Construction system and method of soft science knowledge graph, terminal equipment and storage medium

Abstract

The embodiment of the invention provides a soft science knowledge graph construction system and method, terminal equipment and a storage medium, and relates to the technical field of knowledge graphs. The system comprises a knowledge definition module used for describing the structure of standardized soft science domain knowledge, defining the category and attribute of the soft science domain knowledge, and generating initial data defined by the soft science domain knowledge; the data access module is used for introducing the initial data generated in the knowledge definition module into the knowledge graph platform in a preset mode; the graph construction module is used for converting the initial data into SPO triple data, executing fusion of the SPO triple data and generating a knowledge graph applied to the field of soft science; and the graph application module is used for providing the knowledge graph to the knowledge graph platform for use through the graph calculation output capability component. By adopting the technical scheme provided by the embodiment of the invention, the utilization efficiency of the knowledge in the soft science field can be improved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of knowledge graph, and particularly relate to a soft science knowledge graph construction system and method, a terminal device, and a storage medium. BACKGROUND

[0002] Soft science is a group of emerging disciplines with high comprehensiveness, which is gradually formed by the cross development of modern natural science and social science.

[0003] In related technologies, in the field of soft science, the data sources are extensive and complex, and the formats, qualities, and completeness of the data differ greatly, thereby leading to low utilization efficiency of the knowledge in the field of soft science. SUMMARY

[0004] Embodiments of the present application provide a soft science knowledge graph construction system and method, a terminal device, and a storage medium, which can improve the utilization efficiency of the knowledge in the field of soft science. The technical solutions provided by embodiments of the present application are as follows.

[0005] According to an aspect of embodiments of the present application, a soft science knowledge graph construction system is provided, and the system comprises a knowledge definition module, a data access module, a graph construction module, and a graph application module, wherein,

[0006] The knowledge definition module is configured to describe the structure of standardized soft science field knowledge, define the categories and attributes of the soft science field knowledge, and generate initial data of the soft science field knowledge definition;

[0007] The data access module is configured to introduce the initial data generated by the knowledge definition module into a knowledge graph platform in a preset manner;

[0008] The graph construction module is configured to convert the initial data into structured subject-predicate-object (SPO) triple data, perform fusion of the SPO triple data, and generate a knowledge graph applied to the field of soft science; wherein the SPO triple data comprises entities, attributes, and attribute values;

[0009] The graph application module is configured to provide the knowledge graph to the knowledge graph platform through a graph computation output capability component.

[0010] In some embodiments, the data access module comprises a structured data access unit and an unstructured data access unit; wherein,

[0011] The structured data access unit is configured to perform access of the structured data in the initial data;

[0012] The unstructured data access unit is configured to perform access of unstructured data in the initial data.

[0013] In some embodiments, the graph construction module comprises a knowledge extraction unit, an attribute mapping unit, a data cleaning unit, and a knowledge fusion unit, wherein,

[0014] The knowledge extraction unit is configured to convert the initial data into structured SPO triple data through a knowledge extraction technique.

[0015] The attribute mapping unit is configured to convert all obtained SPO triple data into intermediate data conforming to a data description specification format.

[0016] The data cleaning unit is configured to formulate cleaning rules through data attribute values, eliminate intermediate data that does not conform to data type constraints, and unify the data structure of the intermediate data.

[0017] The knowledge fusion unit is configured to perform fusion operations on the intermediate data to generate a complete knowledge graph.

[0018] In some embodiments, the graph application module comprises a graph question answering unit and a graph visualization unit, wherein,

[0019] The graph question answering unit is configured to interact with a user terminal, automatically extract keywords of a user terminal question, disassemble a user terminal questioning intention, and perform deep reasoning and searching.

[0020] The graph visualization unit is configured to display the correlation between entities, perform switching queries, and display multi-degree relationship path calculations.

[0021] In some embodiments, the system further comprises:

[0022] A graph database module configured to store the knowledge graph in a graph data structure and provide graph data analysis and query functions.

[0023] According to an aspect of an embodiment of the present application, a method for constructing a soft science knowledge graph is provided, the method comprising:

[0024] Determining an application scenario of a soft science field and obtaining initial data, wherein the initial data comprises structured data and unstructured data;

[0025] Performing preprocessing of the structured data based on a first preset rule to obtain first intermediate data;

[0026] Performing preprocessing of the unstructured data based on a second preset rule to obtain second intermediate data;

[0027] Perform a fusion operation on the first intermediate data and the second intermediate data to generate target data to construct a knowledge graph in a soft science field application scenario.

[0028] In some embodiments, the first intermediate data is obtained by performing preprocessing on the structured data based on a first preset rule, including:

[0029] The structured data is converted into first SPO triple data through a knowledge introduction technology, wherein the first SPO triple data includes entities, attributes, and attribute values.

[0030] The first SPO triple data is converted into first intermediate data conforming to a data description specification format through attribute mapping and data cleaning technology.

[0031] In some embodiments, the second intermediate data is obtained by performing preprocessing on the unstructured data based on a second preset rule, including:

[0032] The unstructured data is converted into structured second SPO triple data through a knowledge extraction technology.

[0033] The structured second SPO triple data is converted into second intermediate data conforming to a data description specification format through attribute mapping and data cleaning technology.

