Nursing corpus automatic creation method based on knowledge graph

Abstract

The invention discloses a knowledge graph-based nursing corpus automatic creation method, and relates to the field of artificial intelligence technology and medical information processing, and the method comprises a multi-dimensional data acquisition stage, a knowledge graph construction stage, a corpus automatic generation and updating stage, and a corpus quality control and evaluation stage. According to the automatic corpus generation method driven by the knowledge graph, dynamic generation and real-time updating of corpora are achieved by constructing an ontology structure, a corpus generation template and an updating algorithm in the nursing field; through a deep learning model and a transfer learning technology, new nursing knowledge and a diversified information structure are continuously introduced into the corpus, and the diversity of corpus content is increased.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence technology and medical information processing, and in particular to a method for automatically creating a nursing corpus based on a knowledge graph. Background Art

[0002] With the rapid development of medical informatization, the demand for high-quality, structured nursing corpora in the nursing field has increased significantly. Developing a method that can automatically construct and continuously update nursing corpora has become a key research topic.

[0003] In terms of nursing knowledge graph construction, foreign countries started earlier and have a relatively mature ontology framework. The United States and Japan have built nursing practice ontologies based on the International Classification of Nursing Practice (ICNP), which provides strong support for the in-depth application of nursing information systems and nursing data mining. In contrast, my country's nursing knowledge graph is still in its early stages of development, and it is urgent to further explore how to effectively integrate nursing practice needs with knowledge graph technology to form a domain knowledge graph that adapts to diverse nursing scenarios.

[0004] Corpus, as a collection of natural language texts containing certain language features, is an important tool for linguistics and natural language processing research. Corpus research originated in the United Kingdom and the United States in the 1960s. Initially, it focused on a single language (such as English) and has gone through the following four stages of development: the first stage (1960s) was a small corpus, which was mainly used for basic linguistic research; the second stage (1980s) was a large-scale corpus, with the data volume expanded to millions of times the previous level; the third stage (late 20th century) was the rise of dynamic corpora, which focused on content updates and opened historical data; the fourth stage was an open and dynamic corpus based on the Internet, which gradually became a more extensive language resource. At present, foreign countries pay more attention to the application of corpora in nursing education and training, and use corpora to improve the professional ability and practical level of nursing staff. In China, corpus research has been carried out since the 1990s, mainly from a linguistic perspective, and its application in the field of nursing is still in its infancy. The future development of nursing corpora can help nursing staff make more accurate judgments and decisions in clinical practice by combining artificial intelligence technology to build an intelligent decision support system.

[0005] In summary, it can be seen that the demand for high-quality, structured nursing corpora in the nursing field has increased significantly. Although foreign countries have accumulated relatively mature experience in nursing knowledge graphs and corpus construction, my country's nursing field still lacks automated tools that can efficiently integrate nursing practice needs and knowledge systems. Therefore, it is particularly necessary to develop an automated nursing corpus construction method based on knowledge graphs. Summary of the invention

[0006] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a method for automatically creating a nursing corpus based on a knowledge graph, so as to realize the automatic construction of the nursing corpus and the continuous updating of nursing data resources.

[0007] The objective of the present invention is achieved through the following technical solutions: A method for automatically creating a nursing corpus based on a knowledge graph, comprising: Multi-dimensional data collection stage: crawl data from professional literature databases, electronic medical record systems, and textbooks, perform data preprocessing operations, and perform intelligent analysis and entity recognition on the preprocessed data; Knowledge graph construction stage: construct the nursing domain ontology, including extracting and defining the core concepts, relationships and attributes of entities, mapping the extracted entities, relationships and attributes to the constructed ontology through knowledge representation learning, and constructing and storing the knowledge graph based on entity and relationship vectors; Corpus automatic generation and update stage: Corpus automatic generation, including: designing corpus generation templates and decomposing them into key elements, mapping key elements with entities and relationships in the knowledge graph, generating new nursing cases and simulation scenarios based on entities and relationships, and filling structured data based on entity and attribute information; Corpus automatic update, identifying and updating low-information or outdated information in the corpus based on the information entropy of the corpus; Corpus quality control and evaluation stage: Establish a hierarchical expert review system and quality assessment model, conduct evaluation and re-evaluation at key nodes of graph generation in combination with the latest research; use active learning strategies and transfer technology to achieve automatic correction of the corpus.

