Methods, systems, and devices for constructing domain entity and event dual-center knowledge graphs

Abstract

This invention belongs to the field of data mining and recognition technology, and discloses a method, system, and device for constructing a dual-center knowledge graph of domain entities and events. The knowledge graph construction combines entity-centric and event-centric approaches, depicting static information such as entities, entity attributes, and entity relationships in real-world events, and can also express dynamic information such as event attributes and event relationships. A novel four-tuple data structure is designed to enable source tracing of data in the knowledge graph, supporting practical applications such as data access control, privacy protection, and license management. A novel derivative graph computation module is designed to support operations such as aggregation, statistics, association, and transformation of knowledge graph data, while also supporting intelligent computations such as graph embedding and machine learning models. The processed data is stored in a graph storage engine, accelerating knowledge graph querying and retrieval. This invention improves the fine-grained access control, privacy protection, and data management capabilities of knowledge graphs.

Description

Technical Field

[0001] This invention belongs to the field of data mining and identification technology, and in particular relates to a method, system and device for constructing a dual-center knowledge graph of domain entities and events. Background Technology

[0002] With the rapid improvement of data, model algorithms, and computing power, both computational intelligence and perceptual intelligence have developed rapidly. In some specific fields, machine perception has even surpassed that of humans. However, there are still many challenges in the area of ​​machine cognitive intelligence.

[0003] Knowledge graphs, as an important method for representing, organizing, and managing knowledge, have received significant attention from both academia and industry in recent years. Knowledge bases built using knowledge graph technology are considered the brain of machine cognitive intelligence and a fundamental path to achieving it. With knowledge graphs built upon perceptual intelligence, machines can understand language and data to a certain extent.

[0004] In recent years, knowledge representation methods that use "events" as basic units have gradually attracted attention from the academic community. "An event is all facts, not the sum of things," meaning that real-world events are composed of countless interrelated facts (i.e., events). Therefore, using "events" as units of human knowledge aligns with the laws governing human understanding of events. Knowledge bases using events as knowledge units can facilitate computer information processing similar to the human brain, that is, they can possess human-like cognitive abilities. Supplementing traditional knowledge bases with event knowledge to form event knowledge graphs allows the added event knowledge to accurately represent the relationships between events and entities such as people, places, and times. Utilizing the semantic relationships between different events, the evolutionary patterns and laws governing events can be expressed.

[0005] For a long time, knowledge representation methods, including knowledge graphs, have been based on symbols and have mostly processed structured or text data.

[0006] In many new application scenarios, such as event prediction and personalized recommendations in the financial and consumer sectors, users hope to obtain parent-child relationships between events through knowledge graphs to provide theoretical support for decision-making. An event, in this context, refers to an action that can occur or a state that something can be in. Furthermore, events occur in an orderly manner, with each event having a logical relationship with its predecessor, such as sequential or causal relationships. Therefore, starting from an event, the events that a user is interested in fall into two categories: the parent event (the event that triggered the event, i.e., the upstream event) and the child event (the event that the event will trigger, i.e., the downstream event). Interest in the parent event is equivalent to "finding clues to the occurrence of the event," while interest in the child event is equivalent to "finding the outcome of the event."

[0007] To address the aforementioned problems, prior art (CN202111309154.5) discloses a method for constructing a multimodal event knowledge graph, belonging to the field of knowledge engineering technology. It includes the following steps: data acquisition; construction of a domain event ontology; extraction of multimodal event trigger words; event element extraction; extraction of multimodal event relationships; disambiguation of event coreference based on the event ontology, merging identical sub-events to form a multimodal event subgraph; and merging the subgraphs to obtain a domain multimodal event knowledge graph. This invention improves the quality and efficiency of data acquisition, reduces labor costs, lowers the complexity of knowledge graph construction, and enhances the quality of the constructed knowledge graph.

