, for obtaining all predicted key extraction entities of the industrial chain data of each chemical based on the original knowledge graph of the industrial chain data of each chemical, and obtaining all trunk extraction entities of the industrial chain data of each chemical based on all predicted key extraction entities of the industrial chain data of each chemical;

[0044] The second building module is used to extract entities based on all the backbones of the industrial chain data of each chemical to obtain the final knowledge graph of the industrial chain data of each chemical.

[0045] The beneficial effects of the present invention compared with the prior art are as follows: according to the industrial chain database, all industrial chain data texts of each chemical are obtained, and then according to all industrial chain data texts of each chemical and AI technology, all extraction entities and all extraction relationships in all industrial chain data texts of each chemical are obtained, and according to all extraction entities in all industrial chain data texts of each chemical, all standard extraction entities of industrial chain data of each chemical are obtained, which is convenient for the subsequent construction of the original knowledge graph of industrial chain data of the chemical, according to all standard extraction entities of industrial chain data of each chemical and all extraction relationships in all industrial chain data texts of each chemical, the original knowledge graph of industrial chain data of each chemical is obtained, and the industrial chain data information of each chemical is accurately obtained. A complete knowledge graph representation of the industrial chain data of each chemical is obtained. According to the original knowledge graph of the industrial chain data of each chemical, all predicted key extracted entities of the industrial chain data of each chemical are obtained, and all extracted entities that can reflect the key information in the industrial chain data of each chemical are accurately predicted. According to all predicted key extracted entities of the industrial chain data of each chemical, all backbone extracted entities of the industrial chain data of each chemical are obtained, and all standard extracted entities that can be used as the final knowledge graph nodes of the industrial chain data of the chemical are accurately obtained. Finally, according to all backbone extracted entities of the industrial chain data of each chemical, the final knowledge graph of the industrial chain data of each chemical is obtained, and the refined knowledge graph representation of the industrial chain data information of each chemical is accurately obtained, which improves the readability of the knowledge graph.

[0046] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structure specifically pointed out in the written application document.

[0047] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0049] Figure 1 This is a flow chart of a method for constructing an industrial chain data knowledge graph based on artificial intelligence in an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of a system for building an industrial chain data knowledge graph based on artificial intelligence in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0052] Embodiment 1:

[0053] The present invention provides a method for constructing an industrial chain data knowledge graph based on artificial intelligence. Figure 1 ,include:

[0054] S1: Based on the industrial chain database, all industrial chain data texts of each chemical are obtained. Based on all industrial chain data texts of each chemical and AI technology, all extraction entities and all extraction relationships in all industrial chain data texts of each chemical are obtained. Based on all extraction entities in all industrial chain data texts of each chemical, all standard extraction entities of industrial chain data of each chemical are obtained.

[0055] S2: Based on all standard extraction entities of the industrial chain data of each chemical and all extraction relations in the text of all industrial chain data of each chemical, the original knowledge graph of the industrial chain data of each chemical is obtained;

[0056] S3: Based on the original knowledge graph of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are obtained, and based on all predicted key extraction entities of the industrial chain data of each chemical, all backbone extraction entities of the industrial chain data of each chemical are obtained;

[0057] S4: Based on all the backbone extraction entities of the industrial chain data of each chemical, the final knowledge graph of the industrial chain data of each chemical is obtained.

[0058] In this embodiment, the industrial chain database is a resource database that integrates a large amount of information related to chemicals (such as p-toluenesulfonyl chloride, isophthalonitrile, etc.). The database covers data from raw material procurement, production and processing, product sales to waste disposal.

[0059] In this embodiment, the industrial chain data text is text information about the industrial chain data of each chemical extracted from the industrial chain database, such as product research reports, production instructions, etc.

[0060] In this embodiment, the AI ​​technology is an existing technology used to obtain all extracted entities and all extracted relationships from all industrial chain data texts of each chemical, such as NLP technology.

[0061] In this embodiment, the extracted entities are key entities such as chemical names, company names, technical terms, etc. identified from the industrial chain data text of each chemical.

