Knowledge graph-driven network academic resource association selection optimization method

Abstract

The invention discloses a network academic resource association selection optimization method driven by a knowledge graph, and the method comprises the steps: defining an ontology through a top-down method, and constructing a network academic resource association selection ontology through a seven-step method; preprocessing the network academic resources; identifying entities from the preprocessed network academic resources and determining a relationship between the entities to form a structured knowledge representation; integrating the identified entities and relationships into a knowledge graph, and storing the knowledge graph into a graph database Neo4j; carrying out resource association query and deep analysis by utilizing the constructed knowledge graph, wherein the analysis comprises statistical analysis, path analysis, key object analysis and clustering analysis; and according to the user feedback and the progress of the academic field, optimizing the knowledge graph and the association selection algorithm. According to the method, efficient knowledge graph construction and management, accurate resource association and recommendation, dynamic optimization and continuous improvement can be realized, and scientific research efficiency and knowledge innovation speed are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of network academic resource management, and in particular to a knowledge graph-driven network academic resource association selection optimization method. Background Art

[0002] With the rapid development of Internet technology, online academic resources have exploded. Massive amounts of academic information continue to emerge on the Internet, covering various disciplines and research directions, providing researchers with a rich source of information. However, this rapid growth has also brought about the problem of information overload. It is difficult for users to quickly and accurately find the most valuable resources among the massive amount of academic resources, which seriously affects the efficiency of scientific research and the speed of knowledge innovation.

[0003] Traditional methods for selecting online academic resources mainly rely on keyword retrieval. Users enter keywords and the search engine returns a list of resources related to these keywords. Although this method can help users find the information they need to a certain extent, it has obvious limitations. Keyword retrieval is only based on text matching and lacks an understanding of the deep semantic relationship of online academic resources, making it difficult to achieve accurate resource recommendations and in-depth knowledge exploration.

[0004] In modern scientific research, accurate resource recommendations and in-depth knowledge exploration are essential to improving scientific research efficiency and promoting knowledge innovation. Researchers need to quickly find high-quality academic resources that are highly relevant to their research direction in order to obtain the latest research results, potential cooperation opportunities, and cutting-edge technologies in a short period of time. However, traditional keyword search methods cannot meet this demand. It cannot identify the complex relationships between resources, which makes users spend a lot of time screening and evaluating resources in massive amounts of information, seriously affecting the efficiency of scientific research.

[0005] Therefore, the present invention provides a knowledge graph-driven online academic resource association selection optimization method, which can achieve accurate resource recommendation and in-depth knowledge exploration, and improve the discovery and utilization efficiency of online academic resources. Summary of the invention

[0006] The purpose of the present invention is to solve the problems of insufficient recommendation accuracy, insufficient knowledge exploration depth and difficulty in rapid screening in the prior art, and to propose a knowledge graph-driven network academic resource association selection optimization method.

[0007] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solution: a knowledge graph-driven network academic resource association selection optimization method, comprising the following steps: Step S1, defining the ontology through a top-down approach, and constructing the online academic resource association selection ontology using a seven-step method; Step S2, preprocessing online academic resources, including organization and screening of topics, identification and analysis of topic keywords, selection and evaluation of sample data; Step S3, identifying entities from the preprocessed online academic resources and determining the relationships between entities to form a structured knowledge representation; Step S4, integrating the identified entities and relationships into the knowledge graph and storing them in the graph database Neo4j; Step S5, using the constructed knowledge graph to perform resource association query and in-depth analysis, including statistical analysis, path analysis, key object analysis and cluster analysis; Step S6, optimizing the knowledge graph and associated selection algorithm based on user feedback and progress in the academic field; The seven-step method includes determining the professional field and scope of the ontology, examining the possibility of reusing existing ontologies, listing important terms in the ontology, defining classes and class hierarchies, defining class attributes, defining attribute facets, and creating attributes.

[0008] Furthermore, in step S1, the following sub-steps are also included: S1-1, defines the ontology and data model through a top-down approach using a seven-step method; S1-2, determine the professional field and scope of the ontology, the professional field includes computer science, biology, medicine, physics, chemistry, social science, humanities, and interdisciplinary fields, and the scope includes resource type, data source, time range, geographical range, research method, and application field; S1-3, examine the possibility of reusing existing ontologies, including general ontologies, domain ontologies, task ontologies, and application ontologies. The possibility of reusing existing ontologies includes compatibility, coverage, maintenance status, update frequency, and community support. S1-4, list the conceptual terms in the field, including scientific research subjects, scientific research results, scientific research tools, scientific research exchanges, research fields, research methods, cooperation networks, evaluation indicators, academic identifiers, research trends, metadata, and copyright information; S1-5, define classes and attributes, by creating classes and their hierarchical systems, and defining the attributes of each class, including object attributes and data attributes; S1-6, conduct quality assessment, evaluate the quality of the ontology through the standards of consistency, completeness, reusability, scalability, and accuracy, and update and improve the ontology based on the development of domain knowledge and user feedback.