[0034] According to an aspect of an embodiment of the present application, a soft science knowledge graph construction device is provided, characterized by comprising:

[0035] A data determination module is configured to determine an application scenario of a soft science field and obtain initial data, wherein the initial data includes structured data and unstructured data.

[0036] A preprocessing module is configured to perform preprocessing on the structured data based on a first preset rule to obtain first intermediate data.

[0037] The preprocessing module is further configured to perform preprocessing on the unstructured data based on a second preset rule to obtain second intermediate data.

[0038] A data generation module is configured to perform a fusion operation on the first intermediate data and the second intermediate data to generate target data to construct a knowledge graph in a soft science field application scenario.

[0039] In some embodiments, the preprocessing module is configured to:

[0040] The structured data is converted into first SPO triple data through a knowledge introduction technology, wherein the first SPO triple data includes entities, attributes, and attribute values.

[0041] The first SPO triple data is converted into first intermediate data conforming to a data description specification format by using attribute mapping and data cleaning technology.

[0042] In some embodiments, the preprocessing module is configured to:

[0043] The unstructured data is converted into structured second SPO triple data by using knowledge extraction technology.

[0044] The structured second SPO triple data is converted into second intermediate data conforming to a data description specification format by using attribute mapping and data cleaning technology.

[0045] According to an aspect of the embodiments of the present application, a terminal device is provided, which comprises a processor and a memory, the memory storing a computer program, the computer program being loaded and executed by the processor to implement the method for constructing a soft science knowledge graph.

[0046] According to an aspect of the embodiments of the present application, a computer readable storage medium is provided, which stores a computer program, the computer program being loaded and executed by a processor to implement the method for constructing a soft science knowledge graph.

[0047] According to an aspect of the embodiments of the present application, a computer program product is provided, which is loaded and executed by a processor to implement the method for constructing a soft science knowledge graph.

[0048] The technical solutions provided by the embodiments of the present application can have the following beneficial effects:

[0049] The knowledge graph applied to the soft science field is generated by the knowledge definition module, the data access module, the graph construction module and the graph application module, which has high adaptability to the application scenarios of the soft science field, better meets the needs of knowledge production, organization and application in the soft science field, and improves the utilization efficiency of knowledge in the soft science field.

[0050] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0052] Figure 1 is a block diagram of a terminal device provided by an embodiment of the present application;

[0053] Figure 2 is a block diagram of a soft science knowledge graph construction system provided by an embodiment of the present application;

[0054] Figure 3 is a block diagram of a soft science knowledge graph construction system provided by another embodiment of the present application;

[0055] Figure 4 is a flowchart of a soft science knowledge graph construction method provided by an embodiment of the present application;

[0056] Figure 5 is a block diagram of a soft science knowledge graph construction device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0057] The exemplary embodiments will be described in detail herein with reference to the attached drawings. In the following description, the same numbers are used to indicate the same or similar elements, unless otherwise represented. The embodiments described in the following exemplary embodiments do not represent all the embodiments consistent with the present application. Rather, they are merely examples of methods consistent with some aspects of the present application, as detailed in the appended claims.

[0058] With the development of society, emerging technologies and applications such as mobile Internet applications, artificial intelligence, and autonomous driving are rapidly entering the lives of the public. In the new digital era, it is necessary to understand and process more, faster, and more complex information than ever before in order to utilize the latest technological achievements, improve operational efficiency and decision-making intelligence, and survive and develop in a complex competitive environment.

[0059] Soft science is a group of emerging disciplines with high comprehensiveness that gradually formed from the cross-development of modern natural science and social science. Since soft science is a cross-disciplinary field, its research often involves knowledge from multiple fields and requires the use of multiple research methods. However, there are still some problems in the research methods of the soft science field, such as the selection and use of research methods being not rigorous and standardized enough, and the lack of unified standards and norms, which leads to poor comparability and repeatability of research results, high research costs, and low research reliability, which is a technical problem that needs to be urgently addressed.

[0060] The method provided in this application can be executed by a terminal device, which refers to an electronic device with data computing, processing, and storage capabilities. This terminal device can be a terminal such as a PC (Personal Computer), tablet computer, smartphone, wearable device, or intelligent robot; or it can be a server. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0061] The technical solution of this application will be described and illustrated below through several embodiments.

[0062] Taking running on a terminal device as an example, Figure 1 This is a block diagram of a terminal device for a soft science knowledge graph construction system provided in an embodiment of this application. Figure 1 As shown, the terminal device may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU (Microcontroller Unit) or a programmable gate array (FPGA)) and a memory 104 for storing data are also shown. Optionally, the terminal device may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the terminal device described above. For example, the terminal device may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0063] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / modules corresponding to the soft science knowledge graph construction system in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to terminal devices via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0064] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the terminal device. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0065] The technical solution of this application will be described and illustrated below through several embodiments.

[0066] Please refer to Figure 2 This diagram illustrates a block diagram of a soft science knowledge graph construction system provided in an embodiment of this application. Figure 1 As shown, a soft science knowledge graph construction system 200 is applied to a knowledge graph platform. This system 200 may include: a knowledge definition module 201 with communication connections, a data access module 202, a graph construction module 203, and a graph application module 204. The knowledge definition module 201 describes the structure of standardized soft science knowledge, defines the categories and attributes of soft science knowledge, and generates initial data for soft science knowledge definitions. The data access module 202 imports the initial data generated in the knowledge definition module into the knowledge graph platform through preset methods, including manual upload, API (Application Programming Interface), etc. The knowledge graph is introduced via an application programming interface (API) or a direct database connection. The graph construction module 203 is used to transform the initial data into structured SPO triple data and perform the fusion of the SPO triple data to generate a knowledge graph applicable to the field of soft science. The SPO triple data includes entities, attributes, and attribute values. The knowledge graph application module 204 is used to provide the knowledge graph to the knowledge graph platform through the knowledge graph computing output capability component.

