Relationship graph construction method and system based on semantic analysis of knowledge in electric power field

Abstract

The embodiment of the invention provides a relational graph construction method and system based on semantic analysis of knowledge in the electric power field, and belongs to the technical field of electric power. The relation graph construction method comprises the following steps: acquiring multi-source data about the power field; cleaning and de-noising the multi-source data to obtain corresponding power data; sending the electric power data into a BERT model to complete entity identification of the electric power data; according to the relation between the identified entities, completing extraction of the relation between the entities; and according to the obtained relationship between the entities, constructing a relationship graph about the entities. According to the relation graph construction method, key entity elements can be automatically extracted and the incidence relation of the key entity elements can be intelligently constructed based on the deep knowledge background in the electric power field.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to a method and system for constructing a relationship graph based on semantic analysis of knowledge in the electric power field. Background Art

[0002] With the rapid development of smart grids, distributed energy, and new energy technologies, the processing and utilization of massive data and multi-dimensional information have become the key to improving the operational efficiency of power systems and ensuring the security of energy supply. Traditional data analysis and information management methods can no longer meet the needs of rapid response and accurate decision-making. Therefore, it is necessary to develop a technology that can automatically extract key entity elements and intelligently construct their associations based on the deep knowledge background in the power field. Summary of the invention

[0003] The purpose of the embodiments of the present invention is to provide a method and system for constructing a relationship graph based on semantic analysis of knowledge in the power field. The relationship graph construction method can automatically extract key entity elements and intelligently construct their association relationships based on the deep knowledge background in the power field.

[0004] In order to achieve the above-mentioned purpose, an embodiment of the present invention provides a method for constructing a relationship graph based on semantic analysis of knowledge in the electric power field, and the method for constructing a relationship graph includes:

[0005] Obtain multi-source data on the power sector;

[0006] Cleaning and denoising the multi-source data to obtain corresponding power data;

[0007] The power data is fed into a BERT model to complete entity recognition of the power data;

[0008] Based on the connections between the identified entities, the relationships between the entities are extracted;

[0009] Based on the acquired entities and the relationships between them, a relationship graph about the entities is constructed.

[0010] Optionally, the multi-source data includes: power field documents, technical reports, and standard specifications.

[0011] Optionally, the multi-source data is cleaned and denoised to obtain corresponding power data, including:

[0012] Acquiring the multi-source data;

[0013] Processing the multi-source data to remove spaces, line breaks, and noise in the multi-source data;

[0014] The processed multi-source data is segmented and encoded to obtain corresponding power data.

[0015] Optionally, based on the identified connections between entities, the relationships between entities are extracted, including:

[0016] Acquire entities appearing in the power data;

[0017] According to formula (1), the correlation between any two entities is calculated:

[0018]

[0019] Wherein, K(x) represents the number of times an entity x appears alone in the power data, K(y) represents the number of times another entity y appears alone in the power data, K(x, y) represents the number of times entity x and entity y appear simultaneously in a sentence in the power data, and D(x, y) represents the degree of association;

[0020] Determining whether the correlation degree is greater than a preset threshold;

[0021] When the degree of association is greater than the preset threshold, it is determined that there is a relationship between entity x and entity y, and then a relationship extraction is performed between entity x and entity y.

[0022] Optionally, a relationship graph about the entities is constructed based on the acquired entities and the relationships between the entities, including:

[0023] Performing semantic analysis on the entities in the relationship graph to obtain attributes of each entity and multiple synonyms of the entity;

[0024] According to the attributes and synonyms of the entity, construct a feature vector about the entity, and normalize all the feature vectors;

[0025] Randomly select an entity and its corresponding feature vector;

[0026] Divide the feature vectors of entities other than the selected entity into A groups, each with B feature vectors;

[0027] Concatenate the eigenvectors of each group into a matrix, and the number of rows of the matrix is ​​equal to the number of columns of the eigenvectors of the selected entities;

[0028] Multiplying the feature vector of the selected entity with the matrix to obtain a similarity score for each group;

[0029] Obtaining the similarity score, and filtering entities corresponding to feature vectors whose similarity scores are greater than a preset threshold;

[0030] It is determined that the screened entity has a relationship with the selected entity, an association relationship is established between the two entities, and the relationship diagram is improved according to the association relationship between the entities.