[0008] Furthermore, the multi-dimensional data collection stage specifically includes: It adopts the form of a multi-node collaborative crawler system, automatically crawls the abstracts, texts, and conclusions in professional literature databases through API interfaces or page parsing; automatically extracts nursing records, diagnosis information, and treatment plans from medical records from hospital information systems and patient electronic medical record systems by accessing authorized API interfaces, and desensitizes patients' personal identity information; uses scanning and optical character recognition technology to convert scanned images of paper textbooks into editable text, and obtains digital textbook content from data providers through API interfaces authorized by the platform; Perform data preprocessing on the collected text, including: data format conversion, data cleaning, data screening and merging, and data denoising; Use named entity recognition algorithm to perform intelligent analysis and entity recognition on the preprocessed text, including nursing operation terminology entity recognition and disease or symptom name entity recognition; All data are standardized into a unified format, the data content is labeled using a keyword extraction algorithm, and the pre-processed data is classified and stored in the database according to the label and data type.

[0009] Furthermore, the data preprocessing of the collected text includes: data format conversion, data cleaning, data screening and merging, and data denoising, specifically including: Convert data from different data sources into a standard structured or semi-structured format; Remove invalid or incomplete medical record data information through data cleaning, including useless characters, blank content and incorrectly formatted text; Multiple deduplication algorithms are used to screen and merge duplicate documents, medical records or other redundant information; Perform data denoising, that is, remove meaningless punctuation, advertising information or other useless content in the text.

[0010] Furthermore, the knowledge graph construction stage specifically includes: Constructing nursing knowledge graph ontology: defining core concepts in the nursing field and using them as basic nodes in the knowledge graph, setting specific attributes and their relationships based on the core concepts to form a logical chain; using ontology language to structure the core concepts and relationships; Construct the relationship between nursing knowledge graphs and ontologies: Use deep learning models to learn knowledge representation, model simple and multi-relationships between entities, and convert discrete symbols into continuous vector representations; Generate and store knowledge graphs: store the trained entity and relationship vectors into the graph database, define the query paths for entities and relationships, set indexes for common queries, and generate a nursing knowledge graph; use the built-in visualization tools in the graph database to generate a visual display of the nursing knowledge graph.

[0011] Furthermore, the relationship between the nursing knowledge graph ontology is constructed by using a deep learning model to learn knowledge representation, modeling simple relationships and multiple relationships between entities, and converting discrete symbols into continuous vector representations, specifically including: Extract triples from the knowledge graph as training samples, train the TransE model and adjust its parameters, map the entities and relations in the triples into low-dimensional vectors to model simple relations between entities, and match the relational structure by translating the entity vectors in the low-dimensional vector space; The DistMult model is used to model multiple relationships through a bilinear model to achieve multi-dimensional nursing relationship expression. At the same time, the vectorized relationships are optimized to make the entity vectors with similar relationships closer. The graph convolutional network is used to capture the relationship between neighboring nodes and optimize the node representation. The graph attention network is used to assign different attention weights to each node in the graph, so that the core nodes receive higher attention.

[0012] Furthermore, the automatic generation of the corpus specifically includes: By analyzing common cases and teaching material structures in the nursing field, the corpus template is decomposed into key elements, and connected with the entity and relationship mapping in the knowledge graph to generate template formats for different scenarios; the corpus template includes a symptom description template and a complication treatment template; Generate new nursing cases and simulation scenarios through entities and relationships in the graph using generative adversarial networks and variational autoencoders respectively; Based on the corpus generation template, the corresponding entity and attribute information are extracted from the knowledge graph through semantic analysis to automatically fill in the structured data.

[0013] Furthermore, the generative adversarial network and variational autoencoder are used to generate new nursing cases and simulation scenarios through entities and relationships in the graph, specifically including: Based on the nursing knowledge graph, the generative adversarial network (GAN) is trained to generate nursing descriptions, and then nursing cases are generated based on the nursing descriptions. Use variational autoencoder VAE to generate variant sentences, and use the variant sentences to generate new simulation scenarios.

[0014] Furthermore, the automatic updating of the corpus specifically includes: The TF-IDF algorithm is used to calculate the information content of each corpus, and the corpus with lower information content is marked as replaceable items; Use graph convolutional networks and graph attention network models to automatically analyze and optimize the knowledge graph structure, adjust outdated or newly added nursing knowledge, regularly update the knowledge graph, and regenerate nursing corpus for replaced entries; Use text deduplication algorithms to clean up duplicate entries and ensure that the generated corpus is free of duplication.