[0067] In some embodiments, knowledge graphs describe knowledge and model the relationships between all things in the objective world in a structured, computer-understandable way, expressing information on the Internet in a form that is closer to human cognition of the world, and have advantages such as large scale, high quality, structure-friendly, and strong interpretability.

[0068] In summary, the technical solution provided in this application generates a knowledge graph applicable to the field of soft science through a knowledge definition module, a data access module, a graph construction module, and a graph application module. It has high adaptability to application scenarios in the field of soft science, better meets the needs of knowledge production, organization, and application in the field of soft science, and improves the utilization efficiency of knowledge in the field of soft science.

[0069] In some possible implementations, the knowledge definition module 201 plays a crucial role in the field of soft science, as it is used to describe and standardize the structure of knowledge in this field. By defining the categories and attributes of knowledge in the soft science field, the knowledge definition module 201 can provide a clear and comprehensive classification and description system for knowledge in this field.

[0070] In some embodiments, the knowledge definition module 201 can be used to perform the following tasks:

[0071] Category Definition: Based on the characteristics and needs of the soft science field, knowledge is divided into several categories. These categories can be divided according to research topics, field applications, methodologies, etc., to ensure that the classification of knowledge is both comprehensive and detailed.

[0072] Attribute definition: Define relevant attributes for each category. Attributes are a further description and limitation of the category, used to more accurately describe the characteristics and connotations of knowledge in the field of soft science.

[0073] Initial data generation: Based on the defined categories and attributes, an initial dataset is generated, which provides the basic data source for subsequent knowledge management, querying and application.

[0074] Structured storage: Stores defined categories, attributes, and initial data in a structured manner, facilitating efficient retrieval, querying, and use.

[0075] Dynamic updates and maintenance: Considering the dynamic development characteristics of the soft science field, the knowledge definition module should support the dynamic updates and maintenance of categories and attributes to adapt to the continuous changes and development of domain knowledge.

[0076] Interface design: Provide a user-friendly interface and API interface to enable users to easily query, obtain and use knowledge in the field of soft science.

[0077] In some embodiments, the construction and application of the knowledge definition module 201 enable standardized management of knowledge in the field of soft science, which helps improve the reusability, sharing, and innovation of knowledge. Simultaneously, the knowledge definition module 201 also helps promote the in-depth development of research in the field of soft science and accelerates the transformation and application of research results.

[0078] In some embodiments, the data access module 202 plays a crucial role in the construction of the knowledge graph platform, responsible for introducing the initial data generated in the knowledge definition module 201 into the knowledge graph platform. The data access module 202 provides multiple preset methods, making data introduction more flexible and efficient. The following is a detailed description of each preset method:

[0079] Manual upload: For example, users can manually upload unstructured data or data in a specific format to the knowledge graph platform. This method allows users to control the quality and accuracy of the data, but it may be less efficient with large-scale data.

[0080] API Import: Importing data via Application Programming Interfaces (APIs) is a common approach. Users can write specific programs or scripts to automatically upload data to the knowledge graph platform using the interfaces provided by the API. This method is suitable for automated data integration and batch data processing.

[0081] Direct database connection: For data already stored in the database, the data access module 202 can directly connect to the database to achieve automatic data synchronization and updates. This method can ensure the real-time performance and accuracy of the data, but it is necessary to ensure the security and access permissions of the database.

[0082] In some embodiments, through these preset methods, the data access module 202 can meet the data import needs of different users and scenarios. Users can choose an appropriate method according to the actual situation to import the initial data generated in the knowledge definition module 201 into the knowledge graph platform, providing a complete and accurate data foundation for subsequent knowledge graph construction and application. At the same time, the data access module 202 can also provide corresponding data verification and validation mechanisms to ensure the quality and compliance of the imported data.

[0083] In some embodiments, the data access module 202 may include: a structured data access unit and an unstructured data access unit; wherein the structured data access unit is used to access structured data in the initial data; the unstructured data access unit is used to access unstructured data in the initial data, wherein the unstructured data includes at least one of Word documents, Excel documents, image-type PDFs, and audio / video files.

[0084] In some embodiments, the design of the structured data access unit and the unstructured data access unit of the data access module 202 takes into account the processing needs of different types of data.

[0085] For example, structured data typically has a clear format and specifications, such as tables and data in databases. This type of data is characterized by clearly defined data fields and a unified format, facilitating efficient data processing and analysis. The structured data access unit is specifically designed to process this type of data. It can parse various common data formats, such as CSV, JSON, and XML, and convert them into formats usable by knowledge graph platforms. The structured data access unit also provides data validation functions to ensure data integrity and accuracy.

[0086] In some embodiments, unstructured data refers to data without a fixed format or structure, such as Word documents, Excel documents, images, PDF files, and audio / video files. This type of data typically requires more complex processing and analysis methods. The unstructured data access unit is specifically designed to process this type of data. It can parse various file formats, convert file content into a format usable by the knowledge graph platform, and extract key information from it. For image and audio / video files, the unstructured data access unit can also provide technologies such as OCR (Optical Character Recognition) and speech-to-text to extract meaningful information from these media.

