A case knowledge graph construction platform and method based on big data technology
Through big data technology, the case knowledge graph is constructed, the heterogeneous data processing problem in the case is solved, the dynamic analysis and expansion of the case information chain is realized, the efficiency of obtaining and analyzing case clues is improved, and the poor data analysis effect in the existing technology is solved. question.
Patent Information
- Application Number
- CN202310165592.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-02-15
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2043-02-15
AI Technical Summary
The existing technology is difficult to effectively handle heterogeneous data in cases, resulting in a narrow acquisition of case clues, poor data analysis effect, and the inability to promptly reflect the true situation of the case, and the inability to understand the case from a comprehensive perspective.
Using a case knowledge graph based on big data technology to build a platform, collect and process structured, semi-structured and unstructured data through case data source collection modules, processing modules, information extraction modules and information chain generation modules, collect and process structured, semi-structured and unstructured data, and convert them into metadata. , establish data entities and relationships, and generate case information chains.
It realizes the extraction and dynamic analysis of multi-source heterogeneous data of case information, improves the efficiency of production and expansion of case information chains, enhances the ability to obtain and analyze case clues, and can promptly reflect the real situation of the case.
Smart Images

Figure CN116662559B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of information technology processing technology, and in particular to a case knowledge graph construction platform and method based on big data technology. Background Art
[0002] Currently, there are not many companies engaged in case knowledge graph research on the market, so case knowledge graphs usually appear in the form of platforms or solutions, and basic case analysis capabilities are achieved through a single type of tool. However, the clues in the case have a large number of heterogeneous relationships, with diverse data types and complex association structures; most of the numerous clue information is hidden in some unstructured data of video and text types, which is difficult to retrieve through traditional big data technology, and usually requires manual retrieval of data and extraction of useful information from it; the clues, people, events, and physical evidence in the case are complicated, and it is inefficient to manually sort out the clues and relationships of the case, and it is impossible to understand the case from the overall picture; and as the case goes on over time, the information about people, time, things, events, physical evidence and their relationships also changes accordingly.
[0003] Traditional platform tool systems mainly collect data manually and convert it into structured data for entry into the system for routine data analysis. They cannot timely reflect the true situation of the case and obtain the latest research and analysis data. This results in a narrow scope of case clue acquisition, poor data analysis results, poor connectivity between case clues, and an inability to mine clues based on existing case data. Summary of the invention
[0004] In view of this, an embodiment of the present invention provides a case knowledge graph construction platform and method based on big data technology.
[0005] The first aspect of the present invention provides a case knowledge graph construction platform based on big data technology, including the following modules:
[0006] Case data source collection module: The case data source processing module is used to collect case information, which includes structured data, semi-structured data and unstructured data;
[0007] Case processing module: The case processing module is used to perform structured processing on semi-structured data and unstructured data, so that the semi-structured data and unstructured data are converted into structured data; the structured data contains multiple metadata, and the metadata carries one or more data identifiers; the metadata is used to represent the attributes of various elements of the case;
[0008] Case information extraction module: The information extraction module is used to extract information from the structured data, extract data entities in the form of metadata from the structured data, establish data relationships between multiple data entities, and load the data entities and data relationships into the knowledge graph;
[0009] Case information chain generation module: The case information chain generation module is used to generate a case information chain based on the data entities and data relationships stored in the knowledge graph.
[0010] Furthermore, the structured data includes process information in the case information; the semi-structured data includes text information and table information in the case information; and the unstructured data includes image information and video information in the case information.
[0011] Furthermore, the structuring of the semi-structured data and the unstructured data so as to convert the semi-structured data and the unstructured data into structured data specifically includes the following steps:
[0012] Identify semi-structured and unstructured data to obtain multiple metadata;
[0013] Add data identifiers to each metadata;
[0014] Adding a reference point to the metadata according to the added data identifier, wherein the reference point is used to locate the position of the metadata;
[0015] Extract metadata based on reference points and classify metadata according to data identifiers;
[0016] The metadata is placed into multiple data templates classified based on data identifiers to obtain structured data sets classified by data identifiers.