[0062] In this embodiment, the extracted relationship is the relationship between the extracted entities extracted from the industrial chain data text of each chemical, such as the reaction relationship, production relationship, supply chain relationship, etc. between chemical products.

[0063] In this embodiment, the standard extraction entity is the extraction entity that is most suitable as a knowledge graph node selected from all the extraction entities in all the industrial chain data texts of each chemical.

[0064] In this embodiment, the original knowledge graph is a knowledge graph constructed based on all standard extraction entities of the industrial chain data of each chemical and all extraction relationships within all industrial chain data texts of each chemical, which can fully reflect the industrial chain data information of each chemical.

[0065] In this embodiment, the predicted key extraction entities are the original knowledge graphs based on the industrial chain data of each chemical, and the predicted extraction entities can reflect the key information (such as key raw materials, important intermediates, etc.) in the industrial chain data of each chemical.

[0066] In this embodiment, the backbone extraction entities are all predicted key extraction entities based on the industrial chain data of each chemical, and all standard extraction entities that can be obtained as the final knowledge graph nodes of the industrial chain data of each chemical.

[0067] In this embodiment, the final knowledge graph is a knowledge graph obtained by extracting all the main entities based on the industrial chain data of each chemical, which can refine and reflect the industrial chain data information of each chemical.

[0068] The beneficial effects of the above technology are: according to the industrial chain database, all industrial chain data texts of each chemical are obtained, and then according to all industrial chain data texts of each chemical and AI technology, all extraction entities and all extraction relationships in all industrial chain data texts of each chemical are obtained, and according to all extraction entities in all industrial chain data texts of each chemical, all standard extraction entities of industrial chain data of each chemical are obtained, which is convenient for the subsequent construction of the original knowledge graph of the industrial chain data of the chemical, and according to all standard extraction entities of the industrial chain data of each chemical and all extraction relationships in all industrial chain data texts of each chemical, the original knowledge graph of the industrial chain data of each chemical is obtained, and the complete knowledge graph of the industrial chain data information of each chemical is accurately obtained. The knowledge graph representation is based on the original knowledge graph of the industrial chain data of each chemical, all predicted key extracted entities of the industrial chain data of each chemical are obtained, all extracted entities that can reflect the key information in the industrial chain data of each chemical are accurately predicted, and based on all predicted key extracted entities of the industrial chain data of each chemical, all backbone extracted entities of the industrial chain data of each chemical are obtained, and all standard extracted entities that can be used as the final knowledge graph nodes of the industrial chain data of the chemical are accurately obtained. Finally, based on all backbone extracted entities of the industrial chain data of each chemical, the final knowledge graph of the industrial chain data of each chemical is obtained, and the refined knowledge graph representation of the industrial chain data information of each chemical is accurately obtained, which improves the readability of the knowledge graph.

[0069] Embodiment 2:

[0070] Based on Example 1, an artificial intelligence-based industrial chain data knowledge graph construction method, S1: Based on the industrial chain database, all industrial chain data texts of each chemical are obtained, based on all industrial chain data texts of each chemical and AI technology, all extraction entities and all extraction relationships in all industrial chain data texts of each chemical are obtained, and based on all extraction entities in all industrial chain data texts of each chemical, all standard extraction entities of the industrial chain data of each chemical are obtained, including:

[0071] Obtain all industrial chain data texts of each chemical from the industrial chain database, and based on each industrial chain data text of each chemical and AI technology, obtain all extracted entities and all extracted relationships in each industrial chain data text of each chemical;

[0072] Taking any two different extracted entities from all extracted entities in all industrial chain data texts of each chemical as a set of input groups of the corresponding chemical, and obtaining all group input groups of each chemical;

[0073] Based on all group input groups of each chemical and the preset entity determination model, all standard extraction entities of the industrial chain data of each chemical are obtained.

[0074] In this embodiment, the preset entity judgment model uses a large number of input groups collected in advance as model input, and the manually labeled judgment results of whether two different extracted entities in the corresponding input groups have the same meaning as the model output. The trained model can input a single input group and output the judgment results of whether two different extracted entities in the corresponding input group have the same meaning.