[0087] Through this modular design, the data access module 202 can flexibly process various types of data, meeting the processing needs of the knowledge graph platform for different types of data. At the same time, this design also helps improve the speed and efficiency of data processing, ensuring the stability and efficiency of the knowledge graph platform.

[0088] In some embodiments, the graph construction module 203 is one of the core components of the knowledge graph construction system. Its main task is to transform the initial data into structured SPO triple data and perform data fusion to generate a knowledge graph applicable to the field of soft science.

[0089] In this context, SPO triples are the basic building blocks of a knowledge graph, used to represent relationships between entities. The knowledge graph construction module 203 first transforms the initial data into structured SPO triple data. This process involves entity recognition, relationship extraction, and attribute assignment. Through natural language processing and text mining techniques, the module can extract meaningful information from unstructured data and transform it into a structured SPO triple format. After generating the SPO triple data, the knowledge graph construction module 203 performs data fusion. Since the initial data may come from different data sources, its format, quality, and representation may differ. To ensure the accuracy and consistency of the knowledge graph, the module needs to fuse this data. This includes eliminating redundant information, handling conflicting data, and filling in missing values. Through data fusion, the knowledge graph platform can provide a unified and consistent knowledge representation. Finally, after the transformation and fusion of the SPO triple data, the knowledge graph construction module 203 generates a knowledge graph applicable to the field of soft science. This knowledge graph can graphically represent relationships between entities, providing users with an intuitive and easy-to-use interface for knowledge querying and display. In addition, the knowledge graph platform can also provide API interfaces, allowing users to perform custom queries and data mining operations to meet the application needs of different scenarios.

[0090] Through the aforementioned functions of the knowledge graph construction module 203, the knowledge graph platform can effectively extract, integrate, and display knowledge in the field of soft science from initial data, providing users with comprehensive information support and services. This structured knowledge representation helps improve the accuracy and efficiency of information retrieval, promotes knowledge sharing and dissemination, and drives the research and application development in the field of soft science.

[0091] In some possible implementations, the knowledge graph construction module 203 may include: a knowledge extraction unit, an attribute mapping unit, a data cleaning unit, and a knowledge fusion unit. The knowledge extraction unit is used to transform the initial data into structured SPO triple data using knowledge extraction techniques. The attribute mapping unit is used to transform all the obtained SPO triple data into intermediate data conforming to a data description specification format. The data cleaning unit is used to formulate cleaning rules based on data attribute values, remove intermediate data that does not conform to data type constraints, and unify the data structure of the intermediate data. The knowledge fusion unit is used to perform a fusion operation on the intermediate data to generate a complete knowledge graph.

[0092] In some embodiments, the knowledge extraction unit is responsible for extracting meaningful information from the initial data and transforming it into structured SPO triple data. It utilizes techniques such as natural language processing and text mining to identify entities, extract relationships, and assign attribute values ​​from unstructured data. This unit can handle different types of data sources, extract key information, and provide foundational data for subsequent knowledge graph construction.

[0093] In some embodiments, the attribute mapping unit is responsible for normalizing the SPO triple data. It transforms the obtained SPO triple data into intermediate data conforming to a data description specification format. This process involves mapping and classifying entities, attributes, and attribute values ​​to ensure data consistency and comparability. Through attribute mapping, data from different sources can be unified into a standard format, facilitating subsequent data processing and analysis.

[0094] In some embodiments, the data cleaning unit is responsible for cleaning and organizing intermediate data. It establishes cleaning rules based on data attribute values, removes intermediate data that does not conform to data type constraints, and standardizes the data structure of the intermediate data. This process helps improve data quality and accuracy, providing a guarantee for generating a complete and reliable knowledge graph. The data cleaning unit also supports the detection and processing of abnormal data, ensuring the reliability and stability of the knowledge graph.

[0095] In some embodiments, the knowledge fusion unit is responsible for fusing intermediate data to generate a complete knowledge graph. It utilizes specific algorithms and rules to integrate and correlate intermediate data from different data sources. Through knowledge fusion, operations such as eliminating redundant information, handling conflicting data, and filling in missing values ​​are achieved, thereby constructing a comprehensive and consistent knowledge graph. The knowledge fusion unit also supports the dynamic updating and maintenance of the knowledge graph to adapt to the continuous changes and development of domain knowledge.

[0096] In some embodiments, through the collaborative work of these components, the knowledge mapping module 203 can extract, integrate, and display knowledge in the field of soft science from initial data, providing users with comprehensive information support and services. This structured knowledge representation helps improve the accuracy and efficiency of information retrieval, promotes knowledge sharing and dissemination, and drives the research and application development in the field of soft science.

[0097] In some embodiments, the graph application module 204 is an important component of the knowledge graph platform. It is responsible for providing the constructed knowledge graph to the knowledge graph platform for use and supporting computation and output based on the knowledge graph.

[0098] In some embodiments, the knowledge graph application module 204 can provide access to and query functions for the knowledge graph. Users can query and retrieve the knowledge graph through an intuitive query language or API interface. The knowledge graph application module 204 supports precise queries, fuzzy queries, or queries based on logical rules for entities, attributes, and properties to meet different user needs.

