Knowledge graph-based electric power science and technology project document duplicate checking method and system, and storage medium

Abstract

The invention discloses an electric power science and technology project document duplicate checking method based on a knowledge graph. The method comprises the steps that an existing electric power science and technology project document is acquired and preprocessed; performing knowledge extraction to construct a dictionary and a knowledge graph of the electric power science and technology project document; fusing the dictionary of the electric power science and technology project documents with the universal language lexicon to obtain a professional lexicon of the electric power science and technology project documents; performing word segmentation vectorization processing on the target electric power science and technology project document; calculating the document similarity of the target electric power science and technology project document; and duplicate checking of the target electric power science and technology project document is completed according to the document similarity. The invention also provides a system for realizing the power science and technology project document duplicate checking method based on the knowledge graph, and a storage medium comprising the power science and technology project document duplicate checking method based on the knowledge graph. According to the method, duplicate checking of the electric power science and technology project document is achieved, the reliability is higher, the accuracy is better, and the efficiency is higher.

Description

Technical Field

[0001] The present invention belongs to the field of electrical automation, and specifically relates to a method, system and storage medium for checking duplicate documents of electric power technology projects based on knowledge graphs. Background Art

[0002] With the development of economy and technology and the improvement of people's living standards, electricity has become an indispensable secondary energy source in people's production and life, bringing endless convenience to people's production and life. Therefore, ensuring the stable and reliable supply of electricity has become one of the most important tasks of the power system.

[0003] At present, the power system has invested a lot of R&D time and expenses to ensure its own stable and reliable operation. The audit and approval of scientific and technological projects has a large volume and the requirements are becoming more and more stringent. Therefore, it is particularly important to check the duplicate documents of the power technology project of the power system.

[0004] At present, the traditional solution for checking duplicate documents in power technology projects is still manual checking, that is, the relevant personnel of the power system manually check the duplicate documents of the target power technology project. However, this manual checking solution is not only time-consuming and labor-intensive, but also inefficient, and has poor reliability and accuracy. Summary of the invention

[0005] One of the purposes of the present invention is to provide a method for checking duplicate documents of electric power technology projects based on knowledge graphs with high reliability, good accuracy and high efficiency.

[0006] The second objective of the present invention is to provide a system for implementing the method for checking duplicate documents of electric power technology projects based on knowledge graph.

[0007] The third object of the present invention is to provide a storage medium that includes the method for checking duplicate documents of electric power technology projects based on knowledge graphs.

[0008] The method for checking duplicate documents of electric power technology projects based on knowledge graph provided by the present invention comprises the following steps:

[0009] S1. Obtain existing power technology project documents and perform data preprocessing;

[0010] S2. extract knowledge from the preprocessed power technology project documents to construct a dictionary and knowledge graph of the power technology project documents;

[0011] S3. The dictionary of the power technology project document obtained in step S2 is merged with the general language thesaurus to obtain a professional thesaurus of the power technology project document;

[0012] S4. Based on the professional vocabulary obtained in step S3, the target electric power technology project document is segmented and vectorized;

[0013] S5. Calculate the document similarity of the target electric power technology project document based on the result of word segmentation vectorization processing;

[0014] S6. Based on the document similarity obtained in step S5, complete the duplicate check of the target electric power technology project document.

[0015] The method for checking duplicate documents of electric power technology projects based on knowledge graph also includes the following steps:

[0016] S7. Extract knowledge from the target electric power technology project document to update the dictionary and knowledge graph of the constructed electric power technology project document.

[0017] The step S1 of obtaining existing electric power technology project documents and performing data preprocessing specifically includes the following steps:

[0018] Scan the existing paper-based power technology project documents to obtain image files, and convert them into electronic documents using optical character recognition technology;

[0019] For the obtained electronic documents, a word frequency statistical language model is used to annotate and screen them, and the initial corpus of existing electric power technology project documents is obtained to complete data preprocessing.