[0009] Furthermore, in step S2, the following sub-steps are also included: S2-1, using the Scopus database and subject vocabulary, classify journal articles by subject and screen out the subject directions related to the target research field; S2-2, through Elsevier's fingerprint engine technology, extract topic-related keywords and terms from unstructured text, create a weighted term index, and evaluate the importance of terms through the inverse document frequency statistics method. The specific formula is: Where IDF stands for Inverse Document Frequency, t is the term, and N is the total number of documents. is the number of documents containing term t; S2-3, based on the theory of provenance and the theory of information resource value, evaluate the disciplinary representativeness of the sample data, the authority of the original source, the reputation of the authors and their institutions, the citations of the literature, and the timeliness and completeness of the content.

[0010] Furthermore, in step S3, the following sub-steps are also included: S3-1, identifying entities from the data patterns of online academic resources, including domain experts, work units, journal articles, scientific research projects, patents, research reports, academic conferences, conference reports, academic speeches, consulting reports, project completion reports, books, dissertations, data sets, scientific information, scientific blogs, and reviews; S3-2, determine the relationship between entities, including affiliation, writing, hosting, invention, participation, reporting, contribution, and mentioning; S3-3, entities and relationships are integrated into a structured knowledge representation in the form of RDF triples, where the form of the RDF triples is "entity-relationship-entity".

[0011] Furthermore, in step S4, the following sub-steps are also included: S4-1, clean, remove duplicates, and standardize the extracted entities and relationships; S4-2, integrating entities and relationships from different data sources to make the knowledge graph consistent and complete, the integration including ontology mapping and entity alignment; S4-3, using the attribute graph structure of the Neo4j graph database, the nodes and edges in the knowledge graph are stored in the database, wherein the nodes represent entities and the edges represent relationships; S4-4, establish a data update mechanism, regularly monitor the update status of the data source, and update the knowledge base in an incremental manner.

[0012] Furthermore, in step S5, the following sub-steps are also included: S5-1, Statistical analysis: Statistical analysis of the number, degree distribution, and weighted degree distribution of nodes and edges in the knowledge graph to understand the structure and distribution of network academic resources in the field; S5-2, path analysis: exploring the connection paths between different resources, including shortest path analysis and all path exploration, wherein the shortest path analysis is used to determine the shortest path between two nodes, and the all path exploration is used to find all possible paths between two nodes; S5-3, key object analysis: identifying the most influential nodes through centrality analysis, which includes degree centrality analysis, betweenness centrality analysis, closeness centrality analysis and personalized ranking; S5-4, Cluster analysis: Use the k-means clustering method to group similar nodes in the graph and reveal potential research directions.

[0013] Furthermore, in step S6, the following sub-steps are also included: S6-1, user feedback mechanism, collects users’ evaluations of knowledge graphs and related selection results, analyzes feedback content, and evaluates users’ satisfaction with related selection results; S6-2, Academic Progress Monitoring: Tracking research results, methods and trends in the academic field, and updating knowledge and information in the knowledge graph; S6-3, Algorithm optimization: Based on user feedback and the development of the academic field, optimize the relevance selection algorithm, adjust the algorithm's parameters and strategies, and improve the accuracy and efficiency of relevance selection.

[0014] The beneficial effects brought about by the technical solution provided by the present invention include at least: The method of the present invention defines the ontology through a top-down method, and adopts a seven-step method to construct an ontology for the association selection of network academic resources; preprocesses the network academic resources, including the organization and screening of topics, the identification and analysis of topic keywords, and the selection and evaluation of sample data; identifies entities from the preprocessed network academic resources and determines the relationship between entities to form a structured knowledge representation; integrates the identified entities and relationships into a knowledge graph and stores them in a graph database Neo4j; uses the constructed knowledge graph to perform resource association query and in-depth analysis, including statistical analysis, path analysis, key object analysis and cluster analysis; and optimizes the knowledge graph and the association selection algorithm according to user feedback and progress in the academic field.

[0015] By constructing a knowledge graph, the method of the present invention can deeply explore the deep semantic relationship between online academic resources, achieve accurate resource recommendation and in-depth knowledge exploration, provide users with high-quality academic resources that are highly relevant to research needs, and significantly improve the accuracy and efficiency of resource discovery.

