WO2021175009A1 - Procédé et appareil de construction de graphe d'événement d'alerte précoce, dispositif et support de stockage - Google Patents

Procédé et appareil de construction de graphe d'événement d'alerte précoce, dispositif et support de stockage Download PDF

Info

Publication number
WO2021175009A1
WO2021175009A1 PCT/CN2021/070933 CN2021070933W WO2021175009A1 WO 2021175009 A1 WO2021175009 A1 WO 2021175009A1 CN 2021070933 W CN2021070933 W CN 2021070933W WO 2021175009 A1 WO2021175009 A1 WO 2021175009A1
Authority
WO
WIPO (PCT)
Prior art keywords
event
information
event information
graph
text
Prior art date
Application number
PCT/CN2021/070933
Other languages
English (en)
Chinese (zh)
Inventor
刘康龙
徐国强
Original Assignee
深圳壹账通智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳壹账通智能科技有限公司 filed Critical 深圳壹账通智能科技有限公司
Publication of WO2021175009A1 publication Critical patent/WO2021175009A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/334—Query execution
    • G06F16/3344—Query execution using natural language analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/34—Browsing; Visualisation therefor
    • G06F16/345—Summarisation for human users
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
    • G06F16/367—Ontology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9536—Search customisation based on social or collaborative filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/01—Social networking