[0017] Furthermore, the identification of semi-structured data is specifically completed through semantic segmentation; among the data identifiers added by the metadata, for text information, the metadata referring to the text information is used as the data identifier of the target metadata; for table information, the metadata of the table header is used as the data identifier of the target metadata.
[0018] Furthermore, in the step of identifying the unstructured data to obtain multiple metadata, for the image information, OCR is used to scan the image, and the metadata in the image information is read and output;
[0019] For image information, use the following steps to read the metadata in the image information:
[0020] Divide the image information into multiple video frames, and read the frame information of each video frame;
[0021] Restore the storyboard information according to the adjacent multiple frame information, and restore the scene information according to the adjacent multiple storyboard information;
[0022] The event information in the image is restored based on multiple scene information, and the event information is output as metadata.
[0023] Furthermore, before the step of extracting metadata based on the reference point, the following steps are also included:
[0024] The metadata is unified in form according to the form template corresponding to the data identifier.
[0025] Furthermore, the knowledge extraction module extracts data entities in the form of metadata from the structured data, establishes data relationships between multiple data entities, and loads the data entities and data relationships into the knowledge graph; specifically, the following steps are included:
[0026] Use named entity recognition tools to identify structured data and extract metadata as data entities;
[0027] According to the data identifier of each metadata, an intersection between multiple data entities is established as a data relationship between the data entities;
[0028] Load data entities into the nodes of the knowledge graph and load data relationships into the edges of the knowledge graph.
[0029] Furthermore, generating a case information chain according to the data entities and data relationships stored in the knowledge graph specifically includes the following steps:
[0030] Select the target data identifier corresponding to the target case, and filter the target data entity with the target data identifier in the knowledge graph as the basic link of the case information chain;
[0031] Mining the target data identifiers to indicate the connection between the basic links of the case information chain;
[0032] The case information chain is expanded according to the data relationship of each target data entity.
[0033] Furthermore, it also includes a case information query module; the case information query module is used to query the data entities and data relationships recorded in the knowledge graph according to the input query conditions, and display them as query results.
[0034] The second aspect of the present invention provides a method for constructing a case knowledge graph based on big data technology, comprising the following steps:
[0035] Collecting case information, wherein the case information includes structured data, semi-structured data, and unstructured data;
[0036] Structuring the semi-structured data and the unstructured data so as to convert the semi-structured data and the unstructured data into structured data; the structured data includes multiple metadata, and the metadata carries one or more data identifiers; the metadata is used to represent the attributes of various elements of the case;
[0037] Extracting information from the structured data, extracting data entities in the form of metadata from the structured data, establishing data relationships between multiple data entities, and loading the data entities and data relationships into a knowledge graph;
[0038] Generate a case information chain based on the data entities and data relationships stored in the knowledge graph.
[0039] The embodiments of the present invention have the following beneficial effects: the present invention takes into account the multi-source heterogeneity of case data sources, and provides a case knowledge graph construction platform and method based on big data technology, which can extract case information including unstructured data and semi-structured data, load it into the knowledge graph, and realize the production, improvement and expansion of the case information chain through dynamic analysis of the knowledge graph.
[0040] Additional aspects and advantages of the present invention will be given in part in the following description and, in part, will be obvious from the following description or learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0042] Figure 1 It is a basic structural diagram of a case knowledge graph construction platform and method based on big data technology of the present invention;
[0043] Figure 2 This is a data processing flow diagram of a case knowledge graph construction platform and method based on big data technology in the present invention.
[0044] Figure 3 It is a schematic diagram of semi-structured data conversion in a case knowledge graph construction platform and method based on big data technology of the present invention;
[0045] Figure 4 It is a schematic diagram of unstructured data conversion in a case knowledge graph construction platform and method based on big data technology of the present invention;
[0046] Figure 5 This is a schematic diagram of generating a case information chain in a case knowledge graph construction platform and method based on big data technology in the present invention.
