Data processing method, device and electronic equipment

Abstract

The present application discloses a data processing method, device and electronic device, including: obtaining a target event; determining event elements having a target association relationship with the target event; determining the element association relationship between each of the event elements based on the element characteristics of each of the event elements; determining the event level of the target event based on the target association relationship and the element association relationship, so that the target event is processed based on the event level to obtain an event processing result. The present application can process the event elements corresponding to the target event, determine the relationship between each event element and the target event and the relationship between the elements, avoid the influence of subjective factors on the direct determination of the event level, and improve the accuracy of the event level determination.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and more specifically to a data processing method, device and electronic equipment. Background Art

[0002] Machine learning is the core of artificial intelligence. It is a method that can give machines the ability to complete functions that cannot be completed by direct programming. It can be applied in many aspects. For example, in machine learning tasks such as public opinion analysis of events and crisis warning, it is necessary not only to output the analysis results through the corresponding training model of machine learning, but also to analyze and predict the impact level of the event, such as the severity of product defects, high, medium and low.

[0003] In this scenario, in order to train a machine learning model that outputs event level information, it is necessary to label the event samples. Level information is usually labeled manually, but for the same event, different people may have different cognitive labels on its impact level, which will lead to inconsistent level labeling results, thus affecting the accuracy of the training samples and the accuracy of the model prediction. Summary of the invention

[0004] In view of this, this application provides the following technical solutions:

[0005] A data processing method, comprising:

[0006] Get the target event;

[0007] Determining event elements having a target association relationship with the target event;

[0008] Determining, based on the element characteristics of each of the event elements, element association relationships between the event elements;

[0009] Based on the target association relationship and the element association relationship, an event level of the target event is determined, so that the target event is processed based on the event level to obtain an event processing result.

[0010] Optionally, the determining of the event elements having a target association relationship with the target event includes:

[0011] Extracting entities constituting the target event to obtain entity information;

[0012] Event elements satisfying the target association relationship are determined in the entity information.

[0013] Optionally, the target association relationship includes one or more of a causal relationship, a dependency relationship, or a master-slave relationship.

[0014] Optionally, determining the event level of the target event based on the target association relationship and the element association relationship includes:

[0015] Determine a weight parameter of each event element based on the element association relationship;

[0016] Based on the target association relationship, determining the relationship strength probability of each event element to the target event;

[0017] The event level of the target event is determined according to the weight parameter and the relationship strength probability.

[0018] Optionally, determining a weight parameter of each event element based on the element association relationship includes:

[0019] Based on the element association relationship, clustering each of the event elements to obtain at least one category;

[0020] Based on the relationship between each event element and the category, a weight parameter of each event element is determined.

[0021] Optionally, the method further comprises:

[0022] Acquire event samples, wherein each event sample is annotated with event elements corresponding to the event, element association relationships, and event levels;

[0023] Based on the event samples, a target network graph is determined, wherein nodes without outgoing edges of the target network graph represent event levels, nodes with outgoing edges in the target network graph represent event elements, and directed edges of the target network graph represent element association relationships.

[0024] Optionally, determining the event level of the target event includes:

[0025] Based on the target network graph and the target association relationship, determining a directed subgraph corresponding to the target event;

[0026] Determine the subgraph probability corresponding to the directed subgraph as the level probability of the target event;

[0027] Based on the level probability, an event level of the target event is determined.

[0028] Optionally, it also includes:

[0029] Determining a subgraph probability of the directed subgraph based on a conditional probability of each node in the directed subgraph to a target node it points to;

[0030] The conditional probability of each node to the target node it points to is obtained by learning based on event sample data labeled with the probability of the degree of influence of event elements on events, or the conditional probability is a preset target value.

[0031] A data processing device, comprising:

[0032] An acquisition unit, used for acquiring a target event;

[0033] An element determination unit, used to determine an event element having a target association relationship with the target event;

[0034] A relationship determination unit, configured to determine an element association relationship between each of the event elements based on an element feature of each of the event elements;

[0035] The level determination unit is used to determine the event level of the target event based on the target association relationship and the element association relationship, so that the target event is processed based on the event level to obtain an event processing result.