[0016] The method of the present invention can perform more refined classification and clustering of online academic resources through the structured representation of knowledge graphs, and can also integrate online academic resources across fields, revealing potential connections between different research fields and promoting interdisciplinary research.

[0017] The method of the present invention supports complex association query and analysis through knowledge graph, which can reveal the deep structure and evolution law of online academic resources, provide deeper insights, and help researchers better understand the dynamic changes of academic networks.

[0018] The method of the present invention can accelerate the dissemination and innovation process of scientific knowledge by optimizing the associated selection of online academic resources. By accurately recommending high-quality academic resources, researchers can obtain the latest research results and cutting-edge technologies more quickly, thereby improving scientific research efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 A flow chart of a method provided by an embodiment of the present invention; Figure 2 A diagram of a method for constructing a knowledge graph of online academic resources provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a knowledge graph-driven network academic resource association selection optimization method proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics in one or more embodiments may be combined in any suitable form.

[0022] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0023] The following examples are for illustrative purposes only and are not intended to limit the scope of the present invention.

[0024] The following is a detailed description of a specific scheme of a knowledge graph-driven network academic resource association selection optimization method provided by the present invention in conjunction with the accompanying drawings. Example

[0025] See also Figure 1 and Figure 2 , which shows a method flow chart of a knowledge graph-driven online academic resource association selection optimization method and a method diagram of online academic resource knowledge graph construction provided by an embodiment of the present invention, the method comprising the following steps: Step S1, defining the ontology through a top-down approach, and constructing the online academic resource association selection ontology using a seven-step method, wherein the seven-step method includes determining the professional field and scope of the ontology, examining the possibility of reusing the existing ontology, listing the important terms in the ontology, defining classes and class hierarchies, defining class attributes, defining attribute facets, and creating attributes; Wherein step S1 also includes the following sub-steps: S1-1, defines the ontology and data model through a top-down approach using a seven-step method; S1-2, determine the professional field and scope of the ontology, the professional field includes computer science, biology, medicine, physics, chemistry, social science, humanities, and interdisciplinary fields, and the scope includes resource type, data source, time range, geographical range, research method, and application field; S1-3, examine the possibility of reusing existing ontologies, including general ontologies, domain ontologies, task ontologies, and application ontologies. The possibility of reusing existing ontologies includes compatibility, coverage, maintenance status, update frequency, and community support. S1-4, list the conceptual terms in the field, including scientific research subjects, scientific research results, scientific research tools, scientific research exchanges, research fields, research methods, cooperation networks, evaluation indicators, academic identifiers, research trends, metadata, and copyright information; S1-5, define classes and attributes, by creating classes and their hierarchical systems, and defining the attributes of each class, including object attributes and data attributes; S1-6, conduct quality assessment, evaluate the quality of the ontology through the standards of consistency, completeness, reusability, scalability, and accuracy, and update and improve the ontology based on the development of domain knowledge and user feedback.

[0026] It should be noted that when using a top-down approach to construct a knowledge graph of online academic resources, it is first necessary to define the ontology and data model of the knowledge graph, including defining the rules that need to be followed in constructing the knowledge graph, and using the top-level academic concepts as the basis to establish the top-level ontology of the model layer; secondly, these academic concepts and their hierarchical relationships are further refined, specific instances are extracted from high-quality online academic data sources and added to the knowledge base, and finally a structured knowledge graph of online academic resources is formed.

[0027] Computer science: including software engineering, artificial intelligence, data science, computer networks, and operating systems; biology: including molecular biology, cell biology, genetics, ecology, and bioinformatics; medicine: including clinical medicine, basic medicine, public health, pharmacy, and medical imaging; physics: including theoretical physics, experimental physics, condensed matter physics, particle physics, and astrophysics; chemistry: including organic chemistry, inorganic chemistry, physical chemistry, analytical chemistry, and materials chemistry; social sciences: including economics, sociology, psychology, political science, and law; humanities: including history, literature, philosophy, and art; interdisciplinary fields: including bioinformatics, computational biology, environmental science, materials science, and data science, involving the intersection and integration of multiple disciplines.

[0028] Resource types: including journal articles, conference papers, patents, project reports, data sets, software, books, manuscripts, historical documents, archives; Data sources: including academic databases (PubMed, Web of Science, Scopus), open access resources (arXiv, bioRxiv), academic communities (ResearchGate, Academia.edu), and public data of scientific research institutions; Time range: can be resources within a specific time period, or a combination of historical data and the latest research results; Geographical scope: can be resources worldwide, or resources in a specific country or region; Research methods: including experimental research, theoretical research, case studies, and literature reviews; Application areas: such as industrial applications, medical applications, educational applications, and environmental applications.