[0047] Figure 6 It is an example of a case information chain of a case knowledge graph construction platform and method based on big data technology in the present invention. DETAILED DESCRIPTION
[0048] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0049] The embodiment of the present invention provides a case knowledge graph construction system based on big data technology, which is used to solve the problems existing in the cases in the above background. Figure 1 As shown in the figure, the data processing flow is summarized as follows Figure 2 As shown, it specifically includes the following modules:
[0050] Case data source collection module: The case data source processing module is used to collect case information, which includes structured data, semi-structured data and unstructured data. In this embodiment, after the case occurs, the unit terminals such as the Intelligent Investigation Center will collect the case clues mined by the various assisting case handling units and generate a large amount of case-related information data, including but not including the content of the case event, the time and place of occurrence, relevant personnel, surveillance videos, recordings, transcripts, photos and other information. According to the information type classification, the case information can be divided into three categories: structured data, semi-structured data and unstructured data. Among them, structured data means that it contains metadata that can be read by the case knowledge graph construction system of this embodiment, and conforms to the data processing model format predetermined by the case knowledge graph construction system of this embodiment, that is, case data that can be directly used by the case knowledge graph construction system of this embodiment, such as process information and login identity information in the case data. Since these information are case information directly entered on the unit terminal of the case handling unit, they can be directly used by the case knowledge graph construction system of this embodiment without format conversion.
[0051] Semi-structured data refers to data that contains metadata but does not conform to the predetermined data processing model format. It needs to be converted into structured data before it can be used by the case knowledge graph construction system of this embodiment. For example, text information, table information, etc. in case data are usually case information entered or copied on other terminals. Therefore, although the semi-structured data includes data that can be read by the case knowledge graph construction system of this embodiment, the data format is the data format selected when entering data on other terminals, which is not compatible with the case knowledge graph construction system of this embodiment. Therefore, it is necessary to strip the metadata in the semi-structured data from its original data format, and then fill in the metadata based on the data processing model format predetermined by the case knowledge graph construction system of this embodiment, that is, convert the semi-structured data into structured data before it can be used by the case knowledge graph construction system of this embodiment.
[0052] Unstructured data refers to case information that does not directly contain metadata, and needs to be identified from the metadata through data processing means before it can be used by the case knowledge graph construction system of this embodiment. Such as image information and video information in case data. This information cannot be directly read by the case knowledge graph construction system of this embodiment, so it is necessary to use targeted data recognition methods to identify these unstructured data. After obtaining readable data, structured data can be further generated and used by the case knowledge graph construction system of this embodiment. The structured conversion of unstructured data is more complicated than semi-structured data. For example, video information may come from surveillance videos, driving recorders, or some video promotional materials. The scene information contained therein is diverse, and these scene information is composed of various storyboard information; the information required by the case knowledge graph construction system of this embodiment is the event information contained in the image information, so it is necessary to input the scene information into the system as streaming data, and then extract the event information layer by layer from the image information.
[0053] Case processing module: The case processing module is used to perform structured processing on semi-structured data and unstructured data, so that semi-structured data and unstructured data are converted into structured data.
[0054] In order to make the case information collected by the case data source collection module usable, this embodiment uses the case processing module to perform structured processing on semi-structured data and unstructured data, so that the semi-structured data and unstructured data are converted into structured data. The structured processing flow of semi-structured data is as follows: Figure 3 As shown in the figure, the structured processing flow of unstructured data is as follows Figure 4 shown.
[0055] Exemplarily, in this embodiment, the semi-structured data and the unstructured data are structured so that the semi-structured data and the unstructured data are converted into structured data, which specifically includes the following steps:
[0056] S101. Identify semi-structured data and unstructured data to obtain multiple metadata;
[0057] S102. Add a data identifier for each metadata;
[0058] S103. Add reference points to metadata according to the added data identifier, and the reference points are used to locate the location of the metadata;
[0059] S104. Extracting metadata based on reference points and classifying metadata according to data identifiers;
[0060] S105. Place the metadata into multiple data templates classified based on data identifiers to obtain structured data sets classified by data identifiers.