[0031] Optionally, based on the identified connections between entities, the relationships between entities are extracted, including:

[0032] Build an entity relationship extraction model based on LSTM;

[0033] Obtain two entities with a correlation greater than a preset threshold and sentences containing two entities for feature extraction and use them as input to the LSTM entity relationship extraction model;

[0034] By training the entity relationship extraction model, model parameters of the entity relationship extraction model are obtained;

[0035] Construct an entity relationship extraction model based on LSTM according to the model parameters obtained through training;

[0036] Get two new entities and a statement containing the two entities;

[0037] The two newly acquired entities and the sentence containing the two entities are input into the trained entity relationship extraction model to obtain the relationship between the entities.

[0038] Optionally, the entity relationship extraction model includes a cell state C t 、Forget gate t , input gate i t and output gate o t , the function of constructing the entity relationship extraction model is shown in formula (2):

[0039]

[0040] Among them, f t ,i t ,C t ',O t They represent the forget gate, input gate, memory unit and output gate respectively, σ is the sigmoid activation function, tanh is the hyperbolic tangent activation function, and x t Represents the input to the entity relationship extraction model, h t-1 is the hidden layer state, C t-1 ,C t are the unit states, W f ,W i ,W C ,W O is the weight matrix, b f ,b i ,b C ,b O is the bias matrix, the C t It is the unit state, indicating the output result of the entity relationship extraction model.

[0041] On the other hand, the present invention can also provide a relationship graph construction system based on semantic analysis of power field knowledge, the system comprising:

[0042] A knowledge semantic parsing module, used for acquiring the power data and performing entity extraction on the power data;

[0043] Automatic extraction module, used to automatically extract the relationship between entities from text;

[0044] The association relationship intelligent construction module is used to construct a relationship graph based on the association relationship between entities.

[0045] In one aspect, the present invention may also provide a processor for running a program, wherein the program, when being run, is used to execute the above-mentioned method for constructing a relationship graph based on semantic parsing of knowledge in the electric power field.

[0046] Through the above technical scheme, the present invention provides a method and system for constructing a relationship graph based on semantic analysis of knowledge in the electric power field, which can obtain multi-source data about the electric power field, and then clean and denoise the multi-source data, so as to obtain corresponding electric power data. After obtaining the electric power data, the electric power data can be fed into the BERT model, so that entity recognition of the electric power data can be completed. After completing entity recognition, the relationship between entities can be extracted based on the connection between the identified entities. After obtaining the entity and the relationship related to the entity, a relationship graph about the entity can be constructed. This relationship graph construction method can automatically extract key entity elements and intelligently construct their association relationships based on the deep knowledge background in the electric power field.

[0047] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0049] Figure 1 is a flowchart of a method for constructing a relationship graph based on semantic analysis of knowledge in the electric power field according to an embodiment of the present invention;

[0050] Figure 2 It is a flowchart of cleaning and denoising of a method for constructing a relationship graph based on semantic parsing of knowledge in the electric power field according to an embodiment of the present invention;

[0051] Figure 3It is a first flow chart of performing relationship extraction according to a method for constructing a relationship graph based on semantic analysis of knowledge in the electric power field according to an embodiment of the present invention;

[0052] Figure 4 It is a flowchart of determining the association between entities in a method for constructing a relationship graph based on semantic analysis of knowledge in the electric power field according to an embodiment of the present invention;

[0053] Figure 5 It is a second flow chart for performing relationship extraction according to a method for constructing a relationship graph based on semantic analysis of knowledge in the electric power field according to an embodiment of the present invention. DETAILED DESCRIPTION

[0054] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0055] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application are in compliance with the relevant provisions of national laws and regulations. In the embodiments of this application, some existing solutions in the industry such as certain software, components, and models may be mentioned, which should be considered as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of this application, but it does not mean that the applicant has or will necessarily use the solution.

[0056] Figure 1 The flowchart of a method for constructing a relationship graph based on semantic analysis of knowledge in the electric power field according to an embodiment of the present invention. In the present invention, the process of constructing the relationship graph may include:

[0057] In step S1, multi-source data on the power field is obtained.

[0058] In step S2, the multi-source data is cleaned and denoised to obtain corresponding power data.

[0059] In step S3, the power data is fed into the BERT model to complete entity recognition of the power data.

[0060] In step S4, the relationships between entities are extracted based on the identified connections between the entities.

[0061] In step S5, a relationship graph of entities is constructed based on the acquired entities and the relationships between entities.