[0029] General ontology: aims to cover a wide range of knowledge, not limited to any specific field, and provides a set of general concepts and relationships that can be referenced or inherited by ontologies in different fields.

[0030] Domain ontology: focuses on the knowledge in a specific field, including the concepts, entities and the relationships between them in the field, so that the knowledge in the field can be structured and standardized.

[0031] Task ontology: describes the concepts and relationships needed to complete a specific task, focusing on actions and processes, as well as the objects and goals involved in these processes.

[0032] Application ontology: Ontology combines the requirements of specific applications and is usually a specific instantiation and extension of domain ontology and task ontology, designed to support the functionality of specific applications.

[0033] Research subjects: individuals or organizations involved in scientific research activities, which can be individual researchers, research teams, academic institutions (such as universities, research institutes), laboratories or R&D departments in enterprises; Scientific research results: covers various forms of scientific research outputs, including papers, patents, project reports, and data sets.

[0034] Scientific research tools: instruments, equipment, software, databases, standardized methods or any auxiliary tools and resources used in the process of scientific research.

[0035] Scientific research exchange: the exchange of information and knowledge between research subjects regarding the process, results, theories and methods of scientific research.

[0036] Research field: refers to a specific direction or field of academic research.

[0037] Collaboration network: describes the collaborative relationships between authors based on joint research and paper publication.

[0038] Research methods: involves specific research methods and techniques used in scientific research activities.

[0039] Evaluation indicators: indicators used to evaluate the influence of academic achievements.

[0040] Academic Identifier: Used to uniquely identify online academic resources.

[0041] Research News: Development trends and hot topics in the field of scientific research.

[0042] Metadata: Data that describes the properties and content of online scholarly resources, including creation date, author, and abstract.

[0043] Copyright information: Information on copyright ownership and usage permissions of online academic resources.

[0044] A class is an abstract concept used to describe entities in a knowledge graph, representing a group of objects with similar characteristics and behaviors.

[0045] A hierarchical system is a hierarchical structure between classes, which is used to represent the inheritance relationship between classes. By building a hierarchical system, entities in the knowledge graph can be organized and managed more clearly. It not only helps to reduce duplicate definitions, but also shares attributes and relationships through the inheritance mechanism, thereby improving the scalability and maintainability of the knowledge graph.

[0046] Attributes are characteristic descriptors of a class, used to describe specific information about an instance of a class. Attributes are divided into object properties: describing the relationship between class instances; and data properties: describing the relationship between class instances and data values. The following are examples of object properties and data properties in the online academic resource association selection ontology: # Object properties: # hasAuthor: connects journal articles or any scientific research results to their authors, who can be domain experts or research teams :hasAuthor rdf:type owl:ObjectProperty; rdfs:domain :ScholarlyResource; rdfs:range :ResearchEntity . # affiliatedWith: describes the affiliation between researchers and research institutions :affiliatedWith rdf:type owl:ObjectProperty; rdfs:domain :ResearchEntity ; rdfs:range :ResearchInstitution . # Data attributes: # publicationDate: indicates the publication date of online academic resources :publicationDate rdf:type owl:DatatypeProperty; rdfs:domain :ScholarlyResource; rdfs:range xsd:date . # title: the title of the online academic resource :title rdf:type owl:DatatypeProperty; rdfs:domain :ScholarlyResource; rdfs:range xsd:string . The following is an example of how to implement domain ontology evaluation and evolution in the ontology of online academic resource association selection: @prefix rdf:<http: / / www.w3.org / 1999 / 02 / 22-rdf-syntax-ns#> . @prefix rdfs:<http: / / www.w3.org / 2000 / 01 / rdf-schema#> . @prefix owl:<http: / / www.w3.org / 2002 / 07 / owl#> . @prefix :<http: / / example.org / ontology#> . # Define the ontology version :OntologyVersion rdf:type owl:Ontology; owl:versionInfo "1.0" . # Define ontology evaluation criteria :Consistency rdf:type owl:Class; rdfs:label "consistency" . :Completeness rdf:type owl:Class; rdfs:label "completeness" . :Reusability rdf:type owl:Class; rdfs:label "reusability" . :Extensibility rdf:type owl:Class; rdfs:label "Extensibility" . :Accuracy rdf:type owl:Class; rdfs:label "Accuracy" . # Define the ontology evolution mechanism :RegularAssessment rdf:type owl:Class; rdfs:label "Periodic Evaluation" . :VersionControl rdf:type owl:Class; rdfs:label "version control" . :UserNotification rdf:type owl:Class; rdfs:label "User Notification" . # Define the relationship between ontology evaluation and evolution :OntologyVersion :hasEvaluation :Consistency; :hasEvaluation :Completeness; :hasEvaluation :Reusability ; :hasEvaluation :Extensibility; :hasEvaluation :Accuracy . :OntologyVersion :hasEvolution :RegularAssessment; :hasEvolution :VersionControl; :hasEvolution :UserNotification . Step S2: preprocessing online academic resources, including organizing and screening topics, identifying and analyzing topic keywords, and selecting and evaluating sample data; Wherein step S2 also includes the following sub-steps: S2-1, using the Scopus database and subject vocabulary, classify journal articles by subject and screen out the subject directions related to the target research field; S2-2, through Elsevier's fingerprint engine technology, extract topic-related keywords and terms from unstructured text, create a weighted term index, and evaluate the importance of terms through the inverse document frequency statistics method. The specific formula is: Where IDF stands for Inverse Document Frequency, t is the term, and N is the total number of documents. is the number of documents containing term t; S2-3, based on the theory of provenance and the theory of information resource value, evaluate the disciplinary representativeness of the sample data, the authority of the original source, the reputation of the authors and their institutions, the citations of the literature, and the timeliness and completeness of the content.