[0061] In this embodiment, the case processing module mainly uses the OEM model (object exchange model) to convert semi-structured data. In the OEM model, each object can be represented by a four-tuple (title, label, type, value). The OEM model first finds some features of semi-structured data, such as keywords, through semantic segmentation. Text information is usually in the form of "confession 1:..." In this regard, the metadata representing the text information can be used as the data identifier of the target metadata, the metadata is marked as "confession", and a reference point is added to the ":" in the data to locate it, and then the metadata under the keyword is found and read. For table information, the metadata in the header is used as the data identifier of the target metadata. Metadata is used to represent the attributes of various elements of the case, including case type, time, location, process, cause, personnel, evidence, etc. It is obvious that metadata may not be limited to one layer of structure, that is, there may be metadata under the metadata, so the same method can be used to read keywords, and finally all the case information under the keywords can be read out. In this way, semi-structured data is extracted according to the OEM model and stored in different data templates set by different data identifiers, that is, it has the data processing model format required by the case knowledge graph construction system, forming a structured data set.
[0062] For unstructured data, since the unstructured data mainly involved in case information is image information and video information, the case processing module sets different data conversion methods for image information and video information respectively. For image information, OCR (Optical Character Recognition) is mainly used. The pixel shape is determined by detecting the edge of the pixel in the image, and then the character information contained in the image is converted into metadata by character recognition. After further adding data tags and reference points to the metadata, the metadata is extracted based on the reference points, and the metadata is classified according to the data identifier. The metadata is placed in multiple data templates classified based on the data identifier to obtain a structured data set classified by the data identifier.
[0063] For image information, since the basic information unit of image information is a video frame, that is, a static image, the image information is formed by playing a temporally continuous frame sequence at equal intervals. Combining multiple mutually continuous frame sequences forms storyboard information, which generally represents a sequence of image video frames, and the video frames have content similarities. Scene information represents multiple semantically related storyboard information, such as storyboard information shot from different angles and different techniques for the same scene, which constitutes a scene information. Event information refers to a complete event or story information, which can usually characterize the case itself, including the composition relationship of multiple scene information and a general description of the video. The case processing module of this embodiment first divides the image information into multiple video frames and reads the frame information of each video frame; secondly, the storyboard information is restored based on the adjacent multiple frame information, and the scene information is restored based on the adjacent multiple storyboard information; finally, the event information in the image is restored based on the multiple scene information, and the event information is output as metadata. After further adding data tags and reference points to the metadata, the metadata is extracted based on the reference points, the metadata is classified according to the data identifiers, and the metadata is placed in multiple data templates classified based on the data identifiers to obtain a structured data set classified by the data identifiers.
[0064] In some embodiments, after the case processing module completes the structured conversion, it also includes a step of unifying the format of the metadata. Since the extracted structured data may have problems such as inconsistent length and format, taking time as an example, there may be multiple representations such as 2000 / 1 / 1, 2000-1-1, 01-01-2000, etc., it is necessary to unify the format of the structured data through a preset format template to improve the efficiency of information extraction.
[0065] Case information extraction module: The case information extraction module of this embodiment is used to extract information from structured data, extract data entities in the form of metadata from the structured data, establish data relationships between multiple data entities, and load the data entities and data relationships into the knowledge graph.
[0066] The case information extraction module of this embodiment extracts information from structured data by specifically comprising the following steps:
[0067] S201. Using a named entity recognition tool to identify structured data, extracting metadata therein as data entities;
[0068] S202. Establishing the intersection of multiple data entities according to the data identifier of each metadata as the data relationship between the data entities;
[0069] S203. Load data entities into the nodes of the knowledge graph, and load data relationships into the edges of the knowledge graph.
[0070] In this embodiment, for each structured data, the case information extraction module first uses a named entity recognition tool to identify the metadata therein, including data entities such as case type, time, location, process, cause, personnel, and evidence. For data entities composed of multiple words, multiple data entities are connected by edges in the knowledge graph to form an entity graph; in addition, the intersection between multiple data entities is established based on the data identifier of each metadata, and the intersection represents the knowledge set in the case information. Finally, a knowledge graph is obtained with all metadata in the case information as data entities and the relationships between entities represented by data identifiers as edges.
[0071] Case information chain generation module: The case information chain generation module of this embodiment is used to generate a case information chain based on the data entities and data relationships stored in the knowledge graph. Figure 5 As shown, generating a case information chain can, on the one hand, trace back to the source based on the case information chain to open up the case situation, and on the other hand, it can combine multiple aspects of case information to lock in key targets. The case information chain generation module of this embodiment generates a case information chain specifically including the following steps:
[0072] S301. Select a target data identifier corresponding to the target case, and select a target data entity with the target data identifier in the knowledge graph as a basic link of the case information chain;
[0073] S302. Mining the target data identifier to indicate the connection relationship between the basic links of the case information chain;
[0074] S303. Expand the case information chain according to the data relationship of each target data entity.
