Event circulation method, related device, equipment, system and storage medium

Abstract

The invention discloses an event circulation method, a related device, equipment, a system and a storage medium, and the method comprises the steps: carrying out the element extraction of first description data of a to-be-circulated event based on different element extraction strategies, and obtaining a plurality of element sets; acquiring a generality set of the plurality of element sets and at least one non-generality set; based on the generality set and a sampling set obtained by performing element sampling on the non-generality set by using the element sampling probability, obtaining a target set; at least based on the first description data, a target set and a reference event library, determining a target platform to which the to-be-circulated event is about to flow; and in response to a return instruction of the target platform to the to-be-circulated event, reducing the element sampling probability, and at least performing element sampling on the new to-be-circulated event according to the new element sampling probability. According to the scheme, the event circulation precision and efficiency can be improved, especially under the cross-domain condition of the to-be-circulated event.

Description

Technical Field

[0001] The present application relates to the technical field of event dispatching, and in particular to an event flow method and related devices, equipment, systems and storage media. Background Art

[0002] Event flow has broad and far-reaching significance in many scenarios such as automated office and urban governance. Efficient and accurate event flow can significantly improve event handling efficiency.

[0003] At present, the existing event transfer technology mainly relies on auditors to classify events and then manually transfer events to the corresponding responsible units based on the mapping relationship between "matter classification-responsibility department". However, the accuracy of this event transfer technology depends on the subjective judgment of auditors, which makes it difficult to ensure the accuracy of event transfer and greatly affects the efficiency of event transfer, especially when the events to be transferred are cross-domain. In view of this, how to improve the accuracy and efficiency of event transfer has become an urgent problem to be solved. Summary of the invention

[0004] The main technical problem solved by the present application is to provide an event flow method and related devices, equipment, system and storage medium, which can improve the accuracy and efficiency of event flow.

[0005] In order to solve the above technical problems, the first aspect of the present application provides an event flow method, including: extracting elements from the first description data of the event to be circulated based on different element extraction strategies to obtain several element sets; obtaining a common set and at least one non-common set of the several element sets; obtaining a target set based on the common set and the sampling set obtained by sampling elements of the non-common set using the element sampling probability; determining a target platform to which the event to be circulated is to be flowed based on at least the first description data, the target set and the reference event library; in response to the target platform's return instruction for the event to be circulated, reducing the element sampling probability, and the new element sampling probability at least performs element sampling on the new event to be circulated.

[0006] In order to solve the above technical problems, the second aspect of the present application provides an event flow device, including: an element extraction module, a set processing module, an element sampling module, a platform determination module and a probability update module, the element extraction module is used to extract elements from the first description data of the event to be circulated based on different element extraction strategies to obtain several element sets; the set processing module is used to obtain a common set and at least one non-common set of several element sets; the element sampling module is used to obtain a target set based on the common set and the sampling set obtained by sampling the elements of the non-common set using the element sampling probability; the platform determination module is used to determine the target platform to which the event to be circulated is to flow based on at least the first description data, the target set and the reference event library; the probability update module is used to reduce the element sampling probability in response to the return instruction of the target platform to the flow event, and the new element sampling probability at least performs element sampling on the new event to be circulated.

[0007] In order to solve the above technical problems, the third aspect of the present application provides an electronic device, which at least includes a memory and a processor coupled to each other, the memory at least stores program instructions, and the processor is used to execute the program instructions to implement the event flow method in the above first aspect.

[0008] In order to solve the above-mentioned technical problems, the fourth aspect of the present application provides an event flow system, including a system front-end, a system back-end, a convergence platform and several preset platforms that are communicatively connected. The system front-end is used to receive events reported by user terminals, and the system back-end is used to store events and forward events to the convergence platform. The convergence platform includes at least the electronic device as described in the third aspect above.

[0009] In order to solve the above technical problems, the fifth aspect of the present application provides a computer-readable storage medium, which stores program instructions that can be executed by a processor, and the program instructions are used to implement the event flow method of the first aspect above.