[0020] The step S2 of extracting knowledge from the preprocessed electric power technology project document to construct a dictionary and knowledge graph of the electric power technology project document specifically includes the following steps:

[0021] Extract entities and attributes from preprocessed power technology project documents:

[0022] The initial corpus of electric power technology project documents is matched by templates, entities and attributes of the initial electric power technology project documents are extracted and obtained, and an entity-attribute network is established based on semantic relationships;

[0023] Compute confidence ranking, semantic similarity, and semantic membership;

[0024] Obtain the entries whose confidence ranking, semantic similarity and semantic membership are all greater than the set threshold, and use the obtained entries as entities and attributes of the knowledge graph of the electric power technology project document, and construct the obtained entries into a dictionary of the electric power technology project document;

[0025] Extract the relationship between entities and attributes of the preprocessed power technology project documents:

[0026] The word embedding method combined with the probabilistic topic model is adopted. The word embedding model is used to represent the entity and attribute terms of the power technology project as continuous vectors. The probabilistic topic model is used to calculate the degree of association between the terms and the potential topics. The semantic relationship between the terms is obtained through semantic relationship measurement.

[0027] Perform hierarchical modeling to extract relationships between entities and attributes.

[0028] Step S3 of fusing the dictionary of the electric power technology project document obtained in step S2 with the general language word library to obtain the professional word library of the electric power technology project document specifically includes the following steps:

[0029] The obtained dictionary of the electric power technology project documents is loaded and expanded into the vocabulary of the language processing model. At the same time, the training set of the language processing model is added to achieve fine-tuning and optimization of the model, and the fusion of the dictionary of the electric power technology project documents and the general language vocabulary is achieved, thereby obtaining a professional vocabulary of the electric power technology project documents.

[0030] The step S4 described in which the target electric power technology project document is subjected to word segmentation and vectorization processing based on the professional vocabulary obtained in step S3 specifically includes the following steps:

[0031] According to the professional vocabulary obtained in step S3, the target electric power technology project document and the compared electric power technology project document are subjected to word segmentation and vectorization processing, thereby constructing a corresponding term vectorization matrix.

[0032] The step S5 of calculating the document similarity of the target electric power technology project document according to the result of the word segmentation vectorization processing specifically includes the following steps:

[0033] Select the corresponding similarity among cosine similarity, edit distance similarity, longest common subsequence similarity and mixed similarity, and calculate the document similarity of the target electric power technology project document;

[0034] Among them, the cosine similarity cos(Tx,Ty) is calculated using the following formula:

[0035]

[0036] Where Tx is the target power technology project document; Ty is the compared power technology project document; x is the vector corresponding to the target power technology project document; y is the vector corresponding to the compared power technology project document; |x is the modulus of vector x; |y| is the modulus of vector y;

[0037] The edit distance similarity SimLevenshtein(Tx,Ty) is calculated using the following formula:

[0038]

[0039] Where Levenshtein(Tx,Ty) is the edit distance between document Tx and document Ty; Len is the length of the document;

[0040] The longest common subsequence similarity SimLCS(Tx,Ty) is calculated using the following formula:

[0041]

[0042] Where LCS(Tx,Ty) is the longest common subsequence of document Tx and document Ty;

[0043] The mixed similarity Sim(Tx,Ty) is calculated using the following formula:

[0044] Sim(Tx,Ty)=a·cos(Tx,Ty)+b·SimLevenshtein(Tx,Ty)+c·SimLCS(Tx,Ty), where a is the first weight value; b is the second weight value; and c is the third weight value.

[0045] In the calculation formula of the mixed similarity Sim(Tx,Ty), the custom function Simother(Tx,Ty) and the corresponding weight d are introduced; by setting the custom function and weight, it is possible to implement document duplication checking for special requirements in the future, or to introduce a document duplication checking method with better performance in the future, thereby realizing customized processing of document duplication checking.