[0015] Furthermore, the establishment of a hierarchical expert review system and quality assessment model, at the key nodes of the graph generation, conducts evaluation and re-evaluation in combination with the latest research, specifically including: Establish a hierarchical expert review system, select a number of authoritative experts with 10 years or more of clinical experience and academic background to form a review team, and set up a variety of review modes, including node review, batch review and random sampling; the system automatically assigns key nodes in the process of establishing and updating the map to multiple relevant experts for independent review and evaluation. The review content specifically includes in-depth inspection of node content, definition, and logical relationship; the system automatically invites experts to re-evaluate regularly based on new research, new technologies and new guidelines in the nursing field; A quality assessment model is established to conduct regular multi-dimensional assessments of the corpus, including coverage checks on core nursing knowledge points, problem diagnosis in low-precision areas, and generation of quality assessment reports. The main indicators of the quality assessment model include precision, recall, and F1 score.

[0016] Furthermore, the active learning strategy and transfer technology are used to realize automatic correction of the corpus, specifically including: An active learning algorithm model is introduced, and a sample selection mechanism is adopted to give priority to data with high information gain or large uncertainty for manual labeling, thereby training the quality assessment model and gradually optimizing the quality assessment model; By applying transfer learning technology, introducing knowledge similarity detection and category-based differentiated transfer strategy, the verified knowledge in the existing corpus is transferred to the new corpus verification process, the corpus verification is completed, and the corpus is automatically corrected according to the verification results.

[0017] The beneficial effects of the present invention are: (1) Improve build efficiency and update speed With the help of the automated creation process, the corpus can quickly collect, organize and generate the knowledge content required in the nursing field, greatly shortening the time to build the corpus. At the same time, the automated update mechanism enables the corpus to integrate the latest research and practice results of the nursing industry in real time, ensuring that the corpus always keeps pace with the times. Compared with the traditional manual update method, it significantly improves work efficiency and facilitates the corpus to adapt to new needs and changes in knowledge in a short time.

[0018] (2) Reduce construction and maintenance costs Through the automated corpus generation and update algorithm, the present invention reduces the workload of manual annotation and review, and reduces the investment cost of human resources. Using machine learning and deep learning technology, the system can automatically identify, classify and update content, greatly reducing the need for human review, thereby reducing the long-term maintenance cost of the corpus and making it more economical.

[0019] (3) Double quality assurance The quality of the corpus is doubly guaranteed by the expert review system and the quality assessment model. The expert review system ensures the accuracy and authority of the corpus content by manually verifying the key nodes and relationships in the graph; the quality assessment model regularly tests the comprehensiveness and timeliness of the corpus through comprehensive indicators such as precision, recall rate, and F1 score, and promptly discovers and corrects potential problems. This dual guarantee mechanism significantly improves the reliability of the corpus and its applicability in nursing teaching, research, and practice.

[0020] (4) Improving the diversity and innovation of the corpus Through deep learning models and transfer learning technology, the present invention continuously introduces new nursing knowledge and diversified information structures into the corpus, increasing the diversity of the corpus content. The application of transfer learning technology enables verified knowledge to quickly adapt and expand to new nursing situations, thereby reducing duplication of work while improving the innovation of content and meeting changing practical needs.

[0021] (5) Enhance practicality and application value Based on automated creation, updating and quality assurance, the corpus constructed by the present invention has higher practicality. The corpus can provide timely and accurate knowledge support for nursing practitioners and researchers, and effectively assist nursing education and training. With the help of transfer learning and deep learning models, the content of the corpus can be flexibly applied to different nursing scenarios and cases, improving its application and conversion value in nursing teaching and clinical practice. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A flowchart of a method for automatically creating a nursing corpus based on a knowledge graph provided by the present invention. DETAILED DESCRIPTION

[0023] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0024] See also Figure 1 , the present invention provides a technical solution: A method for automatically creating a nursing corpus based on a knowledge graph, comprising: S1, multi-dimensional data collection stage: capture data from professional literature databases, electronic medical record systems, and textbooks, perform data preprocessing operations, and perform intelligent analysis and entity recognition on the preprocessed data.