[0036] An electronic device, comprising:

[0037] Memory, used to store programs;

[0038] The processor is used to call and execute the program in the memory, and implement the various steps of the data processing method as described in any one of the above by executing the program.

[0039] Through the above technical solutions, it can be known that the present application discloses a data processing method, device and electronic device, including: obtaining a target event; determining event elements that have a target association relationship with the target event; determining the element association relationship between each of the event elements based on the element characteristics of each of the event elements; determining the event level of the target event based on the target association relationship and the element association relationship, so that the target event is processed based on the event level to obtain an event processing result. The present application can process the event elements corresponding to the target event, determine the relationship between each event element and the target event, and the relationship between the elements, avoid the influence of subjective factors on the direct determination of the event level, and improve the accuracy of the event level determination. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] 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 merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0041] Figure 1 A flowchart of a data processing method provided in an embodiment of the present application;

[0042] Figure 2 A schematic diagram of a target network diagram provided in an embodiment of the present application;

[0043] Figure 3 A schematic diagram of an application scenario provided in an embodiment of the present application;

[0044] Figure 4 A schematic diagram of the structure of a data processing device provided in an embodiment of the present application;

[0045] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0047] See also Figure 1 , is a flow chart of a data processing method provided in an embodiment of the present application, the method may include the following steps:

[0048] S101. Obtain target event.

[0049] The target event refers to the event that needs to be processed in the application scenario, that is, the event that needs to determine the event level. For example, the target event can be a product level assessment event in the product fault diagnosis application scenario, or it can be a product production event, that is, analyzing the product production event to determine the evaluation level corresponding to the product generation event.

[0050] S102: Determine event elements that have a target association relationship with the target event.

[0051] Among them, the event elements may be elements that directly or indirectly affect the generation of the target event or the presentation effect of the target event, for example, constituent elements, evaluation elements, effect elements, etc. of the target event. The target association relationship refers to the association relationship corresponding to the determination of the target event level, and the target association relationship may include one or more of a causal relationship, a dependency relationship, or a master-slave relationship. That is, first determine all the elements corresponding to the target event, and then extract the event elements that have a target association relationship with the target event from these elements. For example, the event elements that have a target association relationship with the target event may include constituent elements of the target event, recording elements of the target event generation process, generation means elements corresponding to the target event, etc.

[0052] S103: Determine the element association relationship between the event elements based on the element feature of each event element.

[0053] S104: Determine the event level of the target event based on the target association relationship and the element association relationship, so that the target event is processed based on the event level to obtain an event processing result.

[0054] The element features of event elements refer to the attribute features, intention features, etc. of the event elements obtained by parsing the event elements. For example, when the event element is a customer influencing factor, the customer influencing factor can be identified to obtain the customer portrait features, and the customer intention features can also be obtained, which are determined as the element features of the customer influencing factor. The element association relationship between each event element refers to whether the event elements have a characteristic relationship based on the element features, such as whether it is a progressive relationship, a subordinate relationship, or a relationship belonging to the same element category.

[0055] In the embodiments of the present application, the event level of the target event is not evaluated or determined directly based on subjective experience, but the event elements corresponding to the target event are first obtained, and the event level of the target event is determined by determining the element association relationship between the various event elements, and the target association relationship corresponding to each event element and the target event. The target event is evaluated from a more fine-grained dimension, and the level determination of the target event has more objective evaluation conditions. Among them, the event level refers to the importance of the event under characteristic conditions. For example, under the conditions of fault diagnosis application, the event level can be divided into low risk level, medium risk level, high risk level, etc. In the application scenario of event level annotation, the target event can be level labeled based on the obtained event level to obtain a level annotation result.