[0075] The beneficial effects of the above technology are: according to the industrial chain database, all industrial chain data texts of each chemical are obtained, which is convenient for the subsequent acquisition of extraction entities and extraction relationships, and then according to all industrial chain data texts of each chemical and AI technology, all extraction entities and all extraction relationships in all industrial chain data texts of each chemical are obtained, and according to all extraction entities in all industrial chain data texts of each chemical, all standard extraction entities of the industrial chain data of each chemical are obtained, which is convenient for the subsequent construction of the original knowledge graph of the industrial chain data of the chemical.

[0076] Embodiment 3:

[0077] On the basis of Example 2, the method for constructing the knowledge graph of industrial chain data based on artificial intelligence obtains all standard extraction entities of the industrial chain data of each chemical based on all group input groups and the preset entity judgment model of each chemical, including:

[0078] Based on each input group of each chemical and a preset entity determination model, obtaining entity determination results for each input group of each chemical;

[0079] If the entity determination results of each input group of each chemical are the same, the two extracted entities in the corresponding input group of the corresponding chemical are determined to be mutually converted entities;

[0080] Summarize all the extracted entities that are mutually converted entities in all the group input groups of each chemical as a group of conversion entity groups for each chemical, and obtain all the group conversion entity groups for each chemical;

[0081] The extracted entity with the least number of characters in each conversion entity group of each chemical is regarded as the preferred entity of the corresponding conversion entity group of the corresponding chemical, and the preferred entity of each conversion entity group of each chemical is regarded as the standard extracted entity of the industrial chain data of the corresponding chemical.

[0082] In this embodiment, the entity determination result is a determination result of whether two extracted entities in each input group of each chemical have the same meaning, obtained according to a preset entity determination model.

[0083] In this embodiment, the conversion entity group is a combination of all extracted entities that are mutually convertible entities in all group input groups of each chemical. For example, suppliers and suppliers in input group a are determined to be mutually convertible entities, suppliers and suppliers in input group b are determined to be mutually convertible entities, and suppliers and supply partners in input group c are determined to be mutually convertible entities. Then the conversion entity group is a combination of suppliers, suppliers, suppliers, and supply partners.

[0084] In this embodiment, the preferred entity is the extracted entity with the least number of characters in each group of conversion entities of each chemical.

[0085] The beneficial effects of the above technology are: based on all group input groups of each chemical and a preset entity judgment model, all group conversion entity groups of each chemical are obtained, and then based on all group conversion entity groups of each chemical, the preferred entities of each group conversion entity group of each chemical are obtained, and finally, based on the preferred entities of all group conversion entity groups of each chemical, all standard extraction entities of the industrial chain data of each chemical are obtained. This embodiment gives in detail a specific method for obtaining all standard extraction entities of the industrial chain data of each chemical based on all group input groups of each chemical and a preset entity judgment model.

[0086] Embodiment 4:

[0087] Based on Example 1, the method for constructing a knowledge graph of industrial chain data based on artificial intelligence, S2: based on all standard extraction entities of the industrial chain data of each chemical and all extraction relationships in the text of all industrial chain data of each chemical, the original knowledge graph of the industrial chain data of each chemical is obtained, including:

[0088] Based on all standard extraction entities of the industrial chain data of each chemical, all extraction relationships in all industrial chain data texts of each chemical are simplified to obtain all simplified extraction relationships;

[0089] All standard extracted entities of the industrial chain data of each chemical are taken as nodes, and all simplified extracted relationships are taken as relationship edges to obtain the original knowledge graph of the industrial chain data of each chemical.

[0090] In this embodiment, all extraction relationships in all industrial chain data texts of each chemical are simplified based on all standard extraction entities of the industrial chain data of each chemical to obtain all simplified extraction relationships, namely:

[0091] In each standard extraction entity corresponding group conversion entity group, the remaining extraction entities except the corresponding standard extraction entity are treated as replaceable entities of the corresponding standard extraction entity, the extraction relationships of all replaceable entities of each standard extraction entity are retained on the corresponding standard extraction entity, and the extraction relationships originally possessed by all standard extraction entities and the extraction relationships retained corresponding to all replaceable entities of all standard extraction entities are summarized as simplified extraction relationships between all standard extraction entities.