Definitions

  • This application relates to the field of big data technology, and in particular to a method, device, computer equipment, and storage medium for constructing an early warning event graph.
  • the event map is still a relatively new concept, which records the correlation and causality between various events.
  • each event corresponds to a different result event, and different events have a bearing on the development of the whole thing. Different influences.
  • the event map is also a kind of knowledge map.
  • the knowledge map combines the theories and methods of applied mathematics, graphics, information visualization technology, information science and other disciplines with metrological citation analysis, co-occurrence analysis and other methods, and uses the visualized image of the map. It displays the core structure, development history, frontier fields and overall knowledge structure of the discipline to achieve the modern theory of multi-disciplinary integration.
  • the inventor realizes that due to the different entity knowledge maps with concept-practical examples as the core of the event map, it is difficult to control the effect of event extraction.
  • the event extraction system is difficult to distinguish which events are important and which events are not important. There is no accurate check, which makes the quality of the constructed event map low.
  • This application provides a method for constructing an early warning event map, the method including:
  • the at least one second event information is stored in the corresponding graph database according to the hierarchical relationship to complete the construction of the event graph.
  • This application also provides a device for constructing an early warning event map, the device comprising:
  • the text acquisition module is used to obtain the text information for the event map construction when the event map construction instruction is received;
  • the first processing module is configured to perform event acquisition on the text information based on unsupervised text summarization technology to obtain at least one first event information corresponding to the text information and an early warning label corresponding to the at least one first event information;
  • the second processing module is configured to screen the at least one first event information to obtain at least one second event information for constructing an event map
  • the relationship determination module is configured to determine the hierarchical relationship between the at least one second event information according to the feature information contained in the at least one second event;
  • the graph construction module is configured to store the at least one second event information in a corresponding graph database according to the hierarchical relationship, so as to complete the construction of the event graph.
  • the application also provides a computer device, the computer device includes a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program and realizes when the computer program is executed The following steps:
  • the at least one second event information is stored in the corresponding graph database according to the hierarchical relationship to complete the construction of the event graph.
  • the present application also provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements the following steps:
  • the at least one second event information is stored in the corresponding graph database according to the hierarchical relationship to complete the construction of the event graph.
  • FIG. 1 is a schematic flowchart of a method for constructing an early warning event map in an embodiment of the application
  • FIG. 2 is a schematic flowchart of a step of obtaining second event information in an embodiment of this application
  • FIG. 3 is a schematic flow chart of the steps of processing the first event information after time stamping in an embodiment of the application
  • Fig. 4 is a schematic block diagram of a device for constructing an early warning event graph in an embodiment of the application
  • Fig. 5 is a schematic block diagram of the structure of a computer device in an embodiment.
  • FIG. 1 is a schematic flowchart of a method for constructing an early warning event map in an embodiment of the application.
  • the method includes:
  • Step S101 When receiving the event atlas construction instruction, obtain the text information for the event atlas construction.
  • the obtained text information is a related record of events that occurred in the past, and the event information recorded in the text information is correspondingly recorded. Processing, so that when the same or similar events occur, more reasonable methods can be obtained to obtain better results.
  • the input text information for constructing the event atlas will be received, so as to construct the event atlas according to the received text information.
  • the text information is related to the scene constructed based on the event map.
  • the event map is constructed so that when a risk forecast is needed, the constructed event map and the current events are used to predict the corresponding results.
  • relevant personnel can be reminded in time, and the occurrence of risks can be avoided through corresponding handling operations.
  • risk early warning such as risk early warning of financial events, through the analysis and processing of past financial events, to determine the trigger factors of financial events and how to avoid the generation of trigger factors, etc., when performing a certain operation , How to produce positive results more effectively.
  • Another example is the prediction of the outcome of an event. For an event, the occurrence of a small event will have an impact on the occurrence of the entire event. Therefore, the occurrence of bad things can be avoided through risk warning.
  • the method of obtaining the text information for constructing the event atlas is not limited, and may specifically include: receiving the uploaded text information for constructing the event atlas when the event atlas construction instruction is received; or, when receiving the text information for constructing the event atlas; When the instruction to construct the event map is reached, the input text link is read to obtain the text information for constructing the event map according to the text link.
  • the source of the text information for constructing the event graph can be diversified.
  • the user can upload the corresponding text information, and can also obtain the corresponding text information on the Internet or other platforms, which can be specifically constructed according to the required event graph.
  • the attribute of, that is, the obtained text information is related to the topic information of the currently constructed event graph.
  • Step S102 Perform event acquisition on the text information, and obtain an early warning label corresponding to the at least one first event information level of the at least one first event information corresponding to the text information.
  • the event information After obtaining the text information for the construction of the event graph, the event information will be obtained.
  • the unsupervised text summarization technology can be used to obtain the event information contained in the text information, so as to obtain a certain amount of the first number corresponding to the text information.
  • One event information, and the different early warning labels corresponding to different event information are determined according to the results corresponding to different event information.
  • the first event information is only used for distinguishing from other event information, which means that it is in different processing stages.
  • the warning labels are divided into different levels, and different numbers or texts are used for description.