[0047] It should be noted that Scopus is a large-scale citation and abstract database that is widely used in the field of academic research. It provides extensive coverage of global academic literature and is an important tool for researchers, scholars and academic institutions to obtain and analyze academic information. Scopus provides a wealth of citation analysis tools to help researchers analyze the citations of literature, evaluate academic influence, and support literature in multiple languages, including but not limited to English, Chinese, Japanese, French, etc., to meet the needs of researchers from different countries and regions.

[0048] Elsevier's Fingerprint Engine technology is an advanced text mining and analysis tool used to extract key information from academic literature and generate a weighted term index (also known as a "fingerprint") that reflects the core themes and content of the literature, helping researchers quickly identify and understand the key information of the literature. Provenance theory is a theoretical framework used to analyze and evaluate the source and quality of information resources. It emphasizes evaluating the reliability and credibility of information by tracing the source and dissemination path of information.

[0049] Information resource value theory is a theoretical framework for evaluating the value of information resources, emphasizing that the value of information resources depends not only on their content, but also on the degree to which they meet user needs.

[0050] Discipline representativeness: The traceability theory ensures that the sample data can comprehensively cover the key topics and research directions in the target discipline by tracing the source of information; the information resource value theory evaluates whether the sample data meets the user's needs in the discipline and provides valuable insights and solutions.

[0051] Authority of original source: The traceability theory evaluates whether the original source of the sample data is authoritative, including top journals, well-known conferences, and authoritative databases; the information resource value theory evaluates the authority of the sample data to ensure its degree of satisfaction of user needs.

[0052] Reputation of authors and their institutions: The traceability theory evaluates the influence of authors and institutions in the academic community and the quality of their research results by tracing their backgrounds; the information resource value theory evaluates the reputation of authors and institutions to ensure the extent to which they meet user needs.

[0053] Citations of documents: The source tracing theory analyzes the citations of documents and evaluates their influence in the academic community; the information resource value theory evaluates the citations of documents to ensure the extent to which they meet user needs.

[0054] Timeliness and completeness of content: The provenance theory ensures that the content of sample data is timely and reflects the latest research results; the information resource value theory evaluates the timeliness and completeness of sample data to ensure its degree of satisfaction with user needs.

[0055] Step S3, identifying entities from the preprocessed online academic resources and determining the relationships between entities to form a structured knowledge representation; Wherein step S3 also includes the following sub-steps: S3-1, identifying entities from the data patterns of online academic resources, including domain experts, work units, journal articles, scientific research projects, patents, research reports, academic conferences, conference reports, academic speeches, consulting reports, project completion reports, books, dissertations, data sets, scientific information, scientific blogs, and reviews; S3-2, determine the relationship between entities, including affiliation, writing, hosting, invention, participation, reporting, contribution, and mentioning; S3-3, entities and relationships are integrated into a structured knowledge representation in the form of RDF triples, where the form of the RDF triples is "entity-relationship-entity".

[0056] It should be noted that for field experts: name, title, work experience, educational background, and contact information; Work unit: name, type, address, establishment date, research direction; Journal articles: title, publication date, volume number, issue number, abstract, number of citations; Research project: name, start date, end date, project summary; Patents: patent name, application date, grant date, patent number, patent status, patent description; Research reports: title, publication date, report summary; Academic conferences: name, meeting date, meeting location, list of participants; Conference report: title, report date, report summary; Academic speeches: title, date, location, and abstract; Consulting reports: title, publication date, report summary, and opinion; Final report: title, date, summary; Books: title, publisher, publication date, ISBN number, edition, summary; Dissertation: title, date, abstract; Dataset: name, creation date, data format, data size, data description, access / download link; Science and technology information: title, release date, main content, related pictures, link address; Science blogs: title, publication date, main content, number of comments, number of visits, link address; Comments: content, comment date.