[0046] Step S6, according to the document similarity obtained in step S5, completes the duplicate checking of the target electric power technology project document, which specifically includes the following steps:

[0047] According to the document similarity obtained in step S5, the Myers difference method is used to calculate the similar parts and different parts in the target electric power technology project document and the compared electric power technology project document, and mark them, thereby completing the duplicate check of the target electric power technology project document.

[0048] The present invention also provides a system for implementing the method for checking duplicate documents of electric power technology projects based on knowledge graph, including a data acquisition module, a knowledge extraction module, a fusion module, a word segmentation module, a similarity calculation module, a document duplication checking module and an update module; the data acquisition module, the knowledge extraction module, the fusion module, the word segmentation module, the similarity calculation module, the document duplication checking module and the update module are connected in series in sequence; the data acquisition module is used to acquire existing electric power technology project documents, perform data preprocessing, and upload the data information to the knowledge extraction module; the knowledge extraction module is used to extract knowledge from the preprocessed electric power technology project documents based on the received data information, so as to construct a dictionary and knowledge graph of the electric power technology project documents, and upload the data information to the fusion module; the fusion module is used to compare the obtained dictionary of the electric power technology project documents with the common language based on the received data information The word library is integrated to obtain the professional word library of the electric power technology project document, and the data information is uploaded to the word segmentation module; the word segmentation module is used to perform word segmentation and vectorization processing on the target electric power technology project document based on the received data information and the obtained professional word library, and upload the data information to the similarity calculation module; the similarity calculation module is used to calculate the document similarity of the target electric power technology project document according to the received data information and the result of the word segmentation and vectorization processing, and upload the data information to the document duplicate checking module; the document duplicate checking module is used to complete the duplicate checking of the target electric power technology project document according to the received data information and the obtained document similarity, and upload the data information to the update module; the update module is used to extract knowledge from the target electric power technology project document according to the received data information, so as to update the constructed dictionary and knowledge graph of the electric power technology project document.

[0049] The present invention also provides a storage medium on which a computer program is stored; when the computer program is executed by a processor, the method for checking duplicate documents of electric power technology projects based on knowledge graph is implemented.

[0050] The knowledge graph-based electric power technology project document duplicate checking method, system and storage medium provided by the present invention are based on natural language processing technology, the construction of a knowledge graph and a professional vocabulary, and the calculation of document similarity. It not only realizes the duplicate checking of electric power technology project documents, but also has higher reliability, better accuracy and higher efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 The figure is a schematic diagram of the method flow of the present invention.

[0052] Figure 2 Schematic diagram of the functional modules of the system of the present invention. DETAILED DESCRIPTION

[0053] like Figure 1The method flow diagram of the method of the present invention is shown as follows: the method for checking duplicate documents of electric power technology projects based on knowledge graph provided by the present invention comprises the following steps:

[0054] S1. Obtain existing power technology project documents and perform data preprocessing; specifically, the following steps are included:

[0055] Scan the existing paper power technology project documents to obtain image files, and use optical character recognition technology to convert them into electronic documents; with the help of Python program platform, electronic documents such as Word, PDF, PPT, TXT and Excel can be converted into computer-recognizable data sets with the help of installation packages such as python-word, python-pptx, PyPDF2, pandas, etc.

[0056] For the obtained electronic documents, a word frequency statistical language model is used for annotation and screening to obtain the initial corpus of existing electric power technology project documents and complete data preprocessing; the annotation process is carried out in the word frequency statistical language model, and common ones such as the N-gram language model and the Kneser-Ney smoothing language model can be implemented;

[0057] S2. Extract knowledge from the preprocessed power technology project documents to construct a dictionary and knowledge graph of the power technology project documents; specifically, the steps include:

[0058] Extract entities and attributes from preprocessed power technology project documents:

[0059] The initial corpus of electric power technology project documents is matched by templates, entities and attributes of the initial electric power technology project documents are extracted and obtained, and an entity-attribute network is established based on semantic relationships;