[0075] Figure 6 This is an example of expanding case handling based on information chain and locking in key case targets. Starting from case A itself, the case type, time, location, process, cause, personnel, evidence and other metadata related to case A are searched. When a breakthrough point is found in the cause of the association with case A, metadata such as related personnel, location, case details, etc. are further searched based on the cause of the association. After repeated screening, the new factual information related to case A is finally determined to help solve case A. This embodiment can mine the direct relationship between the case and the event and personnel by generating a case information chain, and use the centrality algorithm and other methods to find key personnel objects so that the staff can conduct effective screening.
[0076] In some embodiments, the knowledge graph construction system also includes a case information query module; the case information query module is used to query the data entities and data relationships recorded in the knowledge graph according to the input query conditions, and display them as query results.
[0077] This embodiment provides a method for constructing a case knowledge graph based on big data technology, including the following steps:
[0078] S301. Collect case information, which includes structured data, semi-structured data and unstructured data;
[0079] S302. Structuring the semi-structured data and the unstructured data so as to convert the semi-structured data and the unstructured data into structured data; the structured data contains multiple metadata, and the metadata contains one or more data identifiers;
[0080] S303. Extract information from structured data, extract data entities in the form of metadata from the structured data, establish data relationships between multiple data entities, and load the data entities and data relationships into the knowledge graph;
[0081] S304. Generate a case information chain based on the data entities and data relationships stored in the knowledge graph.
[0082] The embodiment of the present invention also discloses a computer program product or a computer program, wherein the computer program product or the computer program includes a computer instruction, and the computer instruction is stored in a computer-readable storage medium. A processor of a computer device can read the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes Figure 1 The method shown.
[0083] In some selectable embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the present invention is provided by way of example, for the purpose of providing a more comprehensive understanding of technology. The disclosed method is not limited to the operation and logic flow presented herein. Selectable embodiments are expected, wherein the order of various operations is changed and the sub-operation of a part for which is described as a larger operation is performed independently.
[0084] In addition, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise specified, one or more of the functions and / or features described may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the present invention. More specifically, in view of the properties, functions, and internal relationships of the various functional modules in the device disclosed herein, the actual implementation of the module will be understood within the conventional skills of the engineer. Therefore, those skilled in the art can implement the present invention set forth in the claims without excessive experimentation using ordinary techniques. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.
[0085] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.
[0086] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.
[0087] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the described embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.
Claims
1. A case knowledge graph construction platform based on big data technology, characterized by: Includes the following modules: Case data source collection module: The case data source processing module is used to collect case information, which includes structured data, semi-structured data and unstructured data; The structured data includes process information in the case information; the semi-structured data includes text information and table information in the case information; the unstructured data includes image information and video information in the case information; Case processing module: The case processing module is used to perform structured processing on semi-structured data and unstructured data, so that the semi-structured data and unstructured data are converted into structured data; The structured data includes multiple metadata, each of which carries one or more data identifiers; the metadata is used to represent the attributes of various elements of the case, including the type of case, time, location, process, cause, personnel, and evidence; The structural processing of the semi-structured data and the unstructured data so as to convert the semi-structured data and the unstructured data into structured data specifically includes the following steps: Identify semi-structured and unstructured data to obtain multiple metadata; Add data identifiers to each metadata; Adding a reference point to the metadata according to the added data identifier, wherein the reference point is used to locate the position of the metadata; Extract metadata based on reference points and classify metadata according to data identifiers; The metadata is placed into a plurality of data templates classified based on data identifiers to obtain a structured data set classified by data identifiers; Case information extraction module: The case information extraction module is used to extract information from the structured data, extract data entities in the form of metadata from the structured data, establish data relationships between multiple data entities, and load the data entities and data relationships into the knowledge graph; Case information chain generation module: The case information chain generation module is used to generate and output a case information chain based on the data entities and data relationships stored in the knowledge graph; The case information extraction module extracts data entities in the form of metadata from structured data, establishes data relationships between multiple data entities, and loads the data entities and data relationships into the knowledge graph; specifically, the following steps are included: Use named entity recognition tools to identify structured data and extract metadata as data entities; According to the data identifier of each metadata, an intersection between multiple data entities is established as a data relationship between the data entities; Load data entities into nodes of the knowledge graph and load data relationships into edges of the knowledge graph; The case information chain is generated according to the data entities and data relationships stored in the knowledge graph, and specifically includes the following steps: Select the target data identifier corresponding to the target case, and filter the target data entity with the target data identifier in the knowledge graph as the basic link of the case information chain; Mining the target data identifiers to indicate the connection between the basic links of the case information chain; The case information chain is expanded according to the data relationship of each target data entity.