  • the text information when the text information is processed to obtain the corresponding first event information, it includes: performing word segmentation processing on the text information to obtain a number of word units; Several word units are processed to obtain at least one first event information corresponding to the text information and an early warning label corresponding to the at least one first event information.
  • the unsupervised text summarization technology After receiving the text information for constructing the event graph, the unsupervised text summarization technology will be used to extract the event information from the text information to obtain the corresponding first event information.
  • the TextRank algorithm is used to process the text information, and the text information is divided into several word units, where the word units include words and sentences, and then the corresponding graph model is constructed, and based on the graph model The obtained several word units are sorted to obtain the more important word units in the text information, so as to obtain the first event information corresponding to the text information.
  • the TextRank algorithm When using the TextRank algorithm to process text information, it is used to obtain keyword information and/or keyword group information contained in the text information.
  • the graph model constructed is used to record information and count the occurrence frequency of word units. Filter out word units with extremely low frequency or below a certain threshold. Only when the frequency of occurrence of the word unit reaches a certain requirement, the corresponding word unit is saved, and the text information is obtained according to the relationship between the word units Corresponding several first event information.
  • the weight value of each word unit can also be calculated, and then the effective word unit in the text information is determined by the weight value, and then according to the effective word unit and the relationship between the word units The relationship obtains corresponding first event information.
  • Step S103 Screen the at least one piece of first event information to obtain at least one piece of second event information for building an event map.
  • the first event information will be processed twice to obtain the second event information for constructing the event graph.
  • the event information contained in the obtained first event information may have events such as incomplete or repeated events, etc., so when you get After the first event information, the first event information needs to be processed twice to obtain the second event information, so that the obtained second event information can accurately construct the event map.
  • Step S104 Determine the hierarchical relationship between the at least one second time information according to the feature information included in the at least one second event.
  • the association relationship between each event information can be determined, mainly including the hierarchical relationship.
  • the feature information determines the hierarchical relationship between at least one second event.
  • each event information There will be a certain level relationship between everything.
  • the key words contained in each event information are used to determine the level relationship between events. Taking sports as an example, sports usually include football, basketball, tennis, racing, diving, and table tennis. And so on, there are also certain hierarchical relationships for events. For example, sports events include competition events, competition events, etc. By determining the hierarchical relationship or association relationship before the event, the resulting time map is more hierarchical. Obviously, make the correlation between events clearer.
  • Step S105 Store the at least one second event information in a corresponding graph database according to the hierarchical relationship, so as to complete the construction of the event graph.
  • the second event information is stored in the corresponding graph database to complete the construction of the event graph for risk warning.
  • Neo4j commonly used graph databases include Neo4j, FlockDB, AllegroGrap, GraphDB, InfiniteGraph, etc.
  • the most advanced graph database Neo4j is used in this implementation.
  • the label of each node is determined according to the relationship between the event information, and the attribute information associated with the event information will also be recorded on each node, such as Event information, time information, event category information, etc.
  • the method when using the second event information to complete the construction of the event graph, includes: obtaining a level label corresponding to the second event information, so as to determine the storage level of the second event according to the level label;
  • the second event information is stored according to the storage hierarchy, and an association relationship between the second event information is established to complete the construction of the event map.
  • Levels correspond to nodes. If the hierarchical relationship between events has 3 levels, then these 3 levels correspond to a node respectively. Use the feature description of the event information and the level label to mark the node to obtain the corresponding label, where the different level labels belong
  • the level information is different.
  • the first-level event information is sports
  • the second-level event information is sports
  • the third-level event information is football
  • the level is football.
  • the corresponding 3 nodes are the first node is sports, the second node is sports, and the third node is football.
  • the event graph there are several levels of hierarchical relationships between event information, so there will be several levels of node relationships.
  • an event map for risk early warning when an event map for risk early warning needs to be constructed, the inputted event map construction instruction is received, and the text information for the event map construction is obtained at the same time, and then the unsupervised text summarization technology is used Obtain the event information from the acquired text information to obtain the corresponding first event information, and then process the first event information, including filtering and de-duplication, to obtain the second event information for the construction of the event map, and finally the first event information Second, the event information is stored in the corresponding graph database to complete the construction of the event graph. It is realized that when constructing the event map, the purity of the event map is improved, the noise data in the obtained event map is reduced, and the quality of the event map is improved.
  • FIG. 2 is a schematic flowchart of the step of obtaining second event information in an embodiment of this application.
  • step S103 processing the first text information to obtain second event information for constructing an event graph, includes:
  • Step S201 Acquire time information corresponding to the first event information, so as to use the time information to time stamp the first event information.
  • Step S202 Screen the first event information that is time-stamped to obtain second event information that is to be constructed for the event atlas.
  • the first event information When processing the first event information, first complete the first event information, and then process the event information after the completion of the information. In the evening of information, it is not to complete the completeness of the event. For example, if the information recorded in a certain event information is incomplete, the event information will not be improved at this time. Specifically, after the first event information is obtained , According to the related information of the event information or the related information of the text information, the time information of the obtained first event information will be perfected, and the event occurrence time corresponding to the first event information will be accurately determined.