[0056] The relationship between each event element can be determined through the element association relationship, such as whether they belong to the same category, so as to determine the influence weight parameter of the event element on the event level determination. The target association relationship can determine the relationship between the event element and the target event, so it can be determined to adopt the corresponding target determination mode to determine its influence parameter on the event level of the target event, etc.

[0057] After obtaining the event level, the target event can be processed based on the event level to obtain the event processing result. For example, in a fault diagnosis scenario, it can be determined whether to perform fault processing on the target event based on the fault risk level of the target event, such as whether to send the fault analysis information corresponding to the target event to the fault handling personnel.

[0058] Correspondingly, the data processing method provided in the embodiment of the present application can also be applied in the process of machine learning to automatically label event samples according to event levels, thereby solving the problem of the influence of subjective factors brought about by manual labeling, for example, the problem that the labeling results of the same event corresponding to different labelers are inconsistent due to the inconsistency of labeling standards. This improves the interpretability of event level labeling.

[0059] The embodiment of the present application discloses a data processing method, including: obtaining a target event; determining event elements that have a target association relationship with the target event; determining the element association relationship between each of the event elements based on the element characteristics of each of the event elements; determining the event level of the target event based on the target association relationship and the element association relationship, so that the target event is processed based on the event level to obtain an event processing result. The present application can process the event elements corresponding to the target event, determine the relationship between each event element and the target event, and the relationship between the elements, avoid the influence of subjective factors on the direct determination of the event level, and improve the accuracy of the event level determination.

[0060] In one embodiment of the present application, the event elements having a target association relationship with the target event include: extracting entities constituting the target event to obtain entity information; and determining event elements satisfying the target association relationship in the entity information. In this embodiment, event elements are determined by entity extraction. An entity can be a thing or concept that can be defined within a target event, such as an object, introduction, and processing method in a target event. All elements constituting the target event can be obtained by entity extraction. Since the amount of entity information extracted is large, not every entity will affect the determination of the event level. Therefore, entities satisfying the target association relationship are extracted as event elements. For example, in the record entity recording the target event, entities recording specific links of the target event can be extracted as event elements, while entities recording basic information of the target event, such as the time of occurrence, have a small impact on the event level of the target event and can be ignored.

[0061] Correspondingly, event elements may also include any definable factors such as the conditions, causes, phenomena, and consequences of the event. The relationships between event elements may include causality, dependency, master-slave relationship, inclusion relationship, etc. Correspondingly, event elements and element relationships between elements may be collected and collated by annotating specific instances. For example, the target event may be input into an element recognition model to obtain the corresponding event elements and element association relationships between event elements, and the element recognition model is a neural network model trained based on event training samples annotated with corresponding information about event elements and element association relationships between event elements.

[0062] In one implementation of the embodiment of the present application, determining the event level of the target event based on the target association relationship and the element association relationship includes:

[0063] Based on the element association relationship, a weight parameter of each event element is determined; based on the target association relationship, a relationship strength probability of each event element to the target event is determined; and according to the weight parameter and the relationship strength probability, an event level of the target event is determined.

[0064] The element association relationship can determine the relationship between each event element. Based on the relationship, it can be determined whether each event element belongs to the same category. If they belong to the same category, the corresponding weight parameter is determined based on the relationship with the category. Specifically, based on the element association relationship, each event element can be clustered to obtain at least one category; based on the relationship between each event element and the category, the weight parameter of each event element is determined. It is also possible to determine the degree of influence on the target event level based on the element association relationship, and determine the weight parameter of each event element based on the degree of influence. The probability of the relationship strength of each event element on the target event represents the parameter of the degree of influence of the event element on the target event, that is, the probability of generating a certain effect corresponding to the target event through the event element, and the effect is a presentation effect that matches the target event level.

[0065] In one implementation of the embodiment of the present application, the target event level can also be determined through a target network diagram. Correspondingly, the method further includes:

[0066] S201. Acquire event samples, wherein each event sample is annotated with event elements corresponding to the event, element association relationships, and event levels.