[0025] In a specific embodiment, the multi-dimensional data collection stage specifically includes: S11, using a multi-node collaborative crawler system, automatically crawling abstracts, texts, and conclusions from professional literature databases through API interfaces or page parsing; professional literature databases include but are not limited to PubMed, Wanfang, Weipu, Dingxiangyuan, etc.; By accessing the authorized API interface, the nursing records, diagnosis information, and treatment plans in the medical records are automatically extracted from the hospital information system (HIS) and the patient electronic medical record system (EMR), and the patient's personal identity information is desensitized; Scanning and optical character recognition technology are used to convert the scanned images of paper textbooks into editable text, and the digital textbook content is obtained from the data provider through the API interface through platform authorization. For example, standardized nursing operation procedures and nursing methods can be obtained from relevant nursing professional books, operation manuals, guidelines, etc.

[0026] During the multi-dimensional data collection phase, the structure of data from different sources is analyzed and divided into structured data (such as database records) and unstructured data (such as PDF documents), and appropriate data extraction methods are determined for data with different structures.

[0027] S12, preprocessing the collected text, including: data format conversion, data cleaning, data screening and merging, and data denoising, specifically including: S121, convert data from different data sources into a structured or semi-structured standard format, such as JSON, XML, CSV, etc.; S122, remove invalid or incomplete medical record data information through data cleaning, including useless characters, blank content and incorrectly formatted text; S123, for duplicate documents, medical records or other redundant information, multiple deduplication algorithms are used to screen and merge; S124, performing data denoising processing, that is, removing meaningless punctuation marks, advertising information or other useless content in the text.

[0028] Data preprocessing can improve data availability, ensure that the information in the corpus is unique and non-repetitive, and ensure the purity of the data, laying the foundation for subsequent corpus construction and knowledge graph generation.

[0029] S13, uses the named entity recognition algorithm (NER) to perform intelligent analysis and entity recognition on the preprocessed text, including nursing operation terminology entity recognition and disease or symptom name entity recognition.

[0030] The entity recognition process mainly includes the recognition of nursing operation entities. The system automatically detects and extracts specific nursing operation terms such as "intravenous infusion", "gastric lavage", "indwelling catheterization", "ECG monitoring", etc.; as well as the recognition of disease name entities. By identifying the names of diseases or symptoms involved in the text, such as "pneumonia", "cirrhosis", "hypertension", etc., it ensures that important medical information is accurately extracted.

[0031] S14, standardizes all data into a unified format, uses keyword extraction algorithms to label the data content, and classifies and stores the preprocessed data into a database according to the label and data type.

[0032] For example, standardize the time format and unit of all data into a unified format, use keyword extraction algorithms to label the data content, such as categorizing and storing by labels such as "symptoms", "nursing measures", and "treatment methods" to ensure the efficiency of subsequent processing. Then, according to different data types, store the data in corresponding relational databases (such as MySQL, for structured data) or document databases (such as MongoDB, for unstructured data), and set up regular data backup to ensure data security and support database backup and quick recovery functions.

[0033] S2, knowledge graph construction phase: construct the nursing domain ontology, including extracting and defining the core concepts, relationships and attributes of entities, mapping the extracted entities, relationships and attributes to the constructed ontology through knowledge representation learning, and constructing and storing the knowledge graph based on entity and relationship vectors.

[0034] In a specific embodiment, the knowledge graph construction stage specifically includes: S21, construct the nursing knowledge graph ontology: define the core concepts in the nursing field and use them as the basic nodes in the knowledge graph. Set specific attributes and their relationships based on the core concepts to form a logical chain; use ontology language to structure the core concepts and relationships to ensure the accuracy of the association between data and queries.

[0035] In one example, it specifically includes: (1) Through literature analysis and multiple rounds of expert consultation, the core concepts and secondary concepts in the field of nursing were listed. The core concepts include "disease", "symptoms", "nursing measures", "effects", etc. The secondary concepts of "nursing measures" may include "monitoring", "rehabilitation", "prevention", etc.; (2) Assign attributes and types to each concept. For example, the attributes of “nursing measures” include “implementation time” and “applicable symptoms”. (3) Define the relationship between core concepts and form a logical chain so that the logical associations of different elements in the nursing process can be intuitively displayed, for example, the "disease-symptom" association, such as hypertension-headache, and the "nursing measure-effect" association, such as drug treatment-blood pressure reduction; (4) Use ontology languages ​​(such as OWL, Web Ontology Language) to structure the description of concepts and relationships to ensure the consistency and logic of concept expression, so that the nursing knowledge graph can be standardized and stored in the computer system, providing a basis for further processing.