[0010] The above scheme extracts elements from the first description data of the event to be circulated based on different element extraction strategies, obtains several element sets, obtains common sets and non-common sets of the several element sets, and then obtains the target set based on the common set and the sampling set obtained by sampling elements of the non-common set using the element sampling probability, thereby determining the target platform to which the event to be circulated is to be directed based on at least the first description data, the target set and the reference event library, and then responds to the return instruction of the target platform for the event to be circulated, reduces the element sampling probability, and the new element sampling probability at least samples the elements of the new event to be circulated, so on the one hand Compared with manual event transfer, event transfer can be realized through a series of operations such as feature extraction, feature sampling, and platform determination, which can improve the efficiency of event transfer. On the other hand, since the feedback instructions of the target platform for the transfer event are monitored in real time during the whole process of event transfer, and the feature sampling probability is timely reduced when the return instruction is received to reduce the sampling of features of non-common sets, the feature sampling probability can be adaptively updated and adjusted under the indirect intervention of the target platform, which helps to extract features from non-common sets as much as possible, so as to improve the accuracy of determining the target platform as much as possible, that is, to improve the accuracy of event transfer. Therefore, the accuracy and efficiency of event transfer can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a flowchart of an embodiment of the event flow method of the present application; Figure 2a It is a process diagram of an embodiment of the event flow method of the present application; Figure 2b It is a process schematic diagram of another embodiment of the event flow method of the present application; Figure 3 It is a schematic diagram of the framework of an embodiment of the event flow device of the present application; Figure 4 It is a schematic diagram of the framework of an embodiment of the electronic device of the present application; Figure 5 It is a schematic diagram of the framework of an embodiment of the event flow system of the present application; Figure 6 It is a schematic diagram of a framework of an embodiment of a computer-readable storage medium of the present application. DETAILED DESCRIPTION

[0012] The scheme of the embodiment of the present application is described in detail below in conjunction with the drawings of the specification.

[0013] In the following description, for the purpose of explanation rather than limitation, specific details such as specific system structures, interfaces, and technologies are provided to facilitate a thorough understanding of the present application.

[0014] The terms "system" and "network" are often used interchangeably in this article. The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, the fragment " / " in this article generally indicates that the associated objects before and after are in an "or" relationship. In addition, "many" in this article means two or more than two.

[0015] See also Figure 1 , Figure 1 This is a flow chart of an embodiment of the event transfer method of the present application. Specifically, it may include the following steps: Step S11: extracting elements from the first description data of the flow event based on different element extraction strategies to obtain a plurality of element sets.

[0016] In an implementation scenario, different factor extraction strategies may include, but are not limited to: factor extraction strategies based on neural network models, factor extraction strategies based on set rules, etc., and the specific types of factor extraction strategies are not limited here. It should be noted that the neural network model on which the factor extraction strategy is based can be either an existing neural network or a custom neural network; similarly, the set rules on which the factor extraction strategy is based can be either an existing rule or a custom rule, and are not limited here.

[0017] In a specific implementation scenario, different factor extraction strategies can be based on different underlying technologies. For example, two factor extraction strategies are used, one of which can be based on a neural network model and the other can be based on a set rule.

[0018] In another specific implementation scenario, different factor extraction strategies can also be based on the same underlying technology. Still taking two factor extraction strategies as an example, both can be based on a neural network model, but one can be based on a BERT network model and the other can be based on an LSTM network model.

[0019] In an implementation scenario, the event to be transferred can be reported by the user terminal. The specific reporting process of the event to be transferred can refer to the relevant description in the event transfer system embodiment below, which will not be repeated here. For example, the user terminal can run an APP for event reporting to report events in the APP through text input, voice input, attached images, etc.; or, the user terminal can also report events through a web page, and the web page can also report events through text input, voice input, attached images, etc.

[0020] In one implementation scenario, the first description data of the event to be circulated may include text content of text input; or, in the case of voice input, the first description data of the event to be circulated may also include the recognized text of the voice input; or, in the case of an attached image, the first description data of the event to be circulated may also include the description text obtained by recognizing the attached image. Of course, the above examples are only several possible examples in actual application, and other possible scenarios will not be given one by one here.

[0021] In an implementation scenario, when an event is reported, the reporting time, reporter information, reporting location information, etc. of the event to be circulated may also be attached to the first description data; or, the above-mentioned reporting time, reporter information, reporting location information, etc. may also be packaged together in the first description data, which is not limited here.

[0022] In an implementation scenario, any element set may include element values ​​of several element tags in the first description data. As a possible example, several element tags may include but are not limited to: the place where the event occurred, the time when the event occurred, the key event tag, the person associated with the event, the event organization, the person who reported the event, etc. The above examples are only several possible examples in the actual application process, and other possible situations are not given one by one here. Take the first description data as "X reported in XX city at XX: I temporarily parked in XX section of road, and now I have received an on-site penalty from the traffic police X of the XX traffic police brigade, and was fined XX points and XX yuan for driving license deduction. How to deal with it if I am dissatisfied" as an example, then the element value of the element tag "place where the event occurred" is "XX section", the element value of the element tag "time of the event occurred" is "XX: XX, XX, XX", the element value of the element tag "key event tag" is "on-site penalty", "I am dissatisfied", the element value of the element tag "person associated with the event" is "traffic police X", the element value of the element tag "event organization" is "XX traffic police brigade", and the element value of the element tag "person who reported the event" is "X". Of course, the above examples are only a few possible examples in actual application, and other possible situations will not be given one by one here.