[0067] S202. Determine a target network graph based on the event sample, wherein nodes without outgoing edges of the target network represent event levels, nodes with outgoing edges of the target network represent event elements, and directed edges of the target network represent element association relationships.

[0068] In this embodiment, the target network graph can also be a belief network, which is a directed acyclic graph consisting of variable nodes and directed edges connecting these nodes, and is a graph model representing the probability of variable set connection. Among them, the node represents the abstraction of any problem, and the directed edges between nodes represent the relationship between nodes, such as dependency, causality, premise, etc. The relationship nodes are represented by conditional probabilities, and the nodes without parent nodes are expressed by prior probabilities.

[0069] Corresponding to the embodiment of the present application, a node can be an event element and an event level. The event level is a level determined by the event element, and is not necessarily the overall event level of the target event. Specifically, a node without an outgoing edge can represent an event level, a node with an outgoing edge represents an event element, and a directed edge is an element association relationship. Event elements include any definable factors such as the conditions, causes, phenomena, and consequences of an event. The relationship between nodes can be defined according to actual needs, such as cause-and-effect, dependency, master-slave, and other relationships. Nodes and relationships can be collected and sorted by annotating specific instances.

[0070] When determining the event level of the target event based on the above-mentioned target network graph, the process may include: determining the directed subgraph corresponding to the target event based on the target network graph and the target association relationship; determining the level probability of the target event using the subgraph probability corresponding to the directed subgraph; and determining the event level of the target event based on the level probability. The target network may include multiple directed subgraphs, and a directed subgraph branch corresponding to the target association relationship is selected in the target network. After obtaining the level probability, the event level corresponding to the level range can be determined based on the level range that the level probability falls into.

[0071] Specifically, the subgraph probability of the directed subgraph is determined based on the conditional probability of each node in the directed subgraph to the target node it points to, wherein the conditional probability of each node to the target node it points to is obtained by learning based on event sample data labeled with the probability of the degree of influence of event elements on events, or the conditional probability is a pre-set target value.

[0072] See also Figure 2 , is a schematic diagram of a target network diagram provided in an embodiment of the present application, Figure 2 In , the nodes include x1, x2, x3, x4, x5, x6, and x7. Among them, the nodes representing the event level are the nodes without outgoing nodes, such as Figure 2 In this example, x6 and x7 are nodes with different influence levels. Assuming that node x4 directly affects node x6, a directed edge from x4 to x6 is established, and the connection strength is represented by the conditional probability P(x6|x4).

[0073] For example, X is a directed subgraph consisting of a set of event element nodes and the final level nodes, then:

[0074]

[0075] Among them, x pa(i) is x i The probability of a subgraph formed by the inbound nodes, such as x1, x3, x4, and x6, is p(x1, x3, x4, x6) = P(x1)P(x3)P(x4)P(x6|x4)P(x4|x1, x3).

[0076] The conditional probability can be obtained through event sample learning or by using artificially set prior probabilities. When the network structure is complex and there is a large amount of node data, the complex network can be simplified by clustering the nodes into one node, and then the event level can be determined by the above method. Specifically, when the number of nodes in the target network diagram is greater than the number threshold (such as k), clustering can be used to cluster the nodes, for example, clustering into m categories (m < k), and obtaining m initial centers. Calculate the distance from the node corresponding to the event element to its所属 initial center as the category weight of the event element, which can be used to adjust the calculation of conditional probability or prior probability later.

[0077] After determining the target network diagram (i.e., the belief network) based on the above process, it can be applied to the scenario of event level annotation for events. Specifically, the event elements that appear in the current sample can be selected on the belief network, and the corresponding event level and the probability of the level can be directly deduced by the belief network.

[0078] The following uses a specific application example to illustrate the process of event level annotation in the embodiments of the present application. Taking the annotation task of "the severity level of product problems" as an example, first, 500 sample data can be selected, and the problem summary and key entities can be extracted through the information extraction model. Domain experts label these data with four labels: severe, high, medium, and minor, and fill in the annotation basis in detail at the same time.