  • the text information after obtaining the corresponding event information contained in the text information, it is necessary to accurately determine the time information corresponding to the event information, and then associate the time information with the corresponding time information to obtain the event information containing the time information.
  • time information of the event there are many ways to express the time information of the event. You can use specific time, such as August 8, 2008, etc. to accurately represent the description of the time. You can also use words that contain time information, such as yesterday. , Today and a few days ago can indirectly represent the description of time.
  • when determining the time information corresponding to the event information it includes: when the event information contains time information, obtaining the time information contained in the event information, and comparing the obtained time information with the event information Correlation; when the event information does not contain time information, obtain the recording time information of the text information corresponding to the event information, and the feature words whose word attributes belong to the time attribute in the event information, to determine according to the obtained recording time information and feature words The time information corresponding to the event information. Furthermore, the obtained time information is correlated with the event information to obtain the event information containing the time information.
  • time information is divided into direct time description and indirect time description (ie, indirect time description).
  • the time information is divided into direct time information and indirect time information.
  • direct time description such as August 8, 2008
  • the obtained direct time information will be used directly.
  • the sexual time description marks the event information.
  • corresponding processing will be performed to obtain accurate time information of the event information. For example, for today’s event information, if the event information is “Due to the development needs of 5G, the stock of company A rose sharply yesterday”, and today is October 10, 2020, then the event information corresponding to this event is October 2020 9th.
  • the event information After obtaining the first event information marked by the time, secondary processing will be performed to obtain the second event information for constructing the event graph.
  • the event information When processing the first event information after time stamping, the event information will be filtered and deduplicated, the event information with incomplete event records will be filtered out, and repeated events will be deduplicated to obtain a complete and complete event record. There is no duplicate second event information. By filtering event information to remove duplicates, and deleting and filtering useless events and repeated events in the event information, the purity of the event map construction can be effectively improved.
  • FIG. 3 is a schematic flowchart of the steps of processing the first event information after time stamping in an embodiment of the application.
  • step S202 screening the time-stamped first event information to obtain the second event information for building the event map, includes:
  • Step S301 Perform sparse event removal on the first event information for event marking according to the sparse event removal technology, where the sparse event is an event with incomplete information composition.
  • Step S302 Classify the first event information after the sparse event removal is performed to obtain several categories.
  • Step S303 Perform de-duplication processing on the first event information contained in the several categories according to the entity fusion technology, so as to obtain the second event information for constructing the event graph.
  • the event information containing time information After the event information containing time information is obtained, the event information containing time information will be processed for event removal, and useless event information or event information with incomplete information records in the event information will be deleted to obtain useful events.
  • Information mainly using sparse event removal technology to filter event information after time stamp processing to achieve event removal.
  • screening event information After all event information is obtained based on big data, the keywords contained in the event information are obtained, and the number of keywords corresponding to each event information is recorded, so as to compare the event information according to the number of keywords. Perform classification statistics to obtain the number of events corresponding to different numbers of keywords, and finally remove part of the data according to the large difference in the number of events. For example, when the difference between the number of events corresponding to one keyword and the number of events corresponding to two keywords is the largest, it is determined that the event information corresponding to one keyword is removed as a sparse event.
  • the event information obtained from the processing is deduplicated according to the entity fusion technology, where the entity fusion technology mainly applies the entity fusion and normalization method to combine the event information
  • the event information of the same event recorded in the event information is summarized and normalized, that is, multiple event information that records the same event is deduplicated to avoid the existence of multiple event information that records the same event in the event library.
  • the pre-trained logistic regression model can also be used to judge the event information, determine whether there is repeated event information between the event information, and delete the repeated event information to achieve the de-duplication effect.
  • the process of using entity fusion and normalization for deduplication includes data grouping, data preprocessing, attribute similarity and entity similarity.
  • data grouping In order to find all the same entities, if data grouping is not performed, then the calculation amount at this time will be two-by-two comparison. For massive data, the calculation amount is too large, so grouping must be performed in advance. For each type of data, entity fusion and normalization will be carried out.
  • grouping event information the event information with the same keywords corresponding to the event information is divided into the same group or category. After the grouping or classification is completed, each group of event information is deduplicated, and each group is divided. The repeated event information in the file is filtered and deleted, and a single event information is obtained.
  • Data preprocessing the input original data source often has dirty data and inconsistent data. Through preprocessing and regularization, this step is a time-consuming but very useful work in actual engineering. If there is no good data processing, the subsequent algorithm effects are often It won't be good. For example, like the name of a hospital, some sites are directly xx hospitals, and some sites will have "xx hospitals second-level first-class", "xx hospitals medical insurance designated hospitals” and so on. After we get this kind of data, we will first Process this name and unify the names of each site as "xx hospital", because something similar to "second-level A" will be used as an attribute value of the hospital.
  • the keywords corresponding to the obtained event information are unified.
  • the keywords are unified with the query in the preset list to determine the main keywords corresponding to the keywords, and then Use the obtained main keyword to replace the original keyword.
  • the main keywords corresponding to the above-described "xx hospital”, “xx hospital second-level A”, and "xx hospital medical insurance designated hospital” are "xx hospital”.