[0008] Prior art (CN201910236491.2) provides a method, apparatus, and electronic device for constructing an event knowledge graph. The construction method includes: acquiring multiple event information groups from a target information source; determining multiple target events to be constructed from multiple sub-events and parent events; for each target event, based on the parent-child relationship of each pair of sub-events and parent events, searching for the parent target event and child target events of the target event from the multiple target events to obtain the search results corresponding to the target event; and constructing an event knowledge graph for each target event based on the search results corresponding to each target event. Each node in the event knowledge graph represents a target event, and any two nodes representing target events with a parent-child relationship are connected by a single-arrow connection. Compared with the prior art, the construction method provided by this invention can construct an event knowledge graph that reflects the relationships between various events.

[0009] Based on the above analysis, the problems and shortcomings of the existing technology are as follows:

[0010] (1) Existing general knowledge graphs and domain knowledge graphs mainly focus on static information such as entities and entity attributes, such as a large city, but cannot reflect events that occur in real life, as well as information such as the relationship between events and entities, and between events.

[0011] (2) Existing knowledge graph construction methods do not consider data source management and cannot classify and manage knowledge in the knowledge graph in batches, which are key requirements for data access control, privacy protection and license management.

[0012] (3) In the actual use of knowledge graphs, many applications rely on data statistics, aggregation, association and transformation through knowledge graphs. Existing knowledge graph construction methods lack support for these characteristics, which means that graph calculations need to be performed on the constructed graph when using the graph, which is not suitable for high-throughput applications. Summary of the Invention

[0013] To overcome the problems existing in related technologies, the embodiments disclosed in this invention provide a method for constructing a domain entity and event dual-center knowledge graph for the aggregation and analysis of massive multi-source heterogeneous data. In particular, it relates to a method, system, computer equipment, and medium for constructing a domain entity and event dual-center knowledge graph for the aggregation and analysis of massive multi-source heterogeneous data. It also relates to the fields of artificial intelligence, data mining, information retrieval, knowledge graphs, and recommendation algorithms.

[0014] The technical solution is as follows: A method for constructing a domain entity and event dual-center knowledge graph for aggregation analysis of massive multi-source heterogeneous data includes:

[0015] S1, which uses both entities and events as the core of the ontology to construct a domain event knowledge graph, represents static information such as entities, entity attributes, and entity relationships in real-world events, as well as dynamic information such as event attributes, events and entities, and events and events relationships.

[0016] S2 allows for the configuration of massive, multi-source, heterogeneous data input. Structured data input can be configured by inputting the database name, database type, database address, username, and password information. Unstructured and semi-structured data input can be configured by specifying file addresses, API interfaces, etc.

[0017] S3, according to data type and combined with the domain event knowledge graph constructed in step S1, cleans, extracts, transforms and loads the data, centered on entities and centered on events respectively; the data includes structured data, unstructured data and semi-structured data;

[0018] S4 generates a knowledge graph of the processed domain events according to the knowledge representation format;

[0019] S5 stores the data structure after unified knowledge representation into different types of databases such as relational databases, graph databases, key-value databases, and data warehouses.

[0020] S6, perform aggregation, statistics, association and transformation operations on the domain event knowledge graph data in step S4, and store the fused results back into the knowledge storage engine;

[0021] In one embodiment, in step S1, each piece of knowledge is defined as a quadruple (subject, predicate, object, proofance), i.e. (subject, predicate, object, proofance), where (subject, predicate, object) is the traditional knowledge triple expression, and proofance is a new element used to identify the source of the data.

[0022] In one embodiment, during step S3, which involves cleaning, extracting, transforming, and loading the data, data mapping is performed on the structured data. The relational database tables are mapped to the ontology definitions of entities and events to obtain knowledge triples. The source of the relational data table is used as the data source field to obtain a quadruple.

[0023] In one embodiment, in step S3, during the data cleaning, extraction, transformation, and loading process, unstructured and semi-structured data such as text, images, videos, and audio are classified and processed separately. The processing includes: using natural language processing algorithms to extract entities and events from text data, extracting information on people, time, location, and organizations, extracting relationships and attribute information between people and between people and organizations, and extracting event participants and event relationships, including relationships between events and between events and people; performing OCR recognition on image data and then converting it into text data; extracting keyframe images from video data and then converting it into image data for processing; and transcribing audio data into text data.