[0079] The summary, key entities, and the expert's annotation data are used as event elements, and some evaluation items related to the severity level of product problems, such as event occurrence frequency, symptom manifestations, recovery methods, etc., are also used as event elements. Based on the element association relationships between these event elements and the influence degree of these event elements on the event, calculations are performed, and the median is taken for the ranking of the influence strength of the problem. This process can construct a model based on the belief network and analyze it.

[0080] Calculations are performed according to the ranking of the strength of project evaluation. After calculation, different problems will fall into four score intervals. The scoring results (i.e., the four intervals of severe, high, medium, and minor) are compared with the expert's initial annotation results to determine whether they match and find out the reasons for non - matching; if it belongs to the unreasonable design of scoring items or formulas, the existing design is adjusted until the results are basically the same. When they are basically the same, an automatic annotation model can be obtained. For new data, when the new data is input into this automatic annotation model, the corresponding level annotation results can be obtained.

[0081] In this embodiment, the level benchmark is described through the belief network, which can describe complex event elements and the causal relationship between event elements and event levels. Its description ability is stronger, and the interpretability of event level determination is also stronger, improving the accuracy of event level determination.

[0082] See also Figure 3 , which is a schematic diagram of an application scenario provided by an embodiment of the present application. In this application scenario, the severity of a product failure is determined as an example. The instantiated belief network schematic diagram is as follows Figure 3 As shown. Among them, the severity of product failures can be divided into low, medium, high, and emergency. Among them, the event elements corresponding to the product severity event mainly correspond to the operation aspects, specific event occurrence conditions, event status, customer impact, functional status, etc. Specifically, the event elements can include the results generated for a certain operation, such as whether the daily core operation fails, whether it occurs when using specific options / special preparations, whether it only occurs under specific circumstances, the situation of encountering major customers each time, customer impact, etc.; it also includes the core functions of the enterprise, whether it can operate for a long time, whether the function will fail, the state of the function after failure, the total failure rate, serious violations of regulations, etc., whether the main function will fail, etc. According to the correlation between the above event elements and the relationship with the event itself, the severity of the failure corresponding to the product failure event can be determined.

[0083] For example, if the product partially fails to function after daily core operations fail, and this happens when using specific options / equipment, or only happens under specific circumstances, the product failure severity level is determined to be low. If the enterprise core functions corresponding to the product are slow and only happen under specific circumstances, the product failure severity level is determined to be medium. If the corresponding function fails when the daily core function corresponding to the product fails, and it will still affect customers after retrying in the case of major customers, the product failure severity level is determined to be high. If the product fails to function during long-term operation, always fails and seriously violates specifications, and the main function is invalid after retrying, the product failure level corresponding to the product is determined to be urgent.

[0084] See also Figure 4 In another embodiment of the present application, a data processing device is provided, comprising:

[0085] An acquisition unit 401 is used to acquire a target event;

[0086] An element determination unit 402 is used to determine an event element having a target association relationship with the target event;

[0087] A relationship determination unit 403, configured to determine an element association relationship between each of the event elements based on an element feature of each of the event elements;

[0088] The level determination unit 404 is used to determine the event level of the target event based on the target association relationship and the element association relationship, so as to process the target event based on the event level and obtain an event processing result.

[0089] The embodiment of the present application discloses a data processing device, including: obtaining a target event; determining event elements having a target association relationship with the target event; determining the element association relationship between each of the event elements based on the element characteristics of each of the event elements; determining the event level of the target event based on the target association relationship and the element association relationship, so that the target event is processed based on the event level to obtain an event processing result. The present application can process the event elements corresponding to the target event, determine the relationship between each event element and the target event, and the relationship between the elements, avoid the influence of subjective factors on the direct determination of the event level, and improve the accuracy of the event level determination.

[0090] Optionally, the element determination unit includes:

[0091] An extraction subunit, used to extract entities constituting the target event to obtain entity information;

[0092] The first determining subunit is used to determine event elements that satisfy the target association relationship in the entity information.