  • the attribute similarity includes: pure string: calculate edit distance, levenshtein distance, calculate string A to be transformed into string B by insert/delete/replace operation Distance; Collection type: Calculate Jaccard similarity, calculate the number of collection intersections/collection unions; Document type: Find the keywords of each document through tf-idf, and then calculate the similarity of the keyword collections through cosine similarity , Specifically obtain the keywords in the event information, and then perform the cosine similarity calculation to obtain the corresponding similarity.
  • Entity similarity With the similarity of each attribute of the entity, the entity similarity can be calculated. There are two common methods: regression: directly judge the similarity of the entity through the similarity of each attribute of the entity. The weight of the similarity of each attribute can be directly scored, or the weight of the similarity of each attribute can be calculated by means of logistic regression.
  • Clustering Calculate similar entities directly through clustering operations. Hierarchical clustering, correlation clustering, Canopy+K-means clustering, etc. can be performed.
  • FIG. 4 is a schematic block diagram of a device for constructing an early warning event map in an embodiment of the application, and the device is used to execute the aforementioned method for constructing an early warning event map.
  • the device 400 for constructing an early warning event map includes:
  • the text obtaining module 401 is configured to obtain text information for constructing an event graph when an instruction for constructing an event graph is received;
  • the first processing module 402 is configured to perform event acquisition on the text information to obtain at least one first event information corresponding to the text information and an early warning label corresponding to the at least one first event information;
  • the second processing module 403 is configured to screen the at least one piece of first event information to obtain at least one piece of second event information for building an event graph;
  • the relationship determination module 404 is configured to determine the hierarchical relationship between the at least one second event information according to the feature information contained in the at least one second event;
  • the graph construction module 405 is configured to store the at least one second event information in a corresponding graph database according to the hierarchical relationship to complete the construction of the event graph.
  • the text acquisition module 401 is specifically used to:
  • the first processing module 402 is specifically further configured to:
  • the first processing module 402 is specifically further configured to:
  • Construct a graph model take the several word units as the input of the graph model to obtain the weight value corresponding to each word unit in the several word units; obtain the word units used to form the event information according to the weight value to obtain At least one first event information corresponding to the text information and an early warning label corresponding to the at least one first event information.
  • the second processing module 403 is specifically used to:
  • the second processing module 403 is specifically used to:
  • map building module 405 is specifically used to:
  • the above-mentioned device can be implemented in the form of a computer program, and the computer program can be run on the computer device as shown in FIG. 6.
  • FIG. 5 is a schematic block diagram of the structure of a computer device in an embodiment.
  • the computer device may be a server.
  • the computer device includes a processor, a memory, and a network interface connected through a system bus, where the memory may be volatile or non-volatile.
  • the non-volatile storage medium can store an operating system and a computer program.
  • the computer program includes program instructions, and when the program instructions are executed, the processor can execute any method for constructing a risk early warning event map.
  • the processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.
  • the internal memory provides an environment for the operation of the computer program in the non-volatile storage medium.
  • the processor can execute any method for constructing an early warning event graph.
  • the network interface is used for network communication, such as sending assigned tasks.
  • the network interface is used for network communication, such as sending assigned tasks.
  • FIG. 5 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied.
  • the specific computer device may Including more or fewer parts than shown in the figure, or combining some parts, or having a different arrangement of parts.
  • the processor may be a central processing unit (Central Processing Unit, CPU), and the processor may also be other general-purpose processors or digital signal processors (Digital Signal Processors). Processor, DSP), application specific integrated circuit (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor.
  • the processor is used to run a computer program stored in a memory to implement the following steps:
  • the text information for the event atlas construction When receiving the event atlas construction instruction, obtain the text information for the event atlas construction; perform event acquisition on the text information to obtain at least one first event information corresponding to the text information and the at least one first event information corresponding The early warning label of the; the at least one first event information is screened to obtain at least one second event information for building an event map; the at least one second event information is determined according to the characteristic information contained in the at least one second event The hierarchical relationship between event information; storing the at least one second event information in a corresponding graph database according to the hierarchical relationship to complete the construction of the event graph.
  • the processor is also used to implement:
  • the processor is also used to implement:
  • the processor is also used to implement:
  • Construct a graph model take the several word units as the input of the graph model to obtain the weight value corresponding to each word unit in the several word units; obtain the word units used to form the event information according to the weight value to obtain At least one first event information corresponding to the text information and an early warning label corresponding to the at least one first event information.
  • the processor is also used to implement:
  • the processor is also used to implement:
  • the processor is also used to implement:
  • the embodiments of the present application also provide a computer-readable storage medium, and the computer-readable storage medium may be volatile or non-volatile.
  • the computer-readable storage medium stores a computer program, the computer program includes program instructions, and the processor executes the program instructions to implement any one of the methods for constructing an early warning event graph provided in the embodiments of the present application.
  • the computer-readable storage medium may be the internal storage unit of the computer device described in the foregoing embodiment, for example, the hard disk or memory of the computer device.
  • the computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), and a secure digital (Secure Digital, SD) equipped on the computer device. ) Card, Flash Card, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Computational Linguistics (AREA)
  • Animal Behavior & Ethology (AREA)
  • Human Resources & Organizations (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Economics (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