[0036] S22, constructing the relationship between nursing knowledge graph ontologies: using deep learning models to learn knowledge representation, modeling simple and multi-relationships between entities, and converting discrete symbols into continuous vector representations; specifically, including: S221, extract triples (such as <fever, applicable, cooling care>) from the knowledge graph as training samples, train the TransE model and adjust parameters, map the entities and relationships in the triples into low-dimensional vectors to model simple relationships between entities, and match the relationship structure by translating the entity vectors in the low-dimensional vector space; among them, adjusting model parameters, such as learning rate, vector dimension, etc., can ensure that the trained vectors can effectively represent the semantic associations between concepts.

[0037] S222, the DistMult model is used to perform multi-relationship modeling through a bilinear model to express multi-dimensional nursing relationships (such as "nursing implementation-effect-nursing evaluation"), and at the same time optimize the vectorized relationships to make entity vectors with similar relationships closer, thereby improving the query effect of the knowledge graph.

[0038] S223, through the graph convolution network to capture the relationship between neighboring nodes and optimize the node representation, use the graph attention network to assign different attention weights to each node in the graph, so that the core nodes receive higher attention, increase the priority of the graph in the retrieval and reasoning process, and achieve accurate modeling.

[0039] S23, generate and store knowledge graph: store the trained entity and relationship vectors into a graph database (such as Neo4j), define the query path of entities and relationships, set indexes for common queries, and generate a nursing knowledge graph; use the built-in visualization tool of the graph database to generate a visual display of the nursing knowledge graph. Index settings can improve the query speed of the graph, and the visual display provides support for subsequent inspection and verification of the graph structure.

[0040] S3, corpus automatic generation and update stage: corpus automatic generation, including: designing corpus generation templates and decomposing them into key elements, mapping key elements with entities and relationships in the knowledge graph, generating new nursing cases and simulation scenarios based on entities and relationships, and filling structured data based on entity and attribute information; corpus automatic update, identifying and updating low-information or outdated information in the corpus based on the information entropy of the corpus; Among them, the entities in the knowledge graph include symptoms, nursing intervention plans, treatment methods, etc., and the relationships between entities include causal relationships, dependencies, etc. The generated corpus will include a variety of case scenarios, increase the diversity of cases, and provide rich materials for the application of the corpus.

[0041] In a specific embodiment, the automatic generation of the corpus specifically includes: S311, by analyzing common cases and teaching material structures in the nursing field, decomposing the corpus template into key elements, and connecting with the entity and relationship mapping in the knowledge graph to generate template formats for different scenarios; the corpus template includes a symptom description template and a complication handling template.

[0042] For example: in the symptom description template, "The patient has high fever symptoms and needs to take nursing measures such as ice packs to cool down"; in the complication management template, "Diabetic patients need to be alert to the occurrence of diabetic ketoacidosis, and the recommended nursing measure is timely rehydration."

[0043] S312, using generative adversarial networks and variational autoencoders, respectively, to generate new nursing cases and simulation scenarios through entities and relationships in the graph; specifically including: Based on the nursing knowledge graph, the generative adversarial network (GAN) is trained to generate nursing descriptions, and nursing cases are generated based on the nursing descriptions. The generated description text is such as "the patient has symptoms of Staphylococcus aureus infection, and daily debridement and dressing change are recommended."

[0044] Use variational autoencoder VAE to generate variant sentences, and use the variant sentences to generate new simulation scenarios.

[0045] S313, based on the corpus generation template, extracts corresponding entity and attribute information from the knowledge graph through semantic analysis to automatically fill in structured data.

[0046] The corpus generation template design creates a framework that can generate new corpora that meet the needs of specific fields based on the entities and relationships in the knowledge graph, which can improve the diversity and coverage of the data.

[0047] In a specific embodiment, the automatic updating of the corpus specifically includes: S321, using the TF-IDF algorithm to calculate the information content of each corpus, and marking the corpus with lower information content as replaceable items; S322, using graph convolutional network and graph attention network models to automatically analyze and optimize the knowledge graph structure, adjust outdated or newly added nursing knowledge, regularly update the knowledge graph, and regenerate nursing corpus for replaced entries; S323, using a text deduplication algorithm to clean up duplicate entries to ensure that the generated corpus has no duplication.