[0092] The beneficial effects of the above technology are: based on all standard extraction entities of the industrial chain data of each chemical and all extraction relationships within the text of all industrial chain data of each chemical, the original knowledge graph of the industrial chain data of each chemical is obtained, and the complete knowledge graph representation of the industrial chain data information of each chemical is accurately obtained. This embodiment gives in detail a specific method for obtaining the original knowledge graph of the industrial chain data of each chemical based on all standard extraction entities of the industrial chain data of each chemical and all extraction relationships within the text of all industrial chain data of each chemical.

[0093] Embodiment 5:

[0094] Based on Example 1, an artificial intelligence-based industrial chain data knowledge graph construction method, S3: based on the original knowledge graph of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are obtained, and based on all predicted key extraction entities of the industrial chain data of each chemical, all trunk extraction entities of the industrial chain data of each chemical are obtained, including:

[0095] The number of other standard extraction entities connected to each standard extraction entity in the original knowledge graph of the industrial chain data of each chemical is regarded as the number of connections of the corresponding standard extraction entity of the industrial chain data of the corresponding chemical;

[0096] Based on each standard extraction entity of the industrial chain data of each chemical and a preset entity type scoring model, a type scoring value of a corresponding standard extraction entity of the industrial chain data of the corresponding chemical is obtained;

[0097] The standard extraction entity corresponding to each chemical in the original knowledge graph of the industrial chain data of the corresponding chemical is used as the core extraction entity of the industrial chain data of the corresponding chemical, and the minimum number of intermediate entities between each standard extraction entity of the industrial chain data of each chemical and the core extraction entity of the industrial chain data of the corresponding chemical is used as the minimum number of intermediate entities corresponding to the standard extraction entity of the industrial chain data of the corresponding chemical;

[0098] Based on the number of connections, type score and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are obtained;

[0099] Based on all predicted key extraction entities of the industrial chain data of each chemical, all backbone extraction entities of the industrial chain data of each chemical are obtained.

[0100] In this embodiment, the number of connections of the standard extraction entity is the number of other standard extraction entities directly connected to the standard extraction entity in the original knowledge graph of the industrial chain data of each chemical.

[0101] In this embodiment, the preset entity type scoring model uses a large number of standard extracted entities of a large amount of chemical industry chain data collected in advance as model input, and the type scoring values ​​of the corresponding standard extracted entities of the chemical industry chain data manually labeled as model output. The model can output the type scoring values ​​of the corresponding standard extracted entities by inputting the standard extracted entities obtained through training.

[0102] In this embodiment, the type score assigned to the standard extracted entity is a numerical value representing the importance of the entity type (eg, geographic entity, business entity, etc.) of the standard extracted entity.

[0103] In this embodiment, the minimum number of intermediates of standard extraction entities is the minimum number of intermediate entities between each standard extraction entity of the industrial chain data of each chemical and the core extraction entity of the industrial chain data of the corresponding chemical, wherein the minimum number of intermediate entities is the number of standard extraction entities on the shortest intermediate path between each standard extraction entity in the original knowledge graph of the industrial chain data of each chemical and the core extraction entity of the industrial chain data of the corresponding chemical (the connection path with the least number of nodes among all connection paths between each standard extraction entity and the core extraction entity in the original knowledge graph of the industrial chain data of each chemical).

[0104] The beneficial effects of the above technology are: according to the original knowledge graph of the industrial chain data of each chemical, the number of connections, type score value and the minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical are obtained; according to the number of connections, type score value and the minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are obtained, and all extraction entities that can reflect the key information in the industrial chain data of each chemical are accurately predicted; then, according to all predicted key extraction entities of the industrial chain data of each chemical, all backbone extraction entities of the industrial chain data of each chemical are obtained, and all standard extraction entities that can serve as the final knowledge graph nodes of the industrial chain data of the chemical are accurately obtained.