[0024] In one embodiment, in step S104, the quadruple representation (subject, predicate, object, proofance) is serialized to generate multiple knowledge serialization representations such as JSON-LD / RDFa / MCF / Turtle / N-Triples.

[0025] In one embodiment, in step S106, time series data is aggregated and statistically calculated by week / month / year. For data that requires semantic understanding, machine learning and deep learning methods are used to perform semantic association and entity disambiguation calculations, transform and associate the data, and store the fused results back into the knowledge storage engine.

[0026] In this invention, the aggregation method itself can adopt existing technology, but the application of aggregation, statistics and machine learning to the construction of a dual-center knowledge graph and its storage in the knowledge graph is a technical innovation of this invention.

[0027] Another objective of this invention is to provide a domain entity and event dual-center knowledge graph construction system for massive multi-source heterogeneous data aggregation and analysis. The domain entity and event dual-center knowledge graph construction system for massive multi-source heterogeneous data aggregation and analysis includes:

[0028] The ontology design module is used to build the conceptual layer model of the knowledge graph. It achieves semi-automatic ontology construction through a combination of human participation and machine assistance.

[0029] The data mapping module maps structured data by combining the ontology knowledge defined in the ontology design module.

[0030] The information extraction module provides configurable and scalable information extraction algorithms for unstructured and partially structured data, including entity-centered extraction algorithms and event-centered extraction algorithms.

[0031] The knowledge fusion module performs multi-source knowledge fusion on the data output by the data mapping module and the information extraction module, under the unified representation of the domain event knowledge ontology.

[0032] The knowledge update module is used to ensure the continuous iteration and updating of the domain event knowledge graph, with new knowledge constantly being generated over time;

[0033] The graph storage engine provides storage for domain event ontology and domain event knowledge graph, including relational databases, graph databases, key-value databases, data warehouses, and log databases;

[0034] The derived graph computation module performs aggregation, statistics, association, and transformation operations on knowledge graph data, including but not limited to existing algorithms, and stores the processed data in the graph storage engine;

[0035] The graph query service module provides diverse query interfaces for domain applications;

[0036] The graph visualization engine provides visualization capabilities for domain applications based on graph query services.

[0037] In one embodiment, the output of the ontology design module is a domain event ontology. The domain event ontology defines the conceptual layer of the domain event knowledge graph. By combining entity-centric ontology and event-centric ontology, it reduces the entities, entity attributes, entity relationships, event attributes, and event relationships described by the domain event knowledge graph.

[0038] The knowledge fusion module also supports customized knowledge fusion methods, providing results after ontology alignment and entity disambiguation.

[0039] Another object of the present invention is to provide a computer device, the computer device including a memory and a processor, the memory storing a computer program, the computer program being executed by the processor causing the processor to execute the method for constructing a domain entity and event dual-center knowledge graph for aggregation analysis of massive multi-source heterogeneous data.

[0040] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the described method for constructing a domain entity and event dual-center knowledge graph for aggregation analysis of massive multi-source heterogeneous data.

[0041] Combining all the above technical solutions, the advantages and positive effects of this invention are as follows:

[0042] First, addressing the technical problems existing in the prior art and the difficulty in solving them, and closely combining the technical solution to be protected by this invention with the results and data from the research and development process, this paper provides a detailed and in-depth analysis of how the technical solution of this invention solves the technical problems, and the inventive technical effects brought about after solving the problems. The specific description is as follows:

[0043] This invention enhances the knowledge representation capabilities of knowledge graphs, effectively expressing both static and dynamic information in real-world events.

[0044] This invention improves the fine-grained access control, privacy protection, and data management capabilities of knowledge graphs.

[0045] This invention enhances the high-throughput real-time service support capabilities and intelligent computing capabilities of knowledge graphs.

[0046] Second, considering the technical solution as a whole or from the perspective of the product, the technical effects and advantages of the technical solution to be protected by this invention are specifically described as follows:

[0047] The knowledge graph construction of this invention combines entity-centric and event-centric approaches, which can not only depict static information such as entities, entity attributes, and entity relationships in real-world events, but also express dynamic information such as event attributes and event relationships.