[0093] Furthermore, the target association relationship includes one or more of a causal relationship, a dependency relationship, or a master-slave relationship.

[0094] In one embodiment, the level determination unit includes:

[0095] A second determining subunit, configured to determine a weight parameter of each event element based on the element association relationship;

[0096] A third determining subunit is used to determine the relationship strength probability of each event element to the target event based on the target association relationship;

[0097] The fourth determining subunit is used to determine the event level of the target event according to the weight parameter and the relationship strength probability.

[0098] Optionally, the second determining subunit is specifically configured to:

[0099] Based on the element association relationship, clustering each of the event elements to obtain at least one category;

[0100] Based on the relationship between each event element and the category, a weight parameter of each event element is determined.

[0101] Optionally, the device further comprises:

[0102] A sample acquisition unit, used to acquire event samples, wherein each event sample is annotated with event elements corresponding to the event, element associations, and event levels;

[0103] A network determination unit is used to determine a target network graph based on the event sample, wherein nodes without outgoing edges of the target network represent event levels, nodes with outgoing edges in the target network represent event elements, and directed edges of the target network represent element association relationships.

[0104] Furthermore, the level determination unit includes:

[0105] A fifth determining subunit, configured to determine a directed subgraph corresponding to the target event based on the target network graph and the target association relationship;

[0106] a sixth determining subunit, configured to determine the subgraph probability corresponding to the directed subgraph as the level probability of the target event;

[0107] A seventh determining subunit is configured to determine the event level of the target event based on the level probability.

[0108] Optionally, it also includes:

[0109] An eighth determination subunit, configured to determine a subgraph probability of the directed subgraph based on a conditional probability of each node in the directed subgraph to a target node it points to;

[0110] The conditional probability of each node to the target node it points to is obtained by learning based on event sample data labeled with the probability of the degree of influence of event elements on events, or the conditional probability is a preset target value.

[0111] It should be noted that the specific implementation of each unit and sub-unit in this embodiment can refer to the corresponding content in the previous text and will not be described in detail here.

[0112] In another embodiment of the present application, a readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, each step of the data processing method as described in any one of the above items is implemented.

[0113] See also Figure 5 , is a structural diagram of an electronic device provided in another embodiment of the present application. The technical solution in this embodiment is mainly used to improve the accuracy of event level determination.

[0114] Specifically, the electronic device in this embodiment may include the following structure:

[0115] Memory 501, used for storing programs;

[0116] The processor 502 is configured to call and execute the program in the memory, and implement the following by executing the program:

[0117] Get the target event;

[0118] Determining event elements having a target association relationship with the target event;

[0119] Determining, based on the element characteristics of each of the event elements, element association relationships between the event elements;

[0120] Based on the target association relationship and the element association relationship, an event level of the target event is determined, so that the target event is processed based on the event level to obtain an event processing result.

[0121] Optionally, the determining of the event elements having a target association relationship with the target event includes:

[0122] Extracting entities constituting the target event to obtain entity information;

[0123] Event elements satisfying the target association relationship are determined in the entity information.

[0124] Optionally, the target association relationship includes one or more of a causal relationship, a dependency relationship, or a master-slave relationship.

[0125] Optionally, determining the event level of the target event based on the target association relationship and the element association relationship includes:

[0126] Determine a weight parameter of each event element based on the element association relationship;

[0127] Based on the target association relationship, determining the relationship strength probability of each event element to the target event;

[0128] The event level of the target event is determined according to the weight parameter and the relationship strength probability.

[0129] Optionally, determining a weight parameter of each event element based on the element association relationship includes:

[0130] Based on the element association relationship, clustering each of the event elements to obtain at least one category;

[0131] Based on the relationship between each event element and the category, a weight parameter of each event element is determined.

[0132] Optionally, the method further comprises:

[0133] Acquire event samples, wherein each event sample is annotated with event elements corresponding to the event, element association relationships, and event levels;

[0134] Based on the event samples, a target network graph is determined, wherein nodes without outgoing edges of the target network graph represent event levels, nodes with outgoing edges in the target network graph represent event elements, and directed edges of the target network graph represent element association relationships.