[0048] In order to identify and update low-information or outdated information in the corpus, the present invention designs an analysis algorithm based on information entropy. By evaluating the information entropy of each node in the graph, nodes with low information content or sparse data are identified, and corpus updates are applied to these nodes first to enrich and optimize the graph content. The automatic update and timeliness maintenance of the corpus ensure that the content of the knowledge graph is accurate and up-to-date, and can promptly reflect information changes in the field.

[0049] S4, corpus quality control and evaluation stage: Establish a hierarchical expert review system and quality assessment model, conduct evaluation and re-evaluation at key nodes of graph generation in combination with the latest research; use active learning strategies and transfer technology to achieve automatic correction of the corpus.

[0050] Furthermore, the establishment of a hierarchical expert review system and quality assessment model, at the key nodes of the graph generation, conducts evaluation and re-evaluation in combination with the latest research, specifically including: Establish a hierarchical expert review system, select a number of authoritative experts with 10 years or more of clinical experience and academic background to form a review team, and set up a variety of review modes, including node review, batch review and random sampling; the system automatically assigns key nodes in the process of establishing and updating the map to multiple relevant experts for independent review and evaluation. The review content specifically includes in-depth inspection of node content, definition, and logical relationship; the system automatically invites experts to re-evaluate regularly based on new research, new technologies and new guidelines in the nursing field; A quality assessment model is established to conduct regular multi-dimensional assessments of the corpus, including coverage checks on core nursing knowledge points, problem diagnosis in low-precision areas, and generation of quality assessment reports. The main indicators of the quality assessment model include precision, recall, and F1 score.

[0051] Conduct coverage analysis on common nursing scenarios, symptoms, diagnoses, and treatment plans, detect areas that are not included or have insufficient information, and evaluate the model through coverage checks to ensure that core nursing knowledge points are fully covered and optimize the comprehensiveness of the corpus. The model diagnoses problems in low-precision areas and identifies possible causes of errors, such as missing data and classification errors, to help the development team make accurate corrections. The quality assessment report generated through regular evaluation cycles will cover the changing trends of various indicators in the corpus and will be visualized through charts to facilitate intuitive understanding by the development team and related personnel. Through regular testing and feedback mechanisms, the R&D team can promptly identify potential problems and make adjustments to ensure the stability and scientific nature of the corpus content.

[0052] Furthermore, the active learning strategy and transfer technology are used to realize automatic correction of the corpus, specifically including: An active learning algorithm model is introduced, and a sample selection mechanism is adopted to give priority to data with high information gain or large uncertainty for manual labeling, and then the quality assessment model is trained to gradually optimize the quality assessment model; through active learning, the dependence on a large number of labeled samples is reduced, and the model quality is improved with a small amount of manual verification, which reduces manual workload and costs, while ensuring the accuracy and authority of the corpus, and gradually improving its operability and reliability in practical applications.

[0053] By applying transfer learning technology, introducing knowledge similarity detection and category-based differentiated transfer strategies, the verified knowledge in the existing corpus is transferred to the new corpus verification process to complete the corpus verification, and the corpus is automatically corrected according to the verification results. Transfer learning can effectively improve the verification efficiency, especially in the field of new knowledge, it can significantly reduce the time cost of initial manual verification. Through this technology, the verification process of the corpus can achieve efficient updates while ensuring accuracy, which helps to quickly adapt to new nursing practice needs.

[0054] The differences between the present invention and the prior art are mainly reflected in the following aspects: 1. Automatic creation and update mechanism: The present invention adopts a knowledge graph-driven corpus automatic generation method, which realizes dynamic generation and real-time update of corpus by constructing the ontology structure, corpus generation template and update algorithm in the nursing field. In contrast, the existing technology usually relies on manual or semi-automatic methods to create and update corpora, which is inefficient and difficult to quickly adapt to changes in domain knowledge.

[0055] 2. Structured corpus generation based on knowledge graph: This invention can generate diversified corpus that meets the needs of the nursing field by using structured knowledge graph and corpus generation template, combined with generation models such as GAN or VAE. However, the existing technology is mostly unstructured data, lacks systematic template design and knowledge graph support, and the content of the generated corpus is easily affected by human factors and has low consistency.

[0056] 3. Intelligent corpus update and system maintenance: The present invention uses information entropy or TF-IDF algorithm to identify content with low information density or outdated content, and regularly updates the corpus to ensure the timeliness of the content. This intelligent update mechanism improves the applicability and timeliness of the corpus, so that the generated content can always meet the latest nursing guidelines and standards. However, the existing technology lacks automated timeliness maintenance means, the update frequency is not high, and the content is easy to become outdated.