[0105] Embodiment 6:

[0106] On the basis of Example 5, the method for constructing the knowledge graph of industrial chain data based on artificial intelligence obtains all predicted key extraction entities of the industrial chain data of each chemical based on the number of connections, type score and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical, including:

[0107] Based on the number of connections, type score and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical, the entity evaluation value of each standard extraction entity of the industrial chain data of the corresponding chemical is obtained, which is:

[0108]

[0109] Among them, δ is the entity evaluation value of the current calculation standard extraction entity of the industrial chain data of the current calculation chemical, b is the number of connections of the current calculation standard extraction entity of the industrial chain data of the current calculation chemical, g is the type score value of the current calculation standard extraction entity of the industrial chain data of the current calculation chemical, v is the minimum middle number of the current calculation standard extraction entity of the industrial chain data of the current calculation chemical, ln is the natural logarithm, and the value of the natural constant e is 2.718;

[0110] Based on the entity evaluation values ​​of all standard extracted entities of the industrial chain data of each chemical, all predicted key extracted entities of the industrial chain data of each chemical are obtained.

[0111] In this embodiment, the entity evaluation value of each standard extraction entity of the industrial chain data of each chemical is a numerical value evaluated based on the number of connections, type score value and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical, which can reflect the importance of the corresponding standard extraction entity of the industrial chain data of the corresponding chemical in the industrial chain.

[0112] The beneficial effects of the above technology are: according to the original knowledge graph of the industrial chain data of each chemical, the number of connections, type score value and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical are obtained; according to the number of connections, type score value and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are obtained, and all extraction entities that can reflect the key information in the industrial chain data of each chemical are accurately predicted. This embodiment gives in detail a specific method for obtaining all predicted key extraction entities of the industrial chain data of each chemical based on the number of connections, type score value and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical.

[0113] Embodiment 7:

[0114] On the basis of Example 6, the method for constructing the knowledge graph of industrial chain data based on artificial intelligence is based on the entity evaluation values ​​of all standard extraction entities of the industrial chain data of each chemical, and all predicted key extraction entities of the industrial chain data of each chemical are obtained, including:

[0115] If the entity evaluation value of each standard extraction entity of the industrial chain data of each chemical is greater than a preset threshold, the corresponding standard extraction entity is regarded as a predicted key extraction entity of the industrial chain data of the corresponding chemical.

[0116] In this embodiment, the preset threshold is a threshold set in advance to determine whether each standard extraction entity of the industrial chain data of each chemical is a predicted key extraction entity.

[0117] The beneficial effects of the above technology are: according to the entity evaluation value and preset threshold of each standard extraction entity of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are accurately obtained, and all extraction entities that can reflect the key information in the industrial chain data of each chemical are accurately predicted.

[0118] Embodiment 8:

[0119] On the basis of Example 5, the method for constructing the knowledge graph of industrial chain data based on artificial intelligence is based on all predicted key extraction entities of the industrial chain data of each chemical, and all the main extraction entities of the industrial chain data of each chemical are obtained, including:

[0120] The standard extraction entities on the shortest intermediate paths of all predicted key extraction entities in the original knowledge graph of each chemical's industrial chain data are regarded as the backbone extraction entities of the chemical's industrial chain data.

[0121] In this embodiment, the shortest intermediate path of the predicted key extraction entity is the connection path with the least number of nodes among all connection paths between each predicted key extraction entity and the core extraction entity in the original knowledge graph of the industrial chain data of each chemical.

[0122] The beneficial effects of the above technology are: based on all predicted key extraction entities of the industrial chain data of each chemical, all the main extraction entities of the industrial chain data of each chemical are obtained, which is convenient for the subsequent construction of the final knowledge graph. This embodiment gives in detail a specific method for obtaining all the main extraction entities of the industrial chain data of each chemical based on all predicted key extraction entities of the industrial chain data of each chemical.