[0048] This invention designs a four-tuple data structure to enable source tracking of data in knowledge graphs, supporting practical applications such as data access control, privacy protection, and license management. The data used to construct a knowledge graph may originate from multiple different data sources. When a data source is updated, the knowledge graph with the four-tuple data structure can quickly retrieve the knowledge tuples that need updating. When access control and privacy protection are required for knowledge within the knowledge graph, the four-tuple data structure allows for fine-grained control at the knowledge tuple level. Furthermore, when the open-source license of the knowledge graph data source changes, such as from open source to closed source, the corresponding knowledge tuples can be quickly identified, avoiding potential intellectual property infringement risks.

[0049] This invention designs a derivative graph computation module that supports operations such as aggregation, statistics, association and transformation of knowledge graph data. It also supports intelligent computation such as graph embedding and machine learning models, and saves the processed data in a graph storage engine to accelerate knowledge graph query and retrieval.

[0050] Third, as further evidence of the inventiveness of this invention, it is also reflected in the fact that the technical solution of this invention solves the technical problem of fine-grained access control of knowledge tuples. Attached Figure Description

[0051] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0052] Figure 1 This is a schematic diagram of a domain entity and event dual-center knowledge graph construction system for the aggregation and analysis of massive multi-source heterogeneous data provided in an embodiment of the present invention;

[0053] Figure 2 This is a schematic diagram of a domain entity and event dual-center knowledge graph construction system for aggregation and analysis of massive multi-source heterogeneous data provided in this embodiment of the invention.

[0054] Figure 3 This is a flowchart of a method for constructing a domain entity and event dual-center knowledge graph for aggregation and analysis of massive multi-source heterogeneous data, provided in an embodiment of the present invention.

[0055] Figure 4 This is a schematic diagram of the method for constructing a dual-center knowledge graph of domain entities and events for the aggregation and analysis of massive multi-source heterogeneous data provided in this embodiment of the invention.

[0056] The diagram shows: 1. Ontology design module; 2. Data mapping module; 3. Information extraction module; 4. Knowledge fusion module; 5. Knowledge update module; 6. Graph storage engine; 7. Derivative graph computation module; 8. Graph query service module; 9. Graph visualization engine. Detailed Implementation

[0057] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0058] The domain entity and event dual-center knowledge graph construction method provided by this invention for the aggregation and analysis of massive multi-source heterogeneous data can be applied to any electronic device capable of information analysis and graph construction, such as laptops, tablets, and desktop computers.

[0059] I. Explanation of the Implementation Example:

[0060] Example 1

[0061] like Figure 1 As shown in the figure, the domain entity and event dual-center knowledge graph construction system for massive multi-source heterogeneous data aggregation analysis provided by the embodiments of the present invention includes: ontology design module 1, data mapping module 2, information extraction module 3, knowledge fusion module 4, knowledge update module 5, graph storage engine 6, derivative graph calculation module 7, graph query service module 8, and graph visualization engine 9.

[0062] Ontology Design Module 1 is used to construct the conceptual layer model of the knowledge graph. Through a combination of human involvement and machine assistance, it achieves semi-automated ontology construction through a human-machine collaboration. The output of Ontology Design Module 1 is a domain event ontology, which defines the conceptual layer of the domain event knowledge graph. Its innovation lies in combining entity-centric and event-centric ontology to regulate the entities, entity attributes, entity relationships, event arguments, and event relationships described in the domain event knowledge graph. The ontology of the dual-center knowledge graph can be formally defined as follows:

[0063]

[0064] in Represents an entity, Indicates an event, Represents entity attributes, Represents event arguments, Represents the relationships between entities. Indicates the relationship between entities and events. It indicates the relationship between events.

[0065] This system is designed for massive amounts of heterogeneous data from multiple sources in practical applications. The data input includes structured data, unstructured data, and semi-structured data.

[0066] Data mapping module 2 maps structured data by combining the ontology knowledge defined in ontology design module 1. The mapping mainly involves mapping the relational database to the designed ontology, directly transforming the structured data into a four-tuple dataset.