[0057] Affiliation relationship: the relationship between researchers and scientific research institutions; writing relationship: the relationship between authors and papers; hosting relationship: the relationship between researchers and scientific research projects; invention relationship: the relationship between researchers and patents; participation relationship: the relationship between researchers and academic conferences or research projects; reporting relationship: the relationship between researchers and conference reports or research reports; contribution relationship: the contribution of researchers to research projects; mention relationship: other documents or authors mentioned in the document.

[0058] RDF is a standard model for describing Web resources. It represents knowledge in the form of triples. Each RDF triple consists of three parts: subject: represents the resource being described, usually a URI (Uniform Resource Identifier) ​​or a blank node; predicate: represents the relationship between the subject and the object, usually a URI; object: represents information related to the subject, which can be a URI, a blank node or a literal value (string, number).

[0059] Step S4: Integrate the identified entities and relationships into the knowledge graph and store them in the graph database Neo4j; Wherein step S4 also includes the following sub-steps: S4-1, clean, remove duplicates, and standardize the extracted entities and relationships; S4-2, integrating entities and relationships from different data sources to make the knowledge graph consistent and complete, the integration including ontology mapping and entity alignment; S4-3, using the attribute graph structure of the Neo4j graph database, the nodes and edges in the knowledge graph are stored in the database, wherein the nodes represent entities and the edges represent relationships; S4-4, establish a data update mechanism, regularly monitor the update status of the data source, and update the knowledge base in an incremental manner.

[0060] It should be noted that cleaning: remove erroneous data, incomplete records and duplicates; use data cleaning tools (OpenRefine) or write scripts (Python) to identify and correct errors in the data.

[0061] Deduplication: Identify and merge duplicate entities and relationships; identify duplicates by comparing the attributes of entities and the characteristics of relationships, and merge them.

[0062] Standardization: unify data formats to ensure data consistency; define standard data formats and encoding rules to convert all data into a unified format.

[0063] Ontology mapping: Map the concepts and relationships in different data sources into a unified ontology framework, using ontology mapping tools or manually defining mapping rules, including: PREFIX oap:<http: / / oap.com / ontology#> PREFIX ao:<http: / / aoacademic.com / ontology#> oap:Paper owl:equivalentClass ao:AcademicDocument. Entity alignment: Identify and merge the same entities from different data sources, use similarity calculation methods to identify the same entities, and merge them, including: from difflib import SequenceMatcher def similarity(a, b): return SequenceMatcher(None, a, b).ratio() similarity("John Doe", "John D.") # Output similarity Neo4j is a high-performance graph database designed specifically for storing and querying graph-structured data. It represents and stores data through nodes, relationships, and attributes, making it very suitable for processing complex relationships and network structures. Neo4j's design enables it to efficiently process complex queries in large-scale data sets, especially in the fields of social networks, recommendation systems, and knowledge graphs.

[0064] The incremental method is a data processing and update strategy that processes and updates only the newly added or modified data instead of reprocessing the entire data set each time. It is suitable for processing large-scale data sets, especially in scenarios where data is frequently updated. It can significantly improve efficiency, save computing resources, and maintain the dynamic and timeliness of data.

[0065] Step S5: Use the constructed knowledge graph to perform resource association query and in-depth analysis, including statistical analysis, path analysis, key object analysis and cluster analysis; Wherein, in step S5, the following sub-steps are also included: S5-1, Statistical analysis: Statistical analysis of the number, degree distribution, and weighted degree distribution of nodes and edges in the knowledge graph to understand the structure and distribution of network academic resources in the field; S5-2, path analysis: exploring the connection paths between different resources, including shortest path analysis and all path exploration, wherein the shortest path analysis is used to determine the shortest path between two nodes, and the all path exploration is used to find all possible paths between two nodes; S5-3, key object analysis: identifying the most influential nodes through centrality analysis, which includes degree centrality analysis, betweenness centrality analysis, closeness centrality analysis and personalized ranking; S5-4, Cluster analysis: Use the k-means clustering method to group similar nodes in the graph and reveal potential research directions.

[0066] It should be noted that the number of nodes and edges: counting the number of nodes and edges in the knowledge graph can provide an intuitive understanding of the scale of the graph; degree distribution: analyzing the distribution of node degrees (that is, the number of edges directly connected to the node) helps to identify highly connected nodes in the graph; weighted degree analysis: when edges are given specific weights, weighted degree analysis can reveal the influence of nodes, and the weights are based on the number of citations and the intensity of cooperation.