[0099] In some embodiments, leveraging structured data within the knowledge graph, the graph application module 204 also supports various graph computation and analysis functions. These include path analysis, entity association analysis, and relational reasoning. Through these computations and analyses, users can delve deeper into the hidden information within the knowledge graph and discover connections and patterns between pieces of knowledge.

[0100] In some embodiments, the knowledge graph application module 204 provides a visual display and interactive interface, allowing users to intuitively view the structure and content of the knowledge graph and explore and analyze it through interactive operations. The visualization supports various view and chart formats, such as network diagrams, tree diagrams, and tables, to meet the needs of different users.

[0101] In some embodiments, the knowledge graph application module 204 also supports integration with other application systems to achieve knowledge sharing and transfer. It provides standard API interfaces, allowing third-party applications to access and utilize the functionality of the knowledge graph platform. Furthermore, the knowledge graph application module 204 supports customized development, allowing users to perform secondary development and functional expansion according to specific needs.

[0102] In some embodiments, as knowledge in the field of soft science is continuously updated and developed, the knowledge graph application module 204 supports the dynamic updating and maintenance of the knowledge graph. Through integration with the data access module 202, new data can be automatically or manually introduced into the knowledge graph, maintaining its timeliness and accuracy.

[0103] Through these functions, the knowledge graph application module 204 effectively provides the constructed knowledge graph to the knowledge graph platform, offering users comprehensive knowledge services and support. This helps improve the efficiency and accuracy of knowledge utilization, promoting research and application development in the field of soft science. Simultaneously, through integration with other applications and customized development, the knowledge graph platform can better meet users' actual needs, driving innovation and development in knowledge management.

[0104] In some embodiments, the graph application module 204 may include a graph question-answering unit and a graph visualization unit. The graph question-answering unit interacts with the user, automatically extracting keywords from the user's questions, deconstructing the user's question intent, and performing deep reasoning and searching. The graph visualization unit displays the relationships between entities and performs switching queries to demonstrate multi-degree relationship path calculations. The graph application module 204 can be further subdivided into the graph question-answering unit and the graph visualization unit, each undertaking different functions to meet the different needs of users in knowledge graph applications.

[0105] In some embodiments, the graph-based question-answering unit is primarily responsible for interacting with users. It automatically extracts keywords from user questions, deconstructs the user's intent, performs deep reasoning and searching, and provides accurate and rapid answers. This unit utilizes natural language processing and machine learning techniques to understand the user's natural language questions and transform them into queries on the knowledge graph. Then, it extracts relevant information from the knowledge graph, reasons and integrates it, and generates structured, understandable answers to return to the user. This question-answering function can significantly improve the efficiency and accuracy of users acquiring knowledge.

[0106] In some embodiments, the graph visualization unit is responsible for displaying the relationships between entities, providing functions such as query switching and multi-degree relationship path calculation. This unit utilizes a graphical interface and visualization technology to present data from the knowledge graph to users in an intuitive way. Users can easily view the relationships, relationship paths, and degree relationships between entities. Furthermore, the visualization unit supports query switching, allowing users to perform data filtering, selection, and sorting operations through simple interactive operations. This visualization method helps users better understand the structure and content of the knowledge graph and discover hidden patterns and relationships.

[0107] In some embodiments, through the collaborative work of these two units, the knowledge graph application module 204 can provide comprehensive and efficient knowledge services, meeting users' needs in knowledge acquisition, querying, and analysis. This knowledge graph-based application module can greatly promote knowledge sharing, dissemination, and innovation, driving research and application development in the field of soft science. Simultaneously, through integration and customized development with other applications, the knowledge graph platform can better meet users' actual needs, promoting innovation and development in knowledge management.

[0108] See Figure 3 , Figure 3This is a block diagram of a knowledge graph construction system based on the field of soft science, provided as another embodiment of this application. The knowledge graph construction system 200 based on the field of soft science may further include: a graph database module 205, which stores the knowledge graph in a graph data structure and provides graph data analysis and query functions.

[0109] Graph Database Module 205 is a core component of the knowledge graph platform used for storing and managing knowledge graphs. It uses a graph data structure to store entities, attributes, and relationships, and provides efficient data analysis and query functions.

[0110] The graph database module 205 employs an efficient data storage method, storing entities, attributes, and relationships from the knowledge graph in the database in the form of a graph data structure. This structured storage method ensures data integrity and consistency, and provides fast data access performance.

[0111] The Graph Database module 205 supports rich graph data analysis capabilities, including path analysis, community discovery, and association rule mining. These analytical functions help users delve deeper into hidden information within knowledge graphs, revealing complex relationships and patterns between entities. Through graph data analysis, users can better understand the structure and implications of domain knowledge.

[0112] The graph database module 205 provides flexible graph data query capabilities, supporting precise, fuzzy, and complex queries based on the graph structure. Users can perform multi-dimensional and multi-level queries on the knowledge graph through query languages ​​or API interfaces. Query results are presented to users in a visual manner, facilitating quick retrieval and understanding.

[0113] In some embodiments, to improve the retrieval efficiency of graph data, the graph database module 205 supports an efficient indexing mechanism. By establishing an index, the module can quickly locate relevant entities and relationships, providing a fast query response. Furthermore, the module also supports approximate queries and fuzzy matching of graph data to meet different user query needs.

[0114] Through these functions, the graph database module 205 can store knowledge graphs in a graph data structure and provide efficient data analysis and query functions.