[0123] Embodiment 9:

[0124] Based on Example 1, the method for constructing a knowledge graph of industrial chain data based on artificial intelligence, S4: extracting entities based on all the backbones of the industrial chain data of each chemical to obtain a final knowledge graph of the industrial chain data of each chemical, including:

[0125] In the original knowledge graph of the industrial chain data of each chemical, all standard extraction entities except the backbone extraction entities are deleted to obtain a new knowledge graph, and the new knowledge graph is used as the final knowledge graph of the industrial chain data of the corresponding chemical.

[0126] In this embodiment, deletion is a process of deleting all standard extraction entities except the backbone extraction entity in the original knowledge graph of the industrial chain data of each chemical, and deleting the extraction relationships possessed by all deleted standard extraction entities.

[0127] The beneficial effects of the above technology are: according to all the backbone extraction entities of the industrial chain data of each chemical, the final knowledge graph of the industrial chain data of each chemical is obtained, and the refined knowledge graph representation of the industrial chain data information of each chemical is accurately obtained, thereby improving the readability of the knowledge graph. This embodiment gives in detail a specific method for obtaining the final knowledge graph of the industrial chain data of each chemical based on all the backbone extraction entities of the industrial chain data of each chemical.

[0128] Embodiment 10:

[0129] The present invention provides an artificial intelligence-based industrial chain data knowledge graph construction system, which is used to execute any one of the artificial intelligence-based industrial chain data knowledge graph construction methods in embodiments 1 to 9, with reference to Figure 2 ,include:

[0130] An extraction module is used to obtain all industrial chain data texts of each chemical based on the industrial chain database, obtain all extraction entities and all extraction relationships within all industrial chain data texts of each chemical based on all industrial chain data texts of each chemical and AI technology, and obtain all standard extraction entities of industrial chain data of each chemical based on all extraction entities within all industrial chain data texts of each chemical;

[0131] The first building module is used to obtain the original knowledge graph of the industrial chain data of each chemical based on all standard extraction entities of the industrial chain data of each chemical and all extraction relationships in the text of all industrial chain data of each chemical;

[0132] A calculation module, for obtaining all predicted key extraction entities of the industrial chain data of each chemical based on the original knowledge graph of the industrial chain data of each chemical, and obtaining all trunk extraction entities of the industrial chain data of each chemical based on all predicted key extraction entities of the industrial chain data of each chemical;

[0133] The second building module is used to extract entities based on all the backbones of the industrial chain data of each chemical to obtain the final knowledge graph of the industrial chain data of each chemical.

[0134] The beneficial effects of the above technology are: according to the industrial chain database, all industrial chain data texts of each chemical are obtained, and then according to all industrial chain data texts of each chemical and AI technology, all extraction entities and all extraction relationships in all industrial chain data texts of each chemical are obtained, and according to all extraction entities in all industrial chain data texts of each chemical, all standard extraction entities of industrial chain data of each chemical are obtained, which is convenient for the subsequent construction of the original knowledge graph of the industrial chain data of the chemical, and according to all standard extraction entities of the industrial chain data of each chemical and all extraction relationships in all industrial chain data texts of each chemical, the original knowledge graph of the industrial chain data of each chemical is obtained, and the complete knowledge graph of the industrial chain data information of each chemical is accurately obtained. The knowledge graph representation is based on the original knowledge graph of the industrial chain data of each chemical, all predicted key extracted entities of the industrial chain data of each chemical are obtained, all extracted entities that can reflect the key information in the industrial chain data of each chemical are accurately predicted, and based on all predicted key extracted entities of the industrial chain data of each chemical, all backbone extracted entities of the industrial chain data of each chemical are obtained, and all standard extracted entities that can be used as the final knowledge graph nodes of the industrial chain data of the chemical are accurately obtained. Finally, based on all backbone extracted entities of the industrial chain data of each chemical, the final knowledge graph of the industrial chain data of each chemical is obtained, and the refined knowledge graph representation of the industrial chain data information of each chemical is accurately obtained, which improves the readability of the knowledge graph.

[0135] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention, and the present invention is also intended to include these changes and modifications.