[0067] The following is an example of performing a shortest path analysis: MATCH (a:Author {name: "John Doe"}), (b:Author {name: "Jane Smith"}) MATCH p = shortestPath((a)-[*..15]-(b)) RETURN p Here is an example that does all the path exploration: MATCH (a:Author {name: "John Doe"}), (b:Author {name: "Jane Smith"}) MATCH p = allShortestPaths((a)-[*..15]-(b)) RETURN p Degree centrality: Nodes with high degree centrality have the most direct connections and are often key figures or key papers in the research field.

[0068] Betweenness centrality: This analysis can identify nodes that act as “bridges” in the academic network.

[0069] Closeness centrality: measures the proximity of a node to all other nodes in the graph. Nodes with high closeness centrality are more efficient in knowledge dissemination.

[0070] Personalized ranking: A node importance measurement method based on network structure, which takes into account the connection quality and quantity of nodes.

[0071] K-means clustering is a commonly used cluster analysis method that iteratively divides data points into k categories so that the data points within each category are as similar as possible, while the data points between categories are as different as possible.

[0072] Step S6: Optimize the knowledge graph and related selection algorithm based on user feedback and progress in the academic field; Wherein, in step S6, the following sub-steps are also included: S6-1, user feedback mechanism, collects users’ evaluations of knowledge graphs and related selection results, analyzes feedback content, and evaluates users’ satisfaction with related selection results; S6-2, Academic Progress Monitoring: Tracking research results, methods and trends in the academic field, and updating knowledge and information in the knowledge graph; S6-3, Algorithm optimization: Based on user feedback and the development of the academic field, optimize the relevance selection algorithm, adjust the algorithm's parameters and strategies, and improve the accuracy and efficiency of relevance selection.

[0073] It should be noted that the user feedback mechanism includes feedback channels, feedback content and feedback processing. Feedback channels include online surveys, user forums, and direct feedback buttons; feedback content is provided through the design of structured feedback forms, including ratings, comments, and suggestions; feedback processing is carried out by regularly organizing and analyzing user feedback to identify common problems and improvement points.

[0074] Satisfaction evaluation: Based on user feedback, evaluate the user's satisfaction with the associated selection results. A combination of quantitative methods (rating system) and qualitative methods (user comments) can be used to fully understand the user's needs and expectations.

[0075] Track the academic field: Continuously monitor the latest research results, research methods and research trends in the academic field by subscribing to academic journals, attending conferences, and tracking academic communities and social media.

[0076] Update the knowledge graph: Update the knowledge and information in the knowledge graph in a timely manner based on the monitored academic progress, including adding new entities, relationships and attributes, updating the attribute values ​​of existing entities, and adjusting the structure of the knowledge graph to reflect new research directions.

[0077] Improve accuracy and efficiency: Improve the accuracy and efficiency of association selection by optimizing the algorithm, including introducing new algorithmic technologies (machine learning, deep learning), improving the performance of existing algorithms (reducing computational complexity) and adjusting algorithmic parameters (adjusting the threshold of similarity calculation).

[0078] In this way, a knowledge graph-driven online academic resource association selection optimization method can achieve efficient knowledge graph construction and management, accurate resource association and recommendation, dynamic optimization and continuous improvement, thereby significantly improving scientific research efficiency and the speed of knowledge innovation, and providing strong support for academic research.

[0079] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A knowledge graph-driven online academic resource association selection optimization method, characterized in that: The method includes: Step S1, defining the ontology through a top-down approach, and constructing the online academic resource association selection ontology using a seven-step method; Step S2, preprocessing online academic resources, including organization and screening of topics, identification and analysis of topic keywords, selection and evaluation of sample data; Step S3, identifying entities from the preprocessed online academic resources and determining the relationships between entities to form a structured knowledge representation; Step S4, integrating the identified entities and relationships into the knowledge graph and storing them in the graph database Neo4j; Step S5, using the constructed knowledge graph to perform resource association query and in-depth analysis, including statistical analysis, path analysis, key object analysis and cluster analysis; Step S6, optimizing the knowledge graph and associated selection algorithm based on user feedback and progress in the academic field; The seven-step method includes determining the professional field and scope of the ontology, examining the possibility of reusing existing ontologies, listing important terms in the ontology, defining classes and class hierarchies, defining class attributes, defining attribute facets, and creating attributes.