[0062] In the present invention, when constructing a relationship graph, multi-source data about the power field can be obtained first, and then the multi-source data can be cleaned and denoised, so that the corresponding power data can be obtained. After obtaining the power data, the power data can be fed into the BERT model, so that entity recognition of the power data can be completed. After completing entity recognition, the relationship between entities can be extracted based on the connection between the identified entities. After obtaining the entity and the relationship related to the entity, a relationship graph about the entity can be constructed. This relationship graph construction method can automatically extract key entity elements and intelligently construct their association relationships based on the deep knowledge background in the power field.

[0063] In one embodiment of the present invention, the multi-source data may include data such as power field documents, technical reports, standard specifications, etc.

[0064] In one embodiment of the present invention, Figure 2 As shown, the cleaning and denoising process can include:

[0065] In step S6, the multi-source data is acquired.

[0066] In step S7, the multi-source data is processed to remove spaces, line breaks, and noise in the multi-source data.

[0067] In step S8, the processed multi-source data is segmented and encoded to obtain corresponding power data.

[0068] In the present invention, when cleaning multi-source data, the multi-source data can be acquired first, and then the multi-source data can be processed to remove spaces, line breaks and noise factors in the multi-source data. Then, the processed multi-source data can be segmented and encoded to obtain corresponding power data.

[0069] In one embodiment of the present invention, Figure 3 As shown, the first process of performing relationship extraction may include:

[0070] In step S9, entities appearing in the power data are acquired.

[0071] In step S10, the association degree between any two entities is calculated according to formula (1):

[0072]

[0073] Among them, K(x) represents the number of times an entity x appears alone in the power data, K(y) represents the number of times another entity y appears alone in the power data, K(x, y) represents the number of times entity x and entity y appear together in a sentence in the power data, and D(x, y) represents the degree of association.

[0074] In step S11, it is determined whether the correlation degree is greater than a preset threshold.

[0075] In step S12, when the degree of association is greater than a preset threshold, it is determined that there is a relationship between entity x and entity y, and a relationship extraction is performed between entity x and entity y.

[0076] In the present invention, when extracting the relationship between entities, it can be confirmed that there is indeed a relationship between the two entities. Therefore, the entities appearing in the power data can be obtained first, and then the correlation between any two entities can be calculated according to formula (1). The correlation can be a criterion for judging whether there is a relationship between the two entities. When the correlation is greater than a preset threshold, it can be judged that there is a relationship between the two entities corresponding to the correlation, so the relationship between the two entities can be extracted.

[0077] In one embodiment of the present invention, Figure 4 As shown, the process of determining the association between entities may include:

[0078] In step S13, semantic analysis is performed on the entities in the relationship graph to obtain the attributes of each entity and a plurality of synonyms related to the entity.

[0079] In step S14, a feature vector about the entity is constructed according to the attributes and synonyms of the entity, and all feature vectors are normalized.

[0080] In step S15, an entity and its corresponding feature vector are randomly selected.

[0081] In step S16, the feature vectors of entities other than the selected entity are divided into A groups, each group having B feature vectors.

[0082] In step S17, the feature vectors of each group are concatenated into a matrix, and the number of rows of the matrix is ​​equal to the number of columns of the feature vectors of the selected entities.

[0083] In step S18, the feature vector of the selected entity is multiplied by the matrix to obtain a similarity score for each group.

[0084] In step S19, a similarity score is obtained, and entities corresponding to feature vectors having a similarity score greater than a preset threshold are screened.

[0085] In step S20, it is determined that the screened entity has a relationship with the selected entity, an association relationship is established between the two entities, and the relationship graph is improved according to the association relationship between the entities.

[0086] In the present invention, after the relationship graph is completed, a semantic similarity calculation method can be introduced to perform similarity analysis on different entities and concepts. According to the similarity results, an association relationship is established, for example, similar fault types, equipment types, etc. are associated, so that the relationship graph can be further improved. In the improvement process, the entities in the relationship graph can be semantically analyzed, so that each entity attribute and multiple synonyms about the entity can be obtained, and then the feature vector about the entity can be constructed according to the attributes and synonyms of the entity, and all the feature vectors can be normalized, so that different vectors can be adjusted to a unified format. After normalization, an entity and its corresponding feature vector can be randomly selected, and then the feature vectors of other entities except the selected entity can be divided into A groups, each group of B feature vectors. After grouping, the feature vectors of each group can be spliced ​​into a matrix, and the number of rows of the matrix can be equal to the number of columns of the feature vector of the selected entity. After obtaining the matrix, the feature vector of the selected entity can be multiplied by the matrix, so that the similarity score of each group can be obtained. Compared with the traditional Euclidean distance, this method can find two related entities faster. After obtaining the similarity score, the entities corresponding to the feature vectors with similarity scores greater than a preset threshold can be filtered out, so that it can be determined that there is a relationship between the filtered entity and the selected entity, an association relationship between the two entities can be established, and then the relationship graph can be improved based on the association relationship between the entities.