[0115] As can be seen, this application provides a knowledge graph construction system for the soft science field, consisting of a knowledge definition module, a data access module, a graph construction module, and a graph application module connected by communication. The knowledge definition module describes the structure of standardized knowledge in the soft science field, defines the categories and attributes of soft science knowledge, and generates initial data for the knowledge definition. The data access module imports the initial data generated in the knowledge definition module into the knowledge graph platform through a preset method. The graph construction module transforms the initial data into structured SPO triple data and performs SPO triple data fusion to generate a knowledge graph applicable to the soft science field. The graph application module provides the knowledge graph to the knowledge graph platform for use through a graph computing output capability component. It can generate knowledge graphs applicable to the soft science field using the knowledge definition module, data access module, graph construction module, and graph application module, exhibiting high adaptability to soft science application scenarios and effectively meeting the needs of knowledge production, organization, and application in the soft science field.

[0116] See Figure 4 , Figure 4 The flowchart of a method for constructing a knowledge graph based on the field of soft science provided in one embodiment of this application includes at least one of the following steps (S401-404).

[0117] S401: Determine the application scenarios in the field of soft science and obtain initial data; wherein, the initial data includes structured data and unstructured data.

[0118] In some embodiments, the field of soft science encompasses a wide range of applications, including but not limited to policy research, business management, education, and social research.

[0119] For example, in policy research applications such as policy formulation, policy effectiveness evaluation, and public affairs management, initial data can come from publicly available government data, policy documents, survey data, and statistical data. In enterprise management applications such as strategic planning, market analysis, and human resource management, initial data can come from internal business data, market research data, and employee information.

[0120] In some embodiments, when acquiring initial data, attention should be paid to the integrity and accuracy of the data, as well as its timeliness and updability. Furthermore, since data sources in the field of soft science are diverse, potentially including structured data (such as tables in databases) and unstructured data (such as text, images, audio, etc.), appropriate data cleaning and preprocessing techniques are subsequently required to ensure data usability and accuracy.

[0121] S402: Based on the first preset rule, perform preprocessing on the structured data to obtain the first intermediate data.

[0122] In some embodiments, the structured data is preprocessed based on a first preset rule, mainly by cleaning, transforming and normalizing the data to eliminate redundancy, anomalies and inconsistencies in the data, thereby improving the quality and accuracy of the data.

[0123] Data cleaning is a crucial step in the preprocessing stage, primarily involving identifying and handling outliers, missing values, and redundant information in the data. For example, for numerical data, it may be necessary to handle outliers (values ​​outside the reasonable range), while for text data, it may be necessary to handle spelling errors, formatting errors, or non-standard expressions. Data transformation converts raw data into a format suitable for knowledge extraction and fusion. This may include standardizing date formats, classifying text, and converting strings into entities. Normalization mainly involves standardizing and normalizing the data to ensure consistency across different scales. For example, age data may be converted into a unified age group classification, or income data may be normalized to a relative proportion. Based on predefined rules and classification standards, attribute mapping and classification are performed on the data. This helps to unify data from different sources into a standard format, facilitating subsequent data processing and analysis.

[0124] In some embodiments, the above steps can yield first intermediate data. This data, after preprocessing, has had redundant and anomalous information removed, has been transformed into data conforming to a standardized format, and the relationships between entities have been extracted. This intermediate data can be directly used in subsequent knowledge fusion and graph construction processes.

[0125] In some embodiments, the step of performing preprocessing on the structured data based on a first preset rule to obtain first intermediate data may include:

[0126] 1. The structured data is transformed into first SPO triple data using knowledge introduction technology; wherein the first SPO triple data includes an entity, an attribute, and an attribute value;

[0127] 2. Utilize attribute mapping and data cleaning techniques to transform the first SPO triplet data into first intermediate data that conforms to the data description specification format.

[0128] In some embodiments, structured data is transformed into first SPO triple data using knowledge introduction technology. Here, SPO is a representation of a triple, where S, P, and O represent the subject (entity), predicate (attribute), and object (attribute value), respectively. Through knowledge introduction technology, entities, attributes, and attribute values ​​are identified from the structured data to form the first SPO triple data.

[0129] In some embodiments, after obtaining the first SPO triplet data, attribute mapping technology is used to normalize this data. Attribute mapping is a process of uniformly classifying and identifying entities, attributes, and attribute values ​​to ensure data consistency and comparability. Simultaneously, data cleaning technology is used to clean and organize the data, removing data that does not conform to data type constraints and unifying the data structure. After attribute mapping and data cleaning, the first SPO triplet data is transformed into first intermediate data conforming to the data description specification format.

[0130] Through the above steps, the structured data is preprocessed based on the first preset rule. The first intermediate data obtained is data that conforms to the standard format after being processed by techniques such as knowledge introduction, attribute mapping and data cleaning. This intermediate data provides a high-quality foundation for subsequent knowledge fusion and graph construction.

[0131] S403: Based on the second preset rule, perform preprocessing on the unstructured data to obtain the second intermediate data.

[0132] In some embodiments, the preprocessing of unstructured data is a crucial step in knowledge graph construction because unstructured data typically occupies significant data resources and contains rich semantic information. The preprocessing of unstructured data mainly involves processing data such as text, images, and audio to make it understandable and usable by computers.