2. A knowledge graph-driven online academic resource association selection optimization method as claimed in claim 1, characterized in that: Wherein step S1 also includes the following sub-steps: S1-1, defines the ontology and data model through a top-down approach using a seven-step method; S1-2, determine the professional field and scope of the ontology, the professional field includes computer science, biology, medicine, physics, chemistry, social science, humanities, and interdisciplinary fields, and the scope includes resource type, data source, time range, geographical range, research method, and application field; S1-3, examine the possibility of reusing existing ontologies, including general ontologies, domain ontologies, task ontologies, and application ontologies. The possibility of reusing existing ontologies includes compatibility, coverage, maintenance status, update frequency, and community support. S1-4, list the conceptual terms in the field, including scientific research subjects, scientific research results, scientific research tools, scientific research exchanges, research fields, research methods, cooperation networks, evaluation indicators, academic identifiers, research trends, metadata, and copyright information; S1-5, define classes and attributes, by creating classes and their hierarchical systems, and defining the attributes of each class, including object attributes and data attributes; S1-6, conduct quality assessment, evaluate the quality of the ontology through the standards of consistency, completeness, reusability, scalability, and accuracy, and update and improve the ontology based on the development of domain knowledge and user feedback.

3. A knowledge graph-driven online academic resource association selection optimization method as claimed in claim 1, characterized in that: Wherein step S2 also includes the following sub-steps: S2-1, using the Scopus database and subject vocabulary, classify journal articles by subject and screen out the subject directions related to the target research field; S2-2, through Elsevier's fingerprint engine technology, extract topic-related keywords and terms from unstructured text, create a weighted term index, and evaluate the importance of terms through the inverse document frequency statistics method. The specific formula is: Where IDF stands for Inverse Document Frequency, t is the term, and N is the total number of documents. is the number of documents containing term t; S2-3, based on the theory of provenance and the theory of information resource value, evaluate the disciplinary representativeness of the sample data, the authority of the original source, the reputation of the authors and their institutions, the citations of the literature, and the timeliness and completeness of the content.

4. The knowledge graph-driven online academic resource association selection optimization method according to claim 1, characterized in that: Wherein step S3 also includes the following sub-steps: S3-1, identifying entities from the data patterns of online academic resources, including domain experts, work units, journal articles, scientific research projects, patents, research reports, academic conferences, conference reports, academic speeches, consulting reports, project completion reports, books, dissertations, data sets, scientific information, scientific blogs, and reviews; S3-2, determine the relationship between entities, including affiliation, writing, hosting, inventing, participating, reporting, contributing, and mentioning; S3-3, entities and relationships are integrated into a structured knowledge representation in the form of RDF triples, where the form of the RDF triples is "entity-relationship-entity".

5. The knowledge graph-driven online academic resource association selection optimization method according to claim 1, characterized in that: Wherein step S4 also includes the following sub-steps: S4-1, clean, remove duplicates, and standardize the extracted entities and relationships; S4-2, integrating entities and relationships from different data sources to make the knowledge graph consistent and complete, the integration including ontology mapping and entity alignment; S4-3, using the attribute graph structure of the Neo4j graph database, the nodes and edges in the knowledge graph are stored in the database, wherein the nodes represent entities and the edges represent relationships; S4-4, establish a data update mechanism, regularly monitor the update status of the data source, and update the knowledge base in an incremental manner.

6. A knowledge graph-driven online academic resource association selection optimization method as claimed in claim 1, characterized in that: Wherein, in step S5, the following sub-steps are also included: S5-1, Statistical analysis: Statistical analysis of the number, degree distribution, and weighted degree distribution of nodes and edges in the knowledge graph to understand the structure and distribution of network academic resources in the field; S5-2, path analysis: exploring the connection paths between different resources, including shortest path analysis and all path exploration, wherein the shortest path analysis is used to determine the shortest path between two nodes, and the all path exploration is used to find all possible paths between two nodes; S5-3, key object analysis: identifying the most influential nodes through centrality analysis, which includes degree centrality analysis, betweenness centrality analysis, closeness centrality analysis and personalized ranking; S5-4, Cluster analysis: Use the k-means clustering method to group similar nodes in the graph and reveal potential research directions.

7. The knowledge graph-driven online academic resource association selection optimization method according to claim 1, characterized in that: Wherein, in step S6, the following sub-steps are also included: S6-1, user feedback mechanism, collects users’ evaluations of knowledge graphs and related selection results, analyzes feedback content, and evaluates users’ satisfaction with related selection results; S6-2, Academic Progress Monitoring: Tracking research results, methods and trends in the academic field, and updating knowledge and information in the knowledge graph; S6-3, Algorithm optimization: Based on user feedback and the development of the academic field, optimize the relevance selection algorithm, adjust the algorithm's parameters and strategies, and improve the accuracy and efficiency of relevance selection.

Images