2. According to the case knowledge graph construction platform based on big data technology according to claim 1, it is characterized in that: The identification of semi-structured data is specifically completed through semantic segmentation; in the data identifiers added by the metadata, for text information, the metadata referring to the text information is used as the data identifier of the target metadata; for table information, the metadata of the table header is used as the data identifier of the target metadata.
3. According to the case knowledge graph construction platform based on big data technology as described in claim 1, it is characterized in that: In the step of identifying the unstructured data and obtaining multiple metadata, for the image information, the image is scanned by using OCR to read and output the metadata in the image information; For image information, use the following steps to read the metadata in the image information: Divide the image information into multiple video frames, and read the frame information of each video frame; Restore the storyboard information according to the adjacent multiple frame information, and restore the scene information according to the adjacent multiple storyboard information; The event information in the image is restored based on multiple scene information, and the event information is output as metadata.
4. According to the case knowledge graph construction platform based on big data technology as described in claim 1, it is characterized in that: Before the step of extracting metadata based on the reference point, the following steps are also included: The metadata is unified in form according to the form template corresponding to the data identifier.
5. According to the case knowledge graph construction platform based on big data technology as described in claim 1, it is characterized in that: It also includes a case information query module; the case information query module is used to query the data entities and data relationships recorded in the knowledge graph according to the input query conditions, and display them as query results.
6. A method for constructing a case knowledge graph based on big data technology, characterized in that: The following steps are involved: Collecting case information, wherein the case information includes structured data, semi-structured data, and unstructured data; The structured data includes process information in the case information; the semi-structured data includes text information and table information in the case information; the unstructured data includes image information and video information in the case information; Structuring semi-structured data and unstructured data so as to convert semi-structured data and unstructured data into structured data; The structured data includes multiple metadata, each of which carries one or more data identifiers; the metadata is used to represent the attributes of various elements of the case, including the type of case, time, location, process, cause, personnel, and evidence; The structural processing of the semi-structured data and the unstructured data so as to convert the semi-structured data and the unstructured data into structured data specifically includes the following steps: Identify semi-structured and unstructured data to obtain multiple metadata; Add data identifiers to each metadata; Adding a reference point to the metadata according to the added data identifier, wherein the reference point is used to locate the position of the metadata; Extract metadata based on reference points and classify metadata according to data identifiers; The metadata is placed into a plurality of data templates classified based on data identifiers to obtain a structured data set classified by data identifiers; Extracting information from the structured data, extracting data entities in the form of metadata from the structured data, establishing data relationships between multiple data entities, and loading the data entities and data relationships into a knowledge graph; Generate and output case information chains based on the data entities and data relationships stored in the knowledge graph; The method of extracting data entities in the form of metadata from structured data, establishing data relationships between multiple data entities, and loading the data entities and data relationships into the knowledge graph specifically includes the following steps: Use named entity recognition tools to identify structured data and extract metadata as data entities; According to the data identifier of each metadata, an intersection between multiple data entities is established as a data relationship between the data entities; Load data entities into nodes of the knowledge graph and load data relationships into edges of the knowledge graph; The case information chain is generated according to the data entities and data relationships stored in the knowledge graph, and specifically includes the following steps: Select the target data identifier corresponding to the target case, and filter the target data entity with the target data identifier in the knowledge graph as the basic link of the case information chain; Mining the target data identifiers to indicate the connection between the basic links of the case information chain; The case information chain is expanded according to the data relationship of each target data entity.