[0067] Information extraction module 3 provides configurable and scalable information extraction algorithms for unstructured and partially structured data, including entity-centric extraction algorithms and event-centric extraction algorithms.

[0068] The Knowledge Fusion Module 4 performs multi-source knowledge fusion on the data output from the Data Mapping Module and the Information Extraction Module, under a unified representation of the domain event knowledge ontology. This module supports various ontology alignment and entity disambiguation methods, enabling knowledge fusion of large-scale data. It also supports custom, personalized knowledge fusion methods to ensure the quality of the results after ontology alignment and entity disambiguation. The output of the Knowledge Fusion Module consists of knowledge entries within the domain event knowledge graph.

[0069] The knowledge update module 5 ensures the continuous iteration and updating of the domain event knowledge graph. As time goes by, new knowledge is constantly being generated, and old knowledge may also evolve. The knowledge update module ensures the timeliness of knowledge through updates to the schema layer and the data layer.

[0070] Graph Storage Engine 6 provides a complete storage solution for domain event ontology and domain event knowledge graph, including relational databases, graph databases, key-value databases, data warehouses, log databases, etc.

[0071] The derivative graph calculation module 7 is one of the important innovations of this invention. This module enables the domain event knowledge graph to not only store static ontology information and dynamic event information, but also to perform statistical operations and machine learning, deep learning and other calculations based on these data.

[0072] This derived graph computation module 7 supports operations such as aggregation, statistics, association, and transformation of knowledge graph data, and stores the processed data in the graph storage engine. This reduces the implementation overhead of large-scale graph computation and accelerates knowledge graph querying and retrieval. The module also supports intelligent computing such as graph embedding and machine learning models, providing functionalities not available in traditional knowledge graph queries.

[0073] The graph query service module 8 provides diverse query interfaces for domain applications, while the graph visualization engine 9 provides visualization functions for domain applications based on the graph query service.

[0074] Example 2

[0075] Based on the domain entity and event dual-center knowledge graph construction system for massive multi-source heterogeneous data aggregation and analysis provided in Example 1, the specific working principle is as follows: Figure 2 The following are included:

[0076] Phase 1:

[0077] After the structured data is mapped by the data mapping module 2, it is sent to the knowledge fusion module 4 for entity disambiguation and ontology alignment.

[0078] Unstructured data is extracted by information extraction module 3 with entities as the center, and attribute extraction is performed; then it is sent to knowledge fusion module 4 for entity disambiguation and ontology alignment.

[0079] Semi-structured data is processed by information extraction module 3, which extracts arguments and relations centered on events; then sent to knowledge fusion module 4 for entity disambiguation and ontology alignment.

[0080] Phase Two:

[0081] The ontology design module 1 involves the knowledge fusion module 4 for sending ontology data of local events; it is also stored in the graph storage engine 6. The derived graph calculation module 7 performs operations such as aggregation, statistics, association and transformation on the knowledge graph data, and saves the processed data in the graph storage engine.

[0082] The knowledge fusion module 4 performs multi-source knowledge fusion on the data output from the data mapping module 2 and the information extraction module 3, under the unified representation of the domain event knowledge ontology. It employs various ontology alignment and entity disambiguation methods to complete the knowledge fusion of large-scale data. Finally, it outputs a domain event knowledge graph and stores it in the graph storage engine 6.

[0083] The knowledge update module 5 ensures the continuous iteration and updating of the domain event knowledge graph. As time goes by, new knowledge is constantly generated, and old knowledge may also evolve. Through the updates of the schema layer and data layer, the timeliness of knowledge is ensured, and the updated domain event knowledge graph is stored in the graph storage engine 6.

[0084] Phase Three:

[0085] The knowledge graph query service module 8, after being processed by the derived graph calculation module 7, performs neighborhood applications through the query interface, and the information of the graph query service is visualized through the graph visualization engine 9.