L'invention concerne un procédé et un appareil de construction de graphe d'événement d'alerte précoce, ainsi qu'un dispositif et un support de stockage, qui se rapportent au domaine des mégadonnées. Le procédé consiste à : obtenir des informations textuelles pour une construction de graphe d'événements lorsqu'une instruction de construction de graphe d'événements est reçue (S101) ; effectuer une acquisition d'événement sur les informations textuelles pour obtenir au moins un élément de premières informations d'événement correspondant aux informations textuelles et une étiquette d'alerte précoce correspondant à l'élément ou aux éléments de premières informations d'événement (S102) ; filtrer l'élément ou les éléments de premières informations d'événement afin d'obtenir au moins un élément de secondes informations d'événement pour une construction de graphe d'événement (S103) ; déterminer une relation hiérarchique entre l'élément ou les éléments de secondes informations d'événement en fonction des informations de caractéristiques comprises dans l'élément ou les éléments du second événement (S104) ; et stocker l'élément ou les éléments de secondes informations d'événement dans une base de données de graphe correspondante en fonction de la relation hiérarchique de façon à réaliser la construction d'un graphe d'événement (S105). Le procédé améliore la qualité d'un graphe d'événement lorsque le graphe d'événement est construit.
PCT/CN2021/070933 2020-03-02 2021-01-08 Procédé et appareil de construction de graphe d'événement d'alerte précoce, dispositif et support de stockage WO2021175009A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010136834.0A CN111475612A (zh) 2020-03-02 2020-03-02 预警事件图谱的构建方法、装置、设备及存储介质
CN202010136834.0 2020-03-02