[0087] In one embodiment of the present invention, Figure 5 As shown, the second process of performing relationship extraction may include:

[0088] In step S21, an entity relationship extraction model based on LSTM is constructed.

[0089] In step S22, two entities with a correlation greater than a preset threshold and a sentence containing the two entities are obtained for feature extraction and used as input of the LSTM entity relationship extraction model.

[0090] In step S23, the entity relationship extraction model is trained to obtain model parameters of the entity relationship extraction model.

[0091] In step S24, an LSTM-based entity relationship extraction model is constructed according to the model parameters obtained through training.

[0092] In step S25, two new entities and a statement containing the two entities are obtained.

[0093] In step S26, the two newly acquired entities and the sentence containing the two entities are input into the trained entity relationship extraction model to obtain the relationship between the entities.

[0094] In the present invention, when it is determined that there is a relationship between two entities, it is necessary to extract the relationship between the two entities. In the extraction process, an entity relationship extraction model based on LSTM can be first constructed, and then two entities with a correlation greater than a preset threshold and a statement containing the two entities can be obtained for feature extraction, and the result after feature extraction can be used as the input of the entity relationship extraction model based on LSTM. By training the entity relationship extraction model, parameters about the entity relationship extraction model can be obtained. After obtaining the parameters, an entity relationship extraction model based on LSTM can be constructed according to the parameters obtained by the training, that is, the training of the entity relationship extraction model is completed. After the training is completed, two new entities and a statement containing the two entities can be obtained, and then the obtained two entities and the statement containing the two entities are input into the trained entity relationship extraction model, so that the relationship between the two entities can be obtained.

[0095] In one embodiment of the present invention, the entity relationship extraction model may include a cell state C t 、Forget gate t , input gate i t and output gate o t , the function constructed for the entity relationship extraction model can be shown as formula (2):

[0096]

[0097] Among them, f t ,i t ,C t ',O t They represent the forget gate, input gate, memory unit and output gate respectively, σ is the sigmoid activation function, tanh is the hyperbolic tangent activation function, and x t Represents the input to the entity relationship extraction model, h t-1 is the hidden layer state, C t-1 ,C t are the unit states, W f ,W i ,W C ,W O is the weight matrix, b f ,b i ,b C ,b O is the bias matrix, the C t It is the unit state, indicating the output result of the entity relationship extraction model.

[0098] On the other hand, the present invention can also provide a relationship graph construction system based on semantic analysis of knowledge in the power field, the system comprising: a knowledge semantic analysis module, an automatic extraction module and an association relationship intelligent construction module. The knowledge semantic analysis module is used to obtain the power data and perform entity extraction on the power data. The automatic extraction module is used to automatically extract the association relationship between entities from the text. The association relationship intelligent construction module is used to construct a relationship graph based on the association relationship between entities.

[0099] In one aspect, the present invention may also provide a processor for running a program, wherein the program, when being run, is used to execute the above-mentioned method for constructing a relationship graph based on semantic parsing of knowledge in the electric power field.

[0100] Through the above technical scheme, the present invention provides a method and system for constructing a relationship graph based on semantic analysis of knowledge in the electric power field, which can obtain multi-source data about the electric power field, and then clean and denoise the multi-source data, so as to obtain corresponding electric power data. After obtaining the electric power data, the electric power data can be fed into the BERT model, so that entity recognition of the electric power data can be completed. After completing entity recognition, the relationship between entities can be extracted based on the connection between the identified entities. After obtaining the entity and the relationship related to the entity, a relationship graph about the entity can be constructed. This relationship graph construction method can automatically extract key entity elements and intelligently construct their association relationships based on the deep knowledge background in the electric power field.