[0086] Example 3

[0087] like Figure 3 As shown in the embodiments of the present invention, the method for constructing a domain entity and event dual-center knowledge graph for aggregation analysis of massive multi-source heterogeneous data includes the following steps:

[0088] S101 is a domain event knowledge graph that uses both entities and events as the core of its ontology construction. It represents static information such as entities, entity attributes, and entity relationships in real-world events, as well as dynamic information such as event attributes, events and entities, and events and their relationships.

[0089] S102, configure massive multi-source heterogeneous data input. Configure structured data input by inputting database name, database type, database address, username and password information, and configure unstructured and semi-structured data input by specifying file address, API interface and other methods.

[0090] S103, according to data type and combined with the domain event knowledge graph constructed in step S1, the data is cleaned, extracted, transformed, and loaded, centered on entities and events respectively; the data includes structured data, unstructured data, and semi-structured data;

[0091] S104, Generate the processed domain event knowledge graph data according to the knowledge representation format;

[0092] S105 stores the data structure after unified knowledge representation into different types of databases such as relational databases, graph databases, key-value databases, and data warehouses.

[0093] S106, perform aggregation, statistics, association and transformation operations on the domain event knowledge graph data in step S104, and store the fused results back into the knowledge storage engine;

[0094] S107 is designed for diverse business applications.

[0095] In this embodiment of the invention, step S101 specifically includes: constructing a domain event knowledge graph centered on both entities and events as ontologies; representing static information such as entities, entity attributes, and entity relationships in real-world events; and representing dynamic information such as event attributes, events and entities, and events and events relationships.

[0096] This invention employs an ontology design centered on both entities and events for a domain event knowledge graph. It standardizes the concepts of entities, entity attributes, entity relationships, event arguments, and event relationships described in the knowledge graph, enhancing its knowledge representation capabilities and effectively expressing both static and dynamic information. Each piece of knowledge is abstracted into a four-tuple (subject, predicate, object, proofance), where (subject, predicate, object) is a traditional triple representation of knowledge. Building upon this, the invention adds a new element, proofance, to identify the data source, improving the knowledge graph's fine-grained access control, privacy protection, and data management capabilities.

[0097] In this embodiment of the invention, step S102 specifically includes: configuring massive multi-source heterogeneous data input, which is configured by inputting information such as database name, database type, database address, username and password, and supports multiple types of data input such as relational databases and NoSQL databases.

[0098] In this embodiment of the invention, step S103 specifically includes: according to the data type and in conjunction with the designed ontology, cleaning, extracting, transforming, and loading the data, focusing on entities and events respectively. The specific implementation process of cleaning, extracting, transforming, and loading the data is handled differently for structured and unstructured data types.

[0099] Based on the defined ontology concepts, structured data is mapped by mapping relational database tables to the ontology definitions of entities and events, resulting in (subject, predicate, object) knowledge triples. The source of the relational data table is used as the data source field, resulting in a quadruple (subject, predicate, object, proofance). Unstructured and semi-structured data, such as text, images, videos, and audio, are processed separately. The processing steps are as follows: For text data, natural language processing algorithms are used for entity extraction and event extraction, extracting entity information such as people, time, location, and organization; extracting entity relationships and attribute information such as person-person and person-organization relationships; and extracting event participants and event associations, including event-event relationships and event-person relationships. For image data, OCR recognition is performed, and then it is converted into text data for processing. For video data, keyframe images are extracted and then converted into image data for processing. For audio data, text transcription is performed, converting it into text data for processing. The text data processing process is the same as the steps described above and will not be repeated here.

[0100] In this embodiment of the invention, step S104 specifically includes: generating the processed data according to the knowledge representation format, serializing the quadruple representation (subject, predicate, object, proofance), and generating various knowledge serialization representation files such as JSON-LD / RDFa / MCF / Turtle / N-Triples.

[0101] In this embodiment of the invention, step S105 specifically includes: storing the knowledge representation file into different types of databases such as relational databases, graph databases, key-value databases, and data warehouses.

[0102] In this embodiment of the invention, step S106 specifically includes: performing operations such as aggregation, statistics, association, and transformation on knowledge graph data according to business application needs, as well as intelligent computing such as machine learning; performing aggregation and statistical calculations on time series data by week / month / year; and using machine learning and deep learning methods to perform semantic association and entity disambiguation calculations on data that requires semantic understanding, transforming and associating the data, and storing the fused results back into the knowledge storage engine.