[0060] Compute confidence ranking, semantic similarity, and semantic membership;

[0061] Obtain the entries whose confidence ranking, semantic similarity and semantic membership are all greater than the set threshold, and use the obtained entries as the entities and attributes of the knowledge graph of the electric power technology project document, and at the same time construct the obtained entries into the dictionary of the electric power technology project document; in the specific implementation, the first step is to sort the terms according to the confidence propagation: in the electric power technology project, there is a problem that there is little information available for low-frequency entries, and the extraction through the confidence propagation method has a better effect; for the dictionary to be processed, first calculate the phrase nature by point mutual information, and then build the semantic network of the entries, and then perform confidence sorting, and the entries that meet the set standards enter the similarity judgment, and those that do not meet the standards enter the next entry judgment; there is a problem of diversified expression of professional terms in the electric power technology project, and the similarity judgment can effectively solve this problem. For entries that exceed the set judgment threshold, the similar entries are homogenized and no longer enter the next step; finally, the membership of the entry is judged. The greater the membership of the entry, the greater the relevance of the entry to the electric power technology project. If it is higher than the judgment standard, it can enter the electric power technology project dictionary;

[0062] Extract the relationship between entities and attributes of the preprocessed power technology project documents:

[0063] The word embedding method combined with the probabilistic topic model is adopted. The word embedding model is used to represent the entity and attribute terms of the power technology project as continuous vectors. The probabilistic topic model is used to calculate the degree of association between the terms and the potential topics. The semantic relationship between the terms is obtained through semantic relationship measurement.

[0064] Perform hierarchical modeling to extract relationships between entities and attributes;

[0065] In the specific implementation, the electric power technology project terms may contain multiple potential semantics, so the relationship between the terms must be semantically modeled first; semantic modeling generally uses a statistical probabilistic language model or a language vector space model to fully understand the semantic relationship between terms; word embedding expresses the terms as continuous vectors, and the topic probability calculates the relevance of the terms and potential topics through the Gibbs probability; word embedding and topic probability more fully consider the potential semantics of the terms, and then quantify them to model the language hierarchy; finally, the hierarchical modeling is to establish a knowledge system from the perspective of the entire electric power technology project, that is, the knowledge graph of the electric power technology project; hierarchical modeling is carried out in an active learning framework, and by iteratively optimizing the evaluation indicators, a local optimal system that meets the requirements is established;

[0066] S3. The dictionary of the power technology project document obtained in step S2 is merged with the general language thesaurus to obtain a professional thesaurus of the power technology project document; specifically comprising the following steps:

[0067] The obtained dictionary of the electric power technology project document is loaded and expanded into the vocabulary of the language processing model, and the training set of the language processing model is added to achieve fine-tuning and optimization of the model, and realize the fusion of the dictionary of the electric power technology project document and the general language vocabulary, so as to obtain the professional vocabulary of the electric power technology project document;

[0068] In the specific implementation, the method of expanding the vocabulary is adopted, and the entries in the power technology project dictionary are added to the general language vocabulary HowNet (hownet) and BCC (Beijing Language and Culture University Corpus Center). The new vocabulary constructed has a good coverage of power technology projects, general scientific research terms and general language expressions;

[0069] S4. Based on the professional vocabulary obtained in step S3, the target electric power technology project document is subjected to word segmentation and vectorization processing; specifically, the following steps are included:

[0070] According to the professional vocabulary obtained in step S3, the target electric power technology project document and the compared electric power technology project document are subjected to word segmentation and vectorization processing, thereby constructing a corresponding term vectorization matrix;

[0071] S5. Calculate the document similarity of the target electric power technology project document based on the result of word segmentation vectorization processing; specifically including the following steps:

[0072] Select the corresponding similarity among cosine similarity, edit distance similarity, longest common subsequence similarity and mixed similarity, and calculate the document similarity of the target electric power technology project document;

[0073] Among them, the cosine similarity cos(Tx,Ty) is calculated using the following formula:

[0074]

[0075] Where Tx is the target electric power technology project document; Ty is the compared electric power technology project document; x is the vector corresponding to the target electric power technology project document; y is the vector corresponding to the compared electric power technology project document; |x is the modulus of vector x; |y| is the modulus of vector y; cosine similarity calculation efficiency is high and is applicable to both long and short texts, but it is not sensitive to the word order of this article;

[0076] The edit distance similarity SimLevenshtein(Tx,Ty) is calculated using the following formula:

[0077]

[0078] Where Levenshtein(Tx,Ty) is the edit distance between document Tx and document Ty; Len is the length of the document; edit distance similarity can detect subtle differences in text, but the computational complexity is high and the efficiency is low when processing large-scale text;

[0079] The longest common subsequence similarity SimLCS(Tx,Ty) is calculated using the following formula:

[0080]

[0081] Where LCS(Tx,Ty) is the longest common subsequence of document Tx and document Ty. The longest common subsequence similarity can be applied to any string language and can distinguish the order of characters in the text, but it takes up more computer resources.

[0082] The mixed similarity Sim(Tx,Ty) is calculated using the following formula:

[0083] Sim(Tx,Ty)=a·cos(Tx,Ty)+b·SimLevenshtein(Tx,Ty)+c·SimLCS(Tx,Ty)where a is the first weight value; b is the second weight value; c is the third weight value; if the focus is on project innovation, the calculation weight of cosine similarity is increased; if the focus is on text expression, the weight of edit distance similarity is increased; if the focus is on keyword detection, the weight of the longest common subsequence similarity can be increased;

[0084] In addition, in the calculation formula of the mixed similarity Sim(Tx,Ty), the custom function Simother(Tx,Ty) and the corresponding weight d are introduced; by setting the custom function and weight, it is possible to implement document duplication checking for special requirements in the future, or introduce a document duplication checking method with better performance in the future, thereby realizing customized processing of document duplication checking;

[0085] S6. According to the document similarity obtained in step S5, the target electric power technology project document is checked for duplicates; specifically, the following steps are included:

[0086] According to the document similarity obtained in step S5, the Myers difference method is used to calculate the similar parts and the different parts between the target electric power technology project document and the compared electric power technology project document, and mark them, thereby completing the duplicate check of the target electric power technology project document;

[0087] S7. Extract knowledge from the target electric power technology project document to update the dictionary and knowledge graph of the constructed electric power technology project document.

[0088] like Figure 2The functional module schematic diagram of the system of the present invention is shown as follows: the system disclosed in the present invention for implementing the method for checking duplicate documents of electric power technology projects based on knowledge graph comprises a data acquisition module, a knowledge extraction module, a fusion module, a word segmentation module, a similarity calculation module, a document checking duplicate module and an update module; the data acquisition module, the knowledge extraction module, the fusion module, the word segmentation module, the similarity calculation module, the document checking duplicate module and the update module are connected in series in sequence; the data acquisition module is used to acquire existing electric power technology project documents, perform data preprocessing, and upload the data information to the knowledge extraction module; the knowledge extraction module is used to extract knowledge from the preprocessed electric power technology project documents according to the received data information, so as to construct a dictionary and knowledge graph of the electric power technology project documents, and upload the data information to the fusion module; the fusion module is used to obtain the obtained electric power technology project documents according to the received data information The dictionary of the document is integrated with the general language vocabulary to obtain the professional vocabulary of the electric power technology project document, and the data information is uploaded to the word segmentation module; the word segmentation module is used to perform word segmentation and vectorization processing on the target electric power technology project document based on the received data information and the obtained professional vocabulary, and upload the data information to the similarity calculation module; the similarity calculation module is used to calculate the document similarity of the target electric power technology project document according to the received data information and the result of the word segmentation and vectorization processing, and upload the data information to the document duplication checking module; the document duplication checking module is used to complete the duplication checking of the target electric power technology project document according to the received data information and the obtained document similarity, and upload the data information to the update module; the update module is used to extract knowledge from the target electric power technology project document according to the received data information, so as to update the dictionary and knowledge graph of the constructed electric power technology project document.