[0057] 4. Domain-specific knowledge representation and relationship modeling: Through the ontology construction of the knowledge graph and the relationship modeling based on algorithms such as TransE or DistMult, the present invention can present the core concepts, attributes and their associations in the nursing field in a structured manner, and realize the accurate association and hierarchy of the corpus content. However, the corpus of the prior art usually does not perform specific modeling and structuring of domain knowledge, making it difficult to reflect the professional concept association in the nursing field, and the accuracy and logic of the corpus usage are relatively insufficient.

[0058] 5. Scalability and applicability: Since the present invention uses knowledge graph as a structured representation tool, the corpus has strong scalability and adaptability, and is suitable for different nursing scenarios and user needs. However, the existing technology lacks deep adaptation to nursing scenarios, and its scalability and application scenarios are limited.

[0059] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art shall not deviate from the spirit and scope of the present invention, and shall be within the scope of protection of the claims attached to the present invention.

Claims

1. A method for automatically creating a nursing corpus based on a knowledge graph, characterized in that: include: Multi-dimensional data collection stage: crawl data from professional literature databases, electronic medical record systems, and textbooks, perform data preprocessing operations, and perform intelligent analysis and entity recognition on the preprocessed data; Knowledge graph construction stage: construct the nursing domain ontology, including extracting and defining the core concepts, relationships and attributes of entities, mapping the extracted entities, relationships and attributes to the constructed ontology through knowledge representation learning, and constructing and storing the knowledge graph based on entity and relationship vectors; Corpus automatic generation and update stage: Corpus automatic generation, including: designing corpus generation templates and decomposing them into key elements, mapping key elements with entities and relationships in the knowledge graph, generating new nursing cases and simulation scenarios based on entities and relationships, and filling structured data based on entity and attribute information; Corpus automatic update, identifying and updating low-information or outdated information in the corpus based on the information entropy of the corpus; Corpus quality control and evaluation stage: Establish a hierarchical expert review system and quality assessment model, conduct evaluation and re-evaluation at key nodes of graph generation in combination with the latest research; use active learning strategies and transfer technology to achieve automatic correction of the corpus.

2. The method for automatically creating a nursing corpus based on a knowledge graph according to claim 1, characterized in that: The multi-dimensional data collection stage specifically includes: It adopts the form of a multi-node collaborative crawler system, automatically crawls the abstract, text, and conclusion information in the professional literature database through the API interface or based on page parsing; automatically extracts the nursing records, diagnosis information, and treatment plans in the medical records from the hospital information system and the patient electronic medical record system by accessing the authorized API interface, and desensitizes the patient's personal identity information; uses scanning and optical character recognition technology to convert the scanned images of paper textbooks into editable text, and obtains the content of digital textbooks from the data provider through the API interface authorized by the platform; Perform data preprocessing on the collected text, including: data format conversion, data cleaning, data screening and merging, and data denoising; Use named entity recognition algorithm to perform intelligent analysis and entity recognition on the preprocessed text, including nursing operation terminology entity recognition and disease or symptom name entity recognition; All data are standardized into a unified format, the data content is labeled using a keyword extraction algorithm, and the pre-processed data is classified and stored in the database according to the label and data type.

3. The method for automatically creating a nursing corpus based on a knowledge graph according to claim 2, characterized in that: The data preprocessing of the collected text includes: data format conversion, data cleaning, data screening and merging, and data denoising, specifically including: Convert data from different data sources into a standard structured or semi-structured format; Remove invalid or incomplete medical record data information through data cleaning, including useless characters, blank content and incorrectly formatted text; Multiple deduplication algorithms are used to screen and merge duplicate documents, medical records or other redundant information; Perform data denoising, that is, remove meaningless punctuation, advertising information or other useless content in the text.

4. The method for automatically creating a nursing corpus based on a knowledge graph according to claim 1, characterized in that: The knowledge graph construction stage specifically includes: Constructing nursing knowledge graph ontology: defining core concepts in the nursing field and using them as basic nodes in the knowledge graph, setting specific attributes and their relationships based on the core concepts to form a logical chain; using ontology language to structure the core concepts and relationships; Construct the relationship between nursing knowledge graphs and ontologies: Use deep learning models to learn knowledge representation, model simple and multi-relationships between entities, and convert discrete symbols into continuous vector representations; Generate and store knowledge graphs: store the trained entity and relationship vectors into the graph database, define the query paths for entities and relationships, set indexes for common queries, and generate a nursing knowledge graph; use the built-in visualization tools in the graph database to generate a visual display of the nursing knowledge graph.