[0103] The derived graph computation module performs calculations and saves the computational services required for knowledge graph business in advance, thereby improving the high-throughput real-time business support capability and intelligent computing capability of knowledge graph.

[0104] In this embodiment of the invention, step S107 specifically includes: providing WebService services based on RestfulAPI for diverse business applications.

[0105] Example 4

[0106] like Figure 4 As shown in the embodiments of the present invention, the method for constructing a domain entity and event dual-center knowledge graph for aggregation analysis of massive multi-source heterogeneous data includes the following steps:

[0107] (1) Design an ontology for the domain event knowledge graph with entities and events as the dual centers.

[0108] (2) Configure massive multi-source heterogeneous data input.

[0109] (3) According to the data type and in combination with the designed ontology, the data is cleaned, extracted, transformed and loaded, respectively centered on the entity and centered on the event.

[0110] (4) Generate the processed data according to the knowledge representation format.

[0111] (5) Perform knowledge integration and, based on business application needs, perform operations such as aggregation, statistics, association and transformation of knowledge graph data, as well as intelligent computing such as machine learning. For time series data, perform data aggregation and statistical calculations by week / month / year.

[0112] (6) Store the knowledge representation file in different types of databases such as relational databases, graph databases, key-value databases, and data warehouses.

[0113] (7) Provide WebService services based on RestfulAPI for diverse business applications.

[0114] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0115] The information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of the present invention. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0116] Those skilled in the art will understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the functions described above can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this invention. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0117] II. Application Examples:

[0118] Application Example 1

[0119] An application embodiment of the present invention provides a computer device comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the steps in any of the above-described method embodiments.

[0120] Application Example 2

[0121] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps in the above-described method embodiments.

[0122] Application Example 3

[0123] The present invention also provides an information data processing terminal, which is used to provide a user input interface to implement the steps in the above method embodiments when executed on an electronic device. The information data processing terminal is not limited to mobile phones, computers, and switches.

[0124] Application Example 4

[0125] The present invention also provides a server that, when executed on an electronic device, provides a user input interface to implement the steps described in the above method embodiments.

[0126] Application Example 5

[0127] The present invention provides a computer program product that, when run on an electronic device, enables the electronic device to implement the steps described in the various method embodiments above.

[0128] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying the computer program code to a photographic device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0129] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention and within the spirit and principles of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for constructing a field entity and event bicenter knowledge graph for massive multi-source heterogeneous data aggregation analysis, characterized in that, The method for constructing a domain entity and event dual-center knowledge graph for the aggregation and analysis of massive multi-source heterogeneous data includes: S1. Construct a domain event knowledge graph with entities and events as the core of the ontology construction. Represent static information of entities, entity attributes, and entity relationships in the real world, as well as dynamic information of event attributes, events and entities, and events and events relationships. The ontology of a dual-center knowledge graph is formally defined as: in, Represents an entity, Indicates an event, Represents entity attributes, Represents event arguments, Represents the relationships between entities. Indicates the relationship between entities and events. Indicates the relationship between events; S2 allows for the configuration of massive, multi-source, heterogeneous data input. Structured data input can be configured by inputting the database name, database type, database address, username, and password. Unstructured and semi-structured data input can be configured by specifying the file address and API interface. S3, according to data type and combined with the domain event knowledge graph constructed in step S1, the data is cleaned, extracted, transformed, and loaded, centered on entities and events respectively; the data includes structured data, unstructured data, and semi-structured data; in the data cleaning, extraction, transformation, and loading process, unstructured and semi-structured data such as text, images, videos, and audio are classified and processed separately. The processing includes: for text data, using natural language processing algorithms to extract entities and events, extracting information on people, time, location, and organizations, extracting relationships and attribute information between people and between people and organizations, extracting event participants and event relationships, including relationships between events and between events and people; performing OCR recognition on image data and then converting it into text data; extracting keyframe images from video data and then converting them into image data for processing; and transcribing audio data into text data. S4 generates a knowledge graph of the processed domain events according to the knowledge representation format; S5 stores the data structure after unified knowledge representation into different types of databases such as relational databases, graph databases, key-value databases, and data warehouses. S6, perform aggregation, statistics, association and transformation operations on the domain event knowledge graph data in step S4, and store the fused results back into the knowledge storage engine; S7 is designed for diverse business applications; In step S4, the quadruple representation (subject, predicate, object, provenance) is serialized to generate multiple knowledge serialization representations such as JSON-LD / RDFa / MCF / Turtle / N-Triples; In step S6, time series data is aggregated and statistically calculated by week / month / year. For data that requires semantic understanding, machine learning and deep learning methods are used to perform semantic association and entity disambiguation calculations, transform and associate the data, and store the fused results back into the knowledge storage engine.