[0133] For example, text cleaning and deduplication remove irrelevant characters, stop words, punctuation marks, etc., from the text. Repetitive text content is removed to ensure data consistency and uniqueness. Word segmentation and part-of-speech tagging divide the text into independent words or phrases. Each word is tagged with its part of speech to determine its semantic role in the sentence (e.g., noun, verb, adjective). Relationships between entities are extracted from the text, constructing relation triples to form the SPO representation of unstructured data. The preprocessed unstructured data is mapped to a knowledge graph and associated with structured data, ensuring consistency between the unstructured data and the entities, attributes, and properties in the knowledge graph.

[0134] Therefore, the second intermediate data is unstructured data that has been processed by cleaning, word segmentation, entity recognition, and relation extraction. It can be further integrated into the knowledge graph to enrich its semantic information.

[0135] In some embodiments, the step of performing preprocessing on the unstructured data based on a second preset rule to obtain second intermediate data may include:

[0136] The unstructured data is transformed into structured second SPO triplet data using knowledge extraction technology; the structured second SPO triplet data is then transformed into second intermediate data conforming to the data description specification format using attribute mapping and data cleaning techniques.

[0137] Knowledge extraction involves extracting useful information from unstructured data and transforming it into a structured form. By using techniques such as natural language processing and text mining, entities, attributes, and attribute values ​​are identified from unstructured data, forming second SpO triples. The main purpose of this step is to transform unstructured data such as text, images, and audio into a structured data format to facilitate subsequent data processing and analysis.

[0138] In some embodiments, after obtaining the structured second SPO triplet data, attribute mapping technology is used to normalize this data. Attribute mapping is a process of uniformly classifying and identifying entities, attributes, and attribute values ​​to ensure data consistency and comparability. Simultaneously, data cleaning techniques are used to clean and organize the data, removing data that does not conform to data type constraints and unifying the data structure. After attribute mapping and data cleaning, the structured second SPO triplet data is transformed into second intermediate data conforming to the data description specification format.

[0139] Through the above steps, unstructured data is preprocessed based on the second preset rule to obtain the second intermediate data, which is data that conforms to the standardized format after being processed by techniques such as knowledge extraction, attribute mapping, and data cleaning. This intermediate data provides a high-quality foundation for subsequent knowledge fusion and graph construction.

[0140] S404: Perform a fusion operation on the first intermediate data and the second intermediate data to generate target data, so as to construct a knowledge graph in the application scenario of soft science.

[0141] In some embodiments, entities, attributes, and attribute values ​​in the first and second intermediate data are compared for alignment and matching. For example, similarity algorithms (such as cosine similarity, Jaccard similarity, etc.) are used for entity matching. The first and second intermediate data are integrated into a unified data structure, merging similar entities, attributes, and attribute values ​​from different data sources to eliminate redundant information. Relation completion techniques are used to derive new triplet relations based on known triplet relations, enriching the connections in the knowledge graph through relational reasoning. Graph database technology is used to store the fused data in the form of a graph. By designing a suitable graph structure, including nodes, edges, and attributes, to reflect the knowledge structure and relationships within the domain, a data update mechanism is established to ensure the knowledge graph is continuously updated with the addition of new data. A visual interface for the knowledge graph is designed to facilitate intuitive viewing and interaction by users.

[0142] By providing rich query and analysis tools, users can explore data in the knowledge graph from multiple dimensions. Through the above steps, the first and second intermediate data are fused to generate target data for constructing a knowledge graph for soft science application scenarios. The constructed knowledge graph will encompass rich information from both structured and unstructured data, providing a comprehensive view of knowledge in the soft science field and offering strong support for applications such as decision support, trend analysis, and question answering.

[0143] As can be seen, this application first identifies the application scenarios in the field of soft science and obtains initial data; based on a first preset rule, it performs preprocessing on structured data to obtain first intermediate data; based on a second preset rule, it performs preprocessing on unstructured data to obtain second intermediate data; finally, it performs a fusion operation on the first and second intermediate data to generate target data, thereby constructing a knowledge graph for the application scenarios in the field of soft science. The knowledge graph constructed by the above method can cover rich information from both structured and unstructured data, providing a comprehensive view of knowledge in the field of soft science and providing strong support for applications such as decision support, trend analysis, and question answering.

[0144] The following are embodiments of the apparatus described in this application, which can be used to execute the embodiments of the method described in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in this application.

[0145] Please refer to Figure 5 This diagram illustrates a block diagram of a soft science knowledge graph construction apparatus according to an embodiment of this application. The apparatus has the functionality to implement the aforementioned soft science knowledge graph construction method example; this functionality can be implemented in hardware or by hardware executing corresponding software. The apparatus 500 can be the terminal device described above, or it can be installed on a terminal device. The apparatus 500 may include: a data determination module 510, a preprocessing module 520, and a data generation module 530.

[0146] The data determination module 510 is used to determine the application scenario in the field of soft science and obtain initial data, wherein the initial data includes structured data and unstructured data.

[0147] The preprocessing module 520 is used to perform preprocessing on the structured data based on a first preset rule to obtain first intermediate data.

[0148] The preprocessing module 520 is further configured to perform preprocessing on the unstructured data based on a second preset rule to obtain second intermediate data.

[0149] The data generation module 530 is used to perform a fusion operation on the first intermediate data and the second intermediate data to generate target data, so as to construct a knowledge graph in the application scenario of soft science.