Claims

1. A method for constructing an industrial chain data knowledge graph based on artificial intelligence, characterized in that: include: S1: Based on the industrial chain database, all industrial chain data texts of each chemical are obtained. Based on all industrial chain data texts of each chemical and AI technology, all extraction entities and all extraction relationships in all industrial chain data texts of each chemical are obtained. Based on all extraction entities in all industrial chain data texts of each chemical, all standard extraction entities of industrial chain data of each chemical are obtained. S2: Based on all standard extraction entities of the industrial chain data of each chemical and all extraction relations in the text of all industrial chain data of each chemical, the original knowledge graph of the industrial chain data of each chemical is obtained; S3: Based on the original knowledge graph of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are obtained, and based on all predicted key extraction entities of the industrial chain data of each chemical, all backbone extraction entities of the industrial chain data of each chemical are obtained; specifically including: The number of other standard extraction entities connected to each standard extraction entity in the original knowledge graph of the industrial chain data of each chemical is regarded as the number of connections of the corresponding standard extraction entity of the industrial chain data of the corresponding chemical; Based on each standard extraction entity of the industrial chain data of each chemical and a preset entity type scoring model, a type scoring value of a corresponding standard extraction entity of the industrial chain data of the corresponding chemical is obtained; The standard extraction entity corresponding to each chemical in the original knowledge graph of the industrial chain data of the corresponding chemical is used as the core extraction entity of the industrial chain data of the corresponding chemical, and the minimum number of intermediate entities between each standard extraction entity of the industrial chain data of each chemical and the core extraction entity of the industrial chain data of the corresponding chemical is used as the minimum number of intermediate entities corresponding to the standard extraction entity of the industrial chain data of the corresponding chemical; Based on the number of connections, type score and minimum intermediate number of each standard extraction entity of the industrial chain data of each chemical, the entity evaluation value of each standard extraction entity of the industrial chain data of the corresponding chemical is obtained; based on the entity evaluation values ​​of all standard extraction entities of the industrial chain data of each chemical, all predicted key extraction entities of the industrial chain data of each chemical are obtained; Based on all predicted key extraction entities of the industrial chain data of each chemical, all backbone extraction entities of the industrial chain data of each chemical are obtained, specifically including: taking the standard extraction entities on the shortest intermediate paths of all predicted key extraction entities in the original knowledge graph of the industrial chain data of each chemical as the backbone extraction entities of the industrial chain data of the chemical; S4: Based on all the backbone extraction entities of the industrial chain data of each chemical, the final knowledge graph of the industrial chain data of each chemical is obtained.

2. The method for constructing an industrial chain data knowledge graph based on artificial intelligence according to claim 1 is characterized in that: S1: Based on the industrial chain database, all industrial chain data texts of each chemical are obtained. Based on all industrial chain data texts of each chemical and AI technology, all extraction entities and all extraction relationships in all industrial chain data texts of each chemical are obtained. Based on all extraction entities in all industrial chain data texts of each chemical, all standard extraction entities of industrial chain data of each chemical are obtained, including: Obtain all industrial chain data texts of each chemical from the industrial chain database, and based on each industrial chain data text of each chemical and AI technology, obtain all extracted entities and all extracted relationships in each industrial chain data text of each chemical; Taking any two different extracted entities from all extracted entities in all industrial chain data texts of each chemical as a set of input groups of the corresponding chemical, and obtaining all group input groups of each chemical; Based on all group input groups of each chemical and the preset entity determination model, all standard extraction entities of the industrial chain data of each chemical are obtained.

3. The method for constructing an industrial chain data knowledge graph based on artificial intelligence according to claim 2 is characterized in that: Based on all input groups of each chemical and the preset entity judgment model, all standard extracted entities of the industrial chain data of each chemical are obtained, including: Based on each input group of each chemical and a preset entity determination model, obtaining entity determination results for each input group of each chemical; If the entity determination results of each input group of each chemical are the same, the two extracted entities in the corresponding input group of the corresponding chemical are determined to be mutually converted entities; Summarize all the extracted entities that are mutually converted entities in all the group input groups of each chemical as a group of conversion entity groups for each chemical, and obtain all the group conversion entity groups for each chemical; The extracted entity with the least number of characters in each conversion entity group of each chemical is regarded as the preferred entity of the corresponding conversion entity group of the corresponding chemical, and the preferred entity of each conversion entity group of each chemical is regarded as the standard extracted entity of the industrial chain data of the corresponding chemical.