[0089] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0090] 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.

[0091] 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.

[0092] 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. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0093] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0094] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for checking duplicate content in power technology project documents based on knowledge graph, comprising the following steps: S1. Obtain existing power technology project documents and perform data preprocessing; S2. extract knowledge from the preprocessed power technology project documents to construct a dictionary and knowledge graph of the power technology project documents; S3. The dictionary of the power technology project document obtained in step S2 is merged with the general language thesaurus to obtain a professional thesaurus of the power technology project document; S4. Based on the professional vocabulary obtained in step S3, the target electric power technology project document is segmented and vectorized; S5. Calculate the document similarity of the target electric power technology project document based on the result of word segmentation vectorization processing; S6. Based on the document similarity obtained in step S5, complete the duplicate check of the target electric power technology project document.

2. The method for checking duplicate documents of electric power technology projects based on knowledge graph according to claim 1 is characterized in that: The following steps are also included: S7. Extract knowledge from the target electric power technology project document to update the dictionary and knowledge graph of the constructed electric power technology project document.

3. The method for checking duplicate documents of electric power technology projects based on knowledge graph according to claim 1 is characterized in that The step S1 of obtaining existing electric power technology project documents and performing data preprocessing specifically includes the following steps: Scan the existing paper-based power technology project documents to obtain image files, and convert them into electronic documents using optical character recognition technology; For the obtained electronic documents, a word frequency statistical language model is used to annotate and screen them, and the initial corpus of existing electric power technology project documents is obtained to complete data preprocessing.

4. The method for checking duplicate documents of electric power technology projects based on knowledge graph according to claim 3 is characterized in that The step S2 of extracting knowledge from the preprocessed electric power technology project document to construct a dictionary and knowledge graph of the electric power technology project document specifically includes the following steps: Extract entities and attributes from preprocessed power technology project documents: The initial corpus of electric power technology project documents is matched by templates, entities and attributes of the initial electric power technology project documents are extracted and obtained, and an entity-attribute network is established based on semantic relationships; Compute confidence ranking, semantic similarity, and semantic membership; Obtain the terms whose confidence ranking, semantic similarity and semantic membership are all greater than the set threshold, and use the obtained terms as entities and attributes of the knowledge graph of the electric power technology project document, and construct the obtained terms into a dictionary of the electric power technology project document; Extract the relationship between entities and attributes of the preprocessed power technology project documents: The word embedding method combined with the probabilistic topic model is adopted. The word embedding model is used to represent the entity and attribute terms of the power technology project as continuous vectors. The probabilistic topic model is used to calculate the degree of association between the terms and the potential topics. The semantic relationship between the terms is obtained through semantic relationship measurement. Perform hierarchical modeling to extract relationships between entities and attributes.

5. The method for checking duplicate documents of electric power technology projects based on knowledge graph according to claim 4 is characterized in that Step S3 of fusing the dictionary of the electric power technology project document obtained in step S2 with the general language word library to obtain the professional word library of the electric power technology project document specifically includes the following steps: The obtained dictionary of the electric power technology project documents is loaded and expanded into the vocabulary of the language processing model. At the same time, the training set of the language processing model is added to achieve fine-tuning and optimization of the model, and the fusion of the dictionary of the electric power technology project documents and the general language vocabulary is achieved, thereby obtaining a professional vocabulary of the electric power technology project documents.

6. The method for checking duplicate documents of electric power technology projects based on knowledge graph according to claim 5 is characterized in that The step S4 described in which the target electric power technology project document is subjected to word segmentation and vectorization processing based on the professional vocabulary obtained in step S3 specifically includes the following steps: According to the professional vocabulary obtained in step S3, the target electric power technology project document and the compared electric power technology project document are subjected to word segmentation and vectorization processing, thereby constructing a corresponding term vectorization matrix.