Publications (1)

Publication Number Publication Date
WO2021175009A1 true WO2021175009A1 (fr) 2021-09-10

Family

ID=71747120

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/070933 WO2021175009A1 (fr) 2020-03-02 2021-01-08 Procédé et appareil de construction de graphe d'événement d'alerte précoce, dispositif et support de stockage

Country Status (2)

Country Link
CN (1) CN111475612A (fr)
WO (1) WO2021175009A1 (fr)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113392220A (zh) * 2020-10-23 2021-09-14 腾讯科技（深圳）有限公司 一种知识图谱生成方法、装置、计算机设备及存储介质
CN114647743A (zh) * 2022-05-20 2022-06-21 国网浙江省电力有限公司 电力营销全业务门禁规则图谱生成及处理方法、装置
CN114707004A (zh) * 2022-05-24 2022-07-05 国网浙江省电力有限公司信息通信分公司 基于图像模型和语言模型的事理关系抽取处理方法及系统
CN114880588A (zh) * 2022-06-13 2022-08-09 四川封面传媒科技有限责任公司 基于知识图谱的新闻热度预测方法
CN115146081A (zh) * 2022-08-31 2022-10-04 合肥中科迪宏自动化有限公司 生产设备的故障诊断知识图谱的构建方法及诊断方法
CN115630846A (zh) * 2022-12-07 2023-01-20 速度时空信息科技股份有限公司 适用于自然灾害风险综合监测数据的处理方法
CN116054910A (zh) * 2022-12-20 2023-05-02 中国人民解放军63819部队 基于知识图谱构建的地球站设备故障分析及装置
CN116842200A (zh) * 2023-03-29 2023-10-03 全景智联(武汉)科技有限公司 一种事件聚档管理方法

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111475612A (zh) * 2020-03-02 2020-07-31 深圳壹账通智能科技有限公司 预警事件图谱的构建方法、装置、设备及存储介质
CN114722801A (zh) * 2020-12-22 2022-07-08 航天信息股份有限公司 政务数据分类存储方法及相关装置
CN112749239B (zh) * 2021-01-20 2022-10-21 青岛海信网络科技股份有限公司 一种事件图谱构建方法、装置及计算设备
CN113449116B (zh) * 2021-06-22 2022-12-20 青岛海信网络科技股份有限公司 一种图谱构建、预警方法、设备及介质
CN114706992B (zh) * 2022-02-17 2022-09-30 中科雨辰科技有限公司 一种基于知识图谱的事件信息处理系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107783973A (zh) * 2016-08-24 2018-03-09 慧科讯业有限公司 基于行业知识图谱数据库对互联网媒体事件进行监测的方法、装置和系统
CN108763333A (zh) * 2018-05-11 2018-11-06 北京航空航天大学 一种基于社会媒体的事件图谱构建方法
WO2018209254A1 (fr) * 2017-05-11 2018-11-15 Hubspot, Inc. Procédés et systèmes de génération automatisée de messages personnalisés
CN109977237A (zh) * 2019-05-27 2019-07-05 南京擎盾信息科技有限公司 一种面向法律领域的动态法律事件图谱构建方法
CN110781317A (zh) * 2019-10-29 2020-02-11 北京明略软件系统有限公司 事件图谱的构建方法、装置及电子设备
CN111475612A (zh) * 2020-03-02 2020-07-31 深圳壹账通智能科技有限公司 预警事件图谱的构建方法、装置、设备及存储介质

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107783973A (zh) * 2016-08-24 2018-03-09 慧科讯业有限公司 基于行业知识图谱数据库对互联网媒体事件进行监测的方法、装置和系统
WO2018209254A1 (fr) * 2017-05-11 2018-11-15 Hubspot, Inc. Procédés et systèmes de génération automatisée de messages personnalisés
CN108763333A (zh) * 2018-05-11 2018-11-06 北京航空航天大学 一种基于社会媒体的事件图谱构建方法
CN109977237A (zh) * 2019-05-27 2019-07-05 南京擎盾信息科技有限公司 一种面向法律领域的动态法律事件图谱构建方法
CN110781317A (zh) * 2019-10-29 2020-02-11 北京明略软件系统有限公司 事件图谱的构建方法、装置及电子设备
CN111475612A (zh) * 2020-03-02 2020-07-31 深圳壹账通智能科技有限公司 预警事件图谱的构建方法、装置、设备及存储介质