[0135] Optionally, determining the event level of the target event includes:

[0136] Based on the target network graph and the target association relationship, determining a directed subgraph corresponding to the target event;

[0137] Determine the subgraph probability corresponding to the directed subgraph as the level probability of the target event;

[0138] Based on the level probability, an event level of the target event is determined.

[0139] Optionally, it also includes:

[0140] Determining a subgraph probability of the directed subgraph based on a conditional probability of each node in the directed subgraph to a target node it points to;

[0141] The conditional probability of each node to the target node it points to is obtained by learning based on event sample data labeled with the probability of the degree of influence of event elements on events, or the conditional probability is a preset target value.

[0142] It should be noted that the specific implementation of the processor in this embodiment can refer to the corresponding content in the previous text and will not be described in detail here.

[0143] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0144] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0145] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0146] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A data processing method, comprising: Get the target event; Determining event elements having a target association relationship with the target event; The target association relationship includes one or more of a causal relationship, a dependency relationship, or a master-slave relationship; Determining, based on the element characteristics of each of the event elements, element association relationships between the event elements; Based on the target association relationship and the element association relationship, determine the event level of the target event, so that the target event is processed based on the event level to obtain an event processing result; Wherein, determining the event level of the target event based on the target association relationship and the element association relationship includes: Determine a weight parameter of each event element based on the element association relationship; Based on the target association relationship, determining the relationship strength probability of each event element to the target event; The event level of the target event is determined according to the weight parameter and the relationship strength probability.

2. According to the method of claim 1, the determining of the event elements having a target association relationship with the target event comprises: Extracting entities constituting the target event to obtain entity information; Event elements satisfying the target association relationship are determined in the entity information.

3. The method according to claim 1, wherein determining the weight parameter of each event element based on the element association relationship comprises: Based on the element association relationship, clustering each of the event elements to obtain at least one category; Based on the relationship between each event element and the category, a weight parameter of each event element is determined.

4. The method according to claim 1, further comprising: Acquire event samples, wherein each event sample is annotated with event elements corresponding to the event, element association relationships, and event levels; Based on the event samples, a target network graph is determined, wherein nodes without outgoing edges of the target network represent event levels, nodes with outgoing edges in the target network represent event elements, and directed edges of the target network represent element association relationships.

5. The method according to claim 4, wherein determining the event level of the target event comprises: Based on the target network graph and the target association relationship, determining a directed subgraph corresponding to the target event; Determine the subgraph probability corresponding to the directed subgraph as the level probability of the target event; Based on the level probability, an event level of the target event is determined.

6. The method according to claim 5, further comprising: Determining a subgraph probability of the directed subgraph based on a conditional probability of each node in the directed subgraph to a target node it points to; The conditional probability of each node to the target node it points to is obtained by learning based on event sample data labeled with the probability of the degree of influence of event elements on events, or the conditional probability is a preset target value.

7. A data processing device, comprising: An acquisition unit, used for acquiring a target event; An element determination unit, used to determine an event element having a target association relationship with the target event; The target association relationship includes one or more of a causal relationship, a dependency relationship, or a master-slave relationship; A relationship determination unit, configured to determine an element association relationship between each of the event elements based on an element feature of each of the event elements; A level determination unit, configured to determine the event level of the target event based on the target association relationship and the element association relationship, so as to process the target event based on the event level and obtain an event processing result; Wherein, the level determination unit includes: A second determining subunit, configured to determine a weight parameter of each event element based on the element association relationship; A third determining subunit is used to determine the relationship strength probability of each event element to the target event based on the target association relationship; The fourth determining subunit is used to determine the event level of the target event according to the weight parameter and the relationship strength probability.

8. An electronic device comprising: Memory, used to store programs; A processor is used to call and execute the program in the memory, and implement the various steps of the data processing method as described in any one of claims 1 to 6 by executing the program.

Images