2. The method for constructing a domain entity and event dual-center knowledge graph for aggregation analysis of massive multi-source heterogeneous data as described in claim 1, characterized in that, In step S1, each piece of knowledge is defined as a quadruple, including: subject, predicate, object, and proofance. The subject, predicate, and object are traditional knowledge triples, and proofance is a new element used to identify the source of the data.

3. The method for constructing a domain entity and event dual-center knowledge graph for aggregation analysis of massive multi-source heterogeneous data as described in claim 1, characterized in that, In step S3, during the data cleaning, extraction, transformation, and loading process, data mapping is performed on structured data. Relational database tables are mapped to entity and event ontology definitions to obtain knowledge triples. The source of the relational data table is used as the data source field to obtain a quadruple.

4. A domain entity and event dual-center knowledge graph construction system for the aggregation and analysis of massive multi-source heterogeneous data, characterized in that, The method for constructing a domain entity and event dual-center knowledge graph for massive multi-source heterogeneous data aggregation analysis as described in claim 1, wherein the domain entity and event dual-center knowledge graph construction system for massive multi-source heterogeneous data aggregation analysis comprises: The ontology design module (1) is used to construct the conceptual layer model of the knowledge graph. Through human participation and machine assistance, the semi-automatic construction of the ontology is achieved through human-machine combination. The data mapping module (2) maps structured data by combining the ontology knowledge defined in the ontology design module 1; The information extraction module (3) provides configurable and scalable information extraction algorithms for unstructured and semi-structured data, including entity-centered extraction algorithms and event-centered extraction algorithms. The knowledge fusion module (4) performs multi-source knowledge fusion on the data output by the data mapping module and the information extraction module under the unified representation of the domain event knowledge ontology; The knowledge update module (5) is used to ensure the continuous iteration and updating of the domain event knowledge graph, with new knowledge constantly being generated over time; The graph storage engine (6) provides storage for domain event ontology and domain event knowledge graph, including relational databases, graph databases, key-value databases, data warehouses, and log databases; The derived graph calculation module (7) performs aggregation, statistics, association and transformation operations on knowledge graph data, and saves the processed data in the graph storage engine; The graph query service module (8) provides diverse query interfaces for domain applications; The graph visualization engine (9) provides visualization functions for domain applications based on graph query services.

5. The domain entity and event dual-center knowledge graph construction system for massive multi-source heterogeneous data aggregation and analysis as described in claim 4, characterized in that, The output of the ontology design module (1) is the domain event ontology. The domain event ontology defines the conceptual layer of the domain event knowledge graph. Combining the entity-centered ontology and the event-centered ontology, it modifies the entities, entity attributes, entity relationships, event attributes, and event relationships described by the domain event knowledge graph. The knowledge fusion module (4) also supports custom personalized knowledge fusion methods to provide results after ontology alignment and entity disambiguation.

6. A computer device, characterized in that, The computer device includes a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the processor performs the domain entity and event dual-center knowledge graph construction method for massive multi-source heterogeneous data aggregation analysis as described in any one of claims 1-3.

7. A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the domain entity and event dual-center knowledge graph construction method for aggregation analysis of massive multi-source heterogeneous data as described in any one of claims 1-3.

Images