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113392220B (zh) * 2020-10-23 2024-03-26 腾讯科技（深圳）有限公司 一种知识图谱生成方法、装置、计算机设备及存储介质
CN113392220A (zh) * 2020-10-23 2021-09-14 腾讯科技（深圳）有限公司 一种知识图谱生成方法、装置、计算机设备及存储介质
CN114647743A (zh) * 2022-05-20 2022-06-21 国网浙江省电力有限公司 电力营销全业务门禁规则图谱生成及处理方法、装置
CN114707004A (zh) * 2022-05-24 2022-07-05 国网浙江省电力有限公司信息通信分公司 基于图像模型和语言模型的事理关系抽取处理方法及系统
CN114707004B (zh) * 2022-05-24 2022-08-16 国网浙江省电力有限公司信息通信分公司 基于图像模型和语言模型的事理关系抽取处理方法及系统
CN114880588A (zh) * 2022-06-13 2022-08-09 四川封面传媒科技有限责任公司 基于知识图谱的新闻热度预测方法
CN114880588B (zh) * 2022-06-13 2024-04-26 四川封面传媒科技有限责任公司 基于知识图谱的新闻热度预测方法
CN115146081B (zh) * 2022-08-31 2022-12-09 合肥中科迪宏自动化有限公司 生产设备的故障诊断知识图谱的构建方法及诊断方法
CN115146081A (zh) * 2022-08-31 2022-10-04 合肥中科迪宏自动化有限公司 生产设备的故障诊断知识图谱的构建方法及诊断方法
CN115630846A (zh) * 2022-12-07 2023-01-20 速度时空信息科技股份有限公司 适用于自然灾害风险综合监测数据的处理方法
CN116054910A (zh) * 2022-12-20 2023-05-02 中国人民解放军63819部队 基于知识图谱构建的地球站设备故障分析及装置
CN116054910B (zh) * 2022-12-20 2024-05-14 中国人民解放军63819部队 基于知识图谱构建的地球站设备故障分析及装置
CN116842200A (zh) * 2023-03-29 2023-10-03 全景智联(武汉)科技有限公司 一种事件聚档管理方法
CN116842200B (zh) * 2023-03-29 2024-06-28 全景智联(武汉)科技有限公司 一种事件聚档管理方法

Also Published As

Publication number Publication date
CN111475612A (zh) 2020-07-31

Similar Documents

Publication Publication Date Title
WO2021175009A1 (fr) Procédé et appareil de construction de graphe d'événement d'alerte précoce, dispositif et support de stockage
US20210382878A1 (en) Systems and methods for generating a contextually and conversationally correct response to a query
US8630989B2 (en) Systems and methods for information extraction using contextual pattern discovery
WO2019174132A1 (fr) Procédé de traitement de données, serveur et support de stockage informatique
WO2021042521A1 (fr) Procédé de génération automatique de contrat, dispositif informatique et support de stockage informatique non volatil
WO2021098648A1 (fr) Procédé, appareil et dispositif de recommandation de texte, et support
US8326819B2 (en) Method and system for high performance data metatagging and data indexing using coprocessors
US9922383B2 (en) Patent claims analysis system and method
WO2019218475A1 (fr) Procédé et dispositif d'identification de sujet au comportement anormal, dispositif terminal, et support
CN108121829A (zh) 面向软件缺陷的领域知识图谱自动化构建方法
TWI682287B (zh) 知識圖譜產生裝置、方法及其電腦程式產品
US11100151B2 (en) Interactive patent visualization systems and methods
US20180018355A1 (en) Automated Monitoring and Archiving System and Method
AU2018411565B2 (en) System and methods for generating an enhanced output of relevant content to facilitate content analysis
TW201915777A (zh) 金融非結構化文本分析系統及其方法
CN113239111B (zh) 一种基于知识图谱的网络舆情可视化分析方法及系统
Bykau et al. Fine-grained controversy detection in Wikipedia
CN115358201B (zh) 一种期货领域的投研报告处理方法和系统
US20170286551A1 (en) Scalable processing of heterogeneous user-generated content
Stvilia A model for ontology quality evaluation
US20150154195A1 (en) Method for entity-driven alerts based on disambiguated features
AT&T
CN112231483A (zh) 灾情追踪方法、系统、设备和存储介质
US20240086433A1 (en) Interactive tool for determining a headnote report
US20240086448A1 (en) Detecting cited with connections in legal documents and generating records of same

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21765436

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 11.01.2023)

122 Ep: pct application non-entry in european phase

Ref document number: 21765436

Country of ref document: EP

Kind code of ref document: A1