[0101] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0102] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0103] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0104] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0105] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0106] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0107] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0108] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0109] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for constructing a relationship graph based on semantic analysis of knowledge in the electric power field, characterized in that: The relationship graph construction method comprises: Obtain multi-source data on the power sector; Cleaning and denoising the multi-source data to obtain corresponding power data; The power data is fed into a BERT model to complete entity recognition of the power data; Based on the connections between the identified entities, the relationships between the entities are extracted; Based on the acquired entities and the relationships between them, a relationship graph about the entities is constructed.

2. The relationship graph construction method according to claim 1, characterized in that: The multi-source data includes: documents in the electric power field, technical reports, and standard specifications.

3. The relationship graph construction method according to claim 1, characterized in that: The multi-source data is cleaned and denoised to obtain corresponding power data, including: Acquiring the multi-source data; Processing the multi-source data to remove spaces, line breaks, and noise in the multi-source data; The processed multi-source data is segmented and encoded to obtain corresponding power data.

4. The relationship graph construction method according to claim 1, characterized in that: Based on the connections between the identified entities, the relationships between the entities are extracted, including: Acquire entities appearing in the power data; According to formula (1), the correlation between any two entities is calculated: Wherein, K(x) represents the number of times an entity x appears alone in the power data, K(y) represents the number of times another entity y appears alone in the power data, K(x, y) represents the number of times entity x and entity y appear simultaneously in a sentence in the power data, and D(x, y) represents the degree of association; Determining whether the correlation degree is greater than a preset threshold; When the degree of association is greater than the preset threshold, it is determined that there is a relationship between entity x and entity y, and then a relationship extraction is performed between entity x and entity y.

5. The relationship graph construction method according to claim 1, characterized in that: Based on the acquired entities and the relationships between them, a relationship graph about the entities is constructed, including: Performing semantic analysis on the entities in the relationship graph to obtain attributes of each entity and multiple synonyms of the entity; According to the attributes and synonyms of the entity, construct a feature vector about the entity, and normalize all the feature vectors; Randomly select an entity and its corresponding feature vector; Divide the feature vectors of entities other than the selected entity into A groups, each with B feature vectors; Concatenate the eigenvectors of each group into a matrix, and the number of rows of the matrix is ​​equal to the number of columns of the eigenvectors of the selected entities; Multiplying the feature vector of the selected entity with the matrix to obtain a similarity score for each group; Obtaining the similarity score, and filtering entities corresponding to feature vectors whose similarity scores are greater than a preset threshold; It is determined that the screened entity has a relationship with the selected entity, an association relationship is established between the two entities, and the relationship diagram is improved according to the association relationship between the entities.

6. The relationship graph construction method according to claim 4, characterized in that: Based on the connections between the identified entities, the relationships between the entities are extracted, including: Build an entity relationship extraction model based on LSTM; Obtain two entities with a correlation greater than a preset threshold and sentences containing two entities for feature extraction and use them as input to the LSTM entity relationship extraction model; By training the entity relationship extraction model, model parameters of the entity relationship extraction model are obtained; Construct an entity relationship extraction model based on LSTM according to the model parameters obtained through training; Get two new entities and a statement containing the two entities; The two newly acquired entities and the sentences containing the two entities are input into the trained entity relationship extraction model to obtain the relationship between the entities.

7. The relationship graph construction method according to claim 6, characterized in that: The entity relationship extraction model includes a cell state C t 、Forget gate t , input gate i t and output gate o t , the function of constructing the entity relationship extraction model is shown in formula (2): Among them, f t ,i t ,C t ',O t They represent the forget gate, input gate, memory unit and output gate respectively, σ is the sigmoid activation function, tanh is the hyperbolic tangent activation function, and x t Represents the input to the entity relationship extraction model, h t-1 is the hidden layer state, C t-1 ,C t are the unit states, W f ,W i ,W C ,W O is the weight matrix, b f ,b i ,b C ,b O is the bias matrix, the C t It is the unit state, indicating the output result of the entity relationship extraction model.

8. A relationship graph construction system based on semantic analysis of knowledge in the power field, characterized in that: The system comprises: A knowledge semantic parsing module, used for acquiring the power data and performing entity extraction on the power data; Automatic extraction module, used to automatically extract the relationship between entities from text; The association relationship intelligent construction module is used to construct a relationship graph based on the association relationship between entities.

9. A processor, characterized in that: Used to run a program, wherein the program, when run, is used to execute a method for constructing a relationship graph based on semantic analysis of knowledge in the electric power field as described in any one of claims 1 to 7.

Images