7. The method for checking duplicate documents of electric power technology projects based on knowledge graph according to claim 6 is characterized in that The step S5 of calculating the document similarity of the target electric power technology project document according to the result of the word segmentation vectorization processing specifically includes the following steps: Select the corresponding similarity among cosine similarity, edit distance similarity, longest common subsequence similarity and mixed similarity, and calculate the document similarity of the target electric power technology project document; Among them, the cosine similarity cos(Tx,Ty) is calculated using the following formula: Where Tx is the target power technology project document; Ty is the compared power technology project document; x is the vector corresponding to the target power technology project document; y is the vector corresponding to the compared power technology project document; |x is the modulus of vector x; |y| is the modulus of vector y; The edit distance similarity SimLevenshtein(Tx,Ty) is calculated using the following formula: Where Levenshtein(Tx,Ty) is the edit distance between document Tx and document Ty; Len is the length of the document; The longest common subsequence similarity SimLCS(Tx,Ty) is calculated using the following formula: Where LCS(Tx,Ty) is the longest common subsequence of document Tx and document Ty; The mixed similarity Sim(Tx,Ty) is calculated using the following formula: Sim(Tx,Ty)=a·cos(Tx,Ty)+b·SimLevenshtein(Tx,Ty)+c·SimLCS(Tx,Ty), where a is the first weight value; b is the second weight value; and c is the third weight value.

8. The method for checking duplicate documents of electric power technology projects based on knowledge graph according to claim 7 is characterized in that In the calculation formula of mixed similarity Sim(Tx,Ty), the custom function Simother(Tx,Ty) and the corresponding weight d are introduced to realize the custom processing of mixed similarity; Step S6, according to the document similarity obtained in step S5, completes the duplicate checking of the target electric power technology project document, which specifically includes the following steps: According to the document similarity obtained in step S5, the Myers difference method is used to calculate the similar parts and different parts in the target electric power technology project document and the compared electric power technology project document, and mark them, thereby completing the duplicate check of the target electric power technology project document.

9. A system for implementing the method for checking duplicate documents of electric power technology projects based on knowledge graphs as described in any one of claims 1 to 8, characterized in that It includes a data acquisition module, a knowledge extraction module, a fusion module, a word segmentation module, a similarity calculation module, a document duplicate checking module and an update module; the data acquisition module, the knowledge extraction module, the fusion module, the word segmentation module, the similarity calculation module, the document duplicate checking module and the update module are connected in series in sequence; the data acquisition module is used to acquire existing power technology project documents, perform data preprocessing, and upload data information to the knowledge extraction module; The knowledge extraction module is used to extract knowledge from the pre-processed power technology project documents based on the received data information, so as to construct a dictionary and knowledge graph of the power technology project documents, and upload the data information to the fusion module; The fusion module is used to fuse the dictionary of the electric power technology project document with the general language word library according to the received data information, obtain the professional word library of the electric power technology project document, and upload the data information to the word segmentation module; The word segmentation module is used to perform word segmentation and vectorization processing on the target electric power technology project document based on the received data information and the obtained professional vocabulary, and upload the data information to the similarity calculation module; the similarity calculation module is used to calculate the document similarity of the target electric power technology project document based on the received data information and the result of the word segmentation and vectorization processing, and upload the data information to the document duplication checking module; the document duplication checking module is used to complete the duplication checking of the target electric power technology project document based on the received data information and the obtained document similarity, and upload the data information to the update module; The update module is used to extract knowledge from the target electric power technology project document based on the received data information, so as to update the dictionary and knowledge graph of the constructed electric power technology project document.

10. A storage medium having a computer program stored thereon; when the computer program is executed by a processor, the method for checking duplicate documents of electric power technology projects based on knowledge graphs as described in any one of claims 1 to 8 is implemented.

Images