5. The method for automatically creating a nursing corpus based on a knowledge graph according to claim 4 is characterized in that: The relationship between the nursing knowledge graph and the ontology is constructed by using a deep learning model to learn knowledge representation, modeling simple relationships and multiple relationships between entities, and converting discrete symbols into continuous vector representations, specifically including: Extract triples from the knowledge graph as training samples, train the TransE model and adjust its parameters, map the entities and relations in the triples into low-dimensional vectors to model simple relations between entities, and match the relational structure by translating the entity vectors in the low-dimensional vector space; The DistMult model is used to model multiple relationships through a bilinear model to achieve multi-dimensional nursing relationship expression. At the same time, the vectorized relationships are optimized to make the entity vectors with similar relationships closer. The graph convolutional network is used to capture the relationship between neighboring nodes and optimize the node representation. The graph attention network is used to assign different attention weights to each node in the graph, so that the core nodes receive higher attention.

6. The method for automatically creating a nursing corpus based on a knowledge graph according to claim 1, characterized in that: The automatic generation of the corpus specifically includes: By analyzing common cases and teaching material structures in the nursing field, the corpus template is decomposed into key elements, and connected with the entity and relationship mapping in the knowledge graph to generate template formats for different scenarios; the corpus template includes a symptom description template and a complication treatment template; Generate new nursing cases and simulation scenarios through entities and relationships in the graph using generative adversarial networks and variational autoencoders respectively; Based on the corpus generation template, the corresponding entity and attribute information are extracted from the knowledge graph through semantic analysis to automatically fill in the structured data.

7. The method for automatically creating a nursing corpus based on a knowledge graph according to claim 6, characterized in that: The method uses generative adversarial networks and variational autoencoders to generate new nursing cases and simulation scenarios through entities and relationships in the graph, including: Based on the nursing knowledge graph, the generative adversarial network (GAN) is trained to generate nursing descriptions, and then nursing cases are generated based on the nursing descriptions. Use variational autoencoder VAE to generate variant sentences, and use the variant sentences to generate new simulation scenarios.

8. The method for automatically creating a nursing corpus based on a knowledge graph according to claim 1, characterized in that: The automatic updating of the corpus specifically includes: The TF-IDF algorithm is used to calculate the information content of each corpus, and the corpus with lower information content is marked as replaceable items; Use graph convolutional network and graph attention network models to automatically analyze and optimize the knowledge graph structure, adjust outdated or newly added nursing knowledge, regularly update the knowledge graph, and regenerate nursing corpus for replaced entries; Use text deduplication algorithms to clean up duplicate entries and ensure that the generated corpus is free of duplication.

9. The method for automatically creating a nursing corpus based on a knowledge graph according to claim 1, characterized in that: The above-mentioned establishment of a hierarchical expert review system and quality assessment model will evaluate and re-evaluate the key nodes of the graph generation in combination with the latest research, including: Establish a hierarchical expert review system, select a number of authoritative experts with 10 years or more of clinical experience and academic background to form a review team, and set up a variety of review modes, including node review, batch review and random sampling; the system automatically assigns key nodes in the process of establishing and updating the map to multiple relevant experts for independent review and evaluation. The review content specifically includes in-depth inspection of node content, definition, and logical relationship; the system automatically invites experts to re-evaluate regularly based on new research, new technologies and new guidelines in the nursing field; A quality assessment model is established to conduct regular multi-dimensional assessments of the corpus, including coverage checks on core nursing knowledge points, problem diagnosis in low-precision areas, and generation of quality assessment reports. The main indicators of the quality assessment model include precision, recall, and F1 score.

10. The method for automatically creating a nursing corpus based on a knowledge graph according to claim 9, characterized in that: The active learning strategy and transfer technology are used to realize automatic correction of the corpus, specifically including: An active learning algorithm model is introduced, and a sample selection mechanism is adopted to give priority to data with high information gain or large uncertainty for manual labeling, thereby training the quality assessment model and gradually optimizing the quality assessment model; By applying transfer learning technology, introducing knowledge similarity detection and category-based differentiated transfer strategy, the verified knowledge in the existing corpus is transferred to the new corpus verification process, the corpus verification is completed, and the corpus is automatically corrected according to the verification results.

Images