4. The method for constructing an industrial chain data knowledge graph based on artificial intelligence according to claim 1 is characterized in that: S2: Based on all standard extraction entities of the industrial chain data of each chemical and all extraction relationships in the text of all industrial chain data of each chemical, the original knowledge graph of the industrial chain data of each chemical is obtained, including: Based on all standard extraction entities of the industrial chain data of each chemical, all extraction relationships in all industrial chain data texts of each chemical are simplified to obtain all simplified extraction relationships; All standard extracted entities of the industrial chain data of each chemical are taken as nodes, and all simplified extracted relationships are taken as relationship edges to obtain the original knowledge graph of the industrial chain data of each chemical.

5. The method for constructing an industrial chain data knowledge graph based on artificial intelligence according to claim 1 is characterized in that: Based on the number of connections, type score and minimum number of intermediates of each standard extraction entity of the industrial chain data of each chemical, the entity evaluation value of each standard extraction entity of the industrial chain data of the corresponding chemical is obtained, which is: Among them, δ is the entity evaluation value of the current calculation standard extraction entity of the current calculation chemical industry chain data, b is the number of connections of the current calculation standard extraction entity of the current calculation chemical industry chain data, g is the type score value of the current calculation standard extraction entity of the current calculation chemical industry chain data, v is the minimum middle number of the current calculation standard extraction entity of the current calculation chemical industry chain data, ln is the natural logarithm, and the value of the natural constant e is 2.

718.

6. The method for constructing an industrial chain data knowledge graph based on artificial intelligence according to claim 5 is characterized in that: Based on the entity evaluation values ​​of all standard extracted entities of the industrial chain data of each chemical, all predicted key extracted entities of the industrial chain data of each chemical are obtained, including: If the entity evaluation value of each standard extraction entity of the industrial chain data of each chemical is greater than a preset threshold, the corresponding standard extraction entity is regarded as a predicted key extraction entity of the industrial chain data of the corresponding chemical.

7. The method for constructing an industrial chain data knowledge graph based on artificial intelligence according to claim 1 is characterized in that: S4: Based on all the backbone entities of the industrial chain data of each chemical, the final knowledge graph of the industrial chain data of each chemical is obtained, including: In the original knowledge graph of the industrial chain data of each chemical, all standard extraction entities except the backbone extraction entities are deleted to obtain a new knowledge graph, and the new knowledge graph is used as the final knowledge graph of the industrial chain data of the corresponding chemical.

8. An artificial intelligence-based industrial chain data knowledge graph construction system, characterized in that: A method for constructing an industrial chain data knowledge graph based on artificial intelligence for executing any one of claims 1 to 7, comprising: An extraction module is used to obtain all industrial chain data texts of each chemical based on the industrial chain database, obtain all extraction entities and all extraction relationships within all industrial chain data texts of each chemical based on all industrial chain data texts of each chemical and AI technology, and obtain all standard extraction entities of industrial chain data of each chemical based on all extraction entities within all industrial chain data texts of each chemical; The first building module is used to obtain the original knowledge graph of the industrial chain data of each chemical based on all standard extraction entities of the industrial chain data of each chemical and all extraction relationships in the text of all industrial chain data of each chemical; A calculation module, for obtaining all predicted key extraction entities of the industrial chain data of each chemical based on the original knowledge graph of the industrial chain data of each chemical, and obtaining all trunk extraction entities of the industrial chain data of each chemical based on all predicted key extraction entities of the industrial chain data of each chemical; The second building module is used to extract entities based on all the backbones of the industrial chain data of each chemical to obtain the final knowledge graph of the industrial chain data of each chemical.

Images