[0150] In some embodiments, the preprocessing module 520 is configured to:

[0151] The structured data is transformed into first SPO triple data through knowledge introduction technology, wherein the first SPO triple data includes an entity, an attribute, and an attribute value.

[0152] Using attribute mapping and data cleaning techniques, the first SPO triplet data is transformed into first intermediate data that conforms to the data description specification format.

[0153] In some embodiments, the preprocessing module 520 is configured to:

[0154] The unstructured data is transformed into structured second SPO triplet data through knowledge extraction technology;

[0155] By using attribute mapping and data cleaning techniques, the structured second SPO triplet data is transformed into second intermediate data that conforms to the data description specification format.

[0156] In summary, the technical solution provided in this application generates a knowledge graph applicable to the field of soft science through a knowledge definition module, a data access module, a graph construction module, and a graph application module. It has high adaptability to application scenarios in the field of soft science, better meets the needs of knowledge production, organization, and application in the field of soft science, and improves the utilization efficiency of knowledge in the field of soft science.

[0157] It should be noted that the apparatus provided in the above embodiments is only illustrated by the division of the above functional modules when implementing its functions. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0158] In an exemplary embodiment, a computer-readable storage medium is also provided, wherein a computer program is stored therein, which, when executed by a processor, implements the above-described method for constructing a soft science knowledge graph.

[0159] In an exemplary embodiment, a computer program product is also provided, which is loaded and executed by a processor to implement the above-described method for constructing a soft science knowledge graph.

[0160] It should be understood that "multiple" as used in this article refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0161] The above description is merely an exemplary embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A system for constructing a soft science knowledge graph, characterized in that, The system includes: a knowledge definition module, a data access module, a graph construction module, and a graph application module, wherein... The knowledge definition module is used to describe the structure of standardized soft science knowledge, define the categories and attributes of soft science knowledge, and generate initial data for the definition of soft science knowledge. The data access module is used to import the initial data generated in the knowledge definition module into the knowledge graph platform through a preset method; The graph construction module is used to transform the initial data into structured subject-verb-object SPO triple data, and to perform the fusion of the SPO triple data to generate a knowledge graph applicable to the soft science field; wherein, the SPO triple data includes entities, attributes, and attribute values; The graph application module is used to provide the knowledge graph to the knowledge graph platform through the graph computing output capability component.

2. The system according to claim 1, characterized in that, The data access module includes: a structured data access unit and an unstructured data access unit; wherein, The structured data access unit is used to access the structured data in the initial data; The unstructured data access unit is used to access the unstructured data in the initial data.

3. The system according to claim 2, characterized in that, The graph construction module includes: a knowledge extraction unit, an attribute mapping unit, a data cleaning unit, and a knowledge fusion unit; wherein... The knowledge extraction unit is used to transform the initial data into structured SPO triplet data using knowledge extraction technology. The attribute mapping unit is used to convert all the obtained SPO triplet data into intermediate data that conforms to the data description specification format; The data cleaning unit is used to formulate cleaning rules based on data attribute values, remove intermediate data that does not conform to data type constraints, and unify the data structure of the intermediate data. The knowledge fusion unit is used to perform fusion operations on the intermediate data to generate a complete knowledge graph.

4. The system according to claim 3, characterized in that, The graph application module includes: a graph question-and-answer unit and a graph visualization unit, wherein... The graph-based question-answering unit is used to interact with the user terminal, automatically extract keywords from the user terminal's questions, deconstruct the user terminal's questioning intent, and perform deep reasoning and searching; The graph visualization unit is used to display the relationships between entities, perform switching queries, and display multi-degree relationship path calculations.

5. The system according to any one of claims 1 to 4, characterized in that, The system also includes: The graph database module is used to store the knowledge graph in a graph data structure and to provide graph data analysis and query functions.

6. A method for constructing a soft science knowledge graph, characterized in that, The method includes: Identify application scenarios in the field of soft science and obtain initial data, wherein the initial data includes structured data and unstructured data; Based on the first preset rule, preprocessing of the structured data is performed to obtain the first intermediate data; Based on the second preset rule, preprocessing of the unstructured data is performed to obtain the second intermediate data; Perform a fusion operation on the first intermediate data and the second intermediate data to generate target data, so as to construct a knowledge graph in the application scenario of soft science.

7. The method according to claim 6, characterized in that, The step of preprocessing the structured data based on a first preset rule to obtain first intermediate data includes: The structured data is transformed into first SPO triple data through knowledge introduction technology, wherein the first SPO triple data includes an entity, an attribute, and an attribute value. Using attribute mapping and data cleaning techniques, the first SPO triplet data is transformed into first intermediate data that conforms to the data description specification format.

8. The method according to claim 6, characterized in that, The step of preprocessing the unstructured data based on the second preset rule to obtain the second intermediate data includes: The unstructured data is transformed into structured second SPO triplet data through knowledge extraction technology; By using attribute mapping and data cleaning techniques, the structured second SPO triplet data is transformed into second intermediate data that conforms to the data description specification format.

9. A terminal device, characterized in that, The terminal device includes a processor and a memory, the memory storing a computer program, which is loaded and executed by the processor to implement the method for constructing a soft science knowledge graph as described in any one of claims 6 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which is loaded and executed by a processor to implement the method for constructing a soft science knowledge graph as described in any one of claims 6 to 8.

Images