[0023] For the convenience of description, the following takes the use of two element extraction strategies as an example, which are respectively referred to as the first extraction strategy and the second extraction strategy. The element set extracted by the first extraction strategy can be referred to as the first element set, and the element set extracted by the second extraction strategy can be referred to as the second element set. Among them, the first extraction strategy can be recorded as W1, the second extraction strategy can be recorded as W2, the first element set can be recorded as S1, and the second element set can be recorded as S2. Of course, the above example is only a possible example in the actual application process, and other possible situations are not limited thereto. For example, three element extraction strategies, four element extraction strategies, five element extraction strategies, etc. can also be adopted, and the same can be said accordingly, and examples will not be given one by one here.

[0024] Step S12: Obtain a commonality set and at least one non-commonality set of several element sets.

[0025] Specifically, the commonality set of several element sets represents an element set formed by element values ​​existing in all element sets, and the non-commonality set of several element sets represents an element set formed by element values ​​existing in one of the element sets but not in at least one other element set. Still taking the two element extraction strategies as an example, the commonality set S3 of the first element set S1 and the second element set S2 can be obtained, and the first non-commonality set S1-S3 formed by element values ​​existing in the first element set S1 but not in the second element set S2 can be obtained, and the second non-commonality set S2-S3 formed by element values ​​not existing in the first element set S1 but existing in the second element set S2 can be obtained. Of course, the above example is only a possible example in the actual application process, and other possible situations will not be given examples one by one here.

[0026] Step S13: Obtain a target set based on the commonality set and a sampling set obtained by sampling elements of the non-commonality set using element sampling probability.

[0027] In an implementation scenario, the union of the commonality set and the sampling set obtained by sampling the elements of the non-commonality set using the element sampling probability can be used as the target set. Still taking the two element extraction strategies as an example, the target set St can be obtained based on the commonality set S3, the first sampling set obtained by sampling the elements of the first non-commonality set S1-S3 using the first sampling probability W1, and the second sampling set obtained by sampling the elements of the second non-commonality set S2-S3 using the second sampling probability W2: St=W1*(S1-S3)+W2*(S2-S3)+S3……(1) In the above formula (1), W1*(S1-S3) indicates random sampling in the first non-commonality set S1-S3 with the sampling probability of W1, and W2*(S2-S3) indicates random sampling in the second non-commonality set S2-S3 with the sampling probability of W2. Of course, the above example is only a possible example in the actual application process, and other possible situations can be deduced by analogy to obtain the target set. Other possible situations will not be given examples one by one here.

[0028] In an implementation scenario, as mentioned above, if any element set contains element values ​​about several element tags in the first description data, it is also possible to detect whether the element sets extracted by each element extraction strategy completely cover several element tags, and in response to the element set extracted by any element extraction strategy completely covering several element tags, increase the element sampling probability corresponding to the element extraction strategy. Still taking the use of two element extraction strategies as an example, if the first element set extracted by the first extraction strategy completely covers all element tags, the element sampling probability corresponding to the first extraction strategy, that is, the first sampling probability W1, can be increased; similarly, if the second element set extracted by the second extraction strategy completely covers all element tags, the element sampling probability corresponding to the second extraction strategy, that is, the second sampling probability W2, can be increased. Of course, the above example is only a possible example in the actual application process, and other possible situations will not be given one by one here. In the above method, when the feature set extracted by the feature extraction strategy completely covers all feature labels, the feature sampling probability corresponding to the feature extraction strategy is increased. Therefore, the feature sampling probability can be adaptively improved according to the reliability of the feature extraction strategy to increase the feature values ​​specially extracted by the feature extraction strategy in the target set, which helps to improve the reliability of event flow based on this.

[0029] In a specific implementation scenario, when it is necessary to increase the sampling probability of the element corresponding to the element extraction strategy, the probability increase amplitude can be obtained based on the current average daily number of events, and the probability increase amplitude is negatively correlated with the current average daily number of events. That is to say, the larger the current average daily number of events, the smaller the probability increase amplitude, and conversely, the smaller the current average daily number of events, the larger the probability increase amplitude. Exemplarily, the probability increase amplitude can be obtained by dividing the preset probability amplitude (such as 0.5) by the current average daily number of events. On this basis, the probability increase amplitude can be added to the current element sampling probability of the element extraction strategy. The above method, which adaptively determines the probability increase amplitude in combination with the current average daily number of events, helps to improve the reliability of the probability increase amplitude.

[0030] In another specific implementation scenario, different from the aforementioned implementation method of increasing the element sampling probability, a fixed increase amplitude (such as 0.00001, etc.) can be added to the current element sampling probability of the element extraction strategy. The specific value of the fixed increase amplitude is not limited here.

[0031] It should be noted that in the embodiment of the present disclosure, once the element sampling probability is adjusted, it can take effect on the next execution of the event flow method process in the embodiment of the present disclosure. Of course, it is also possible to customize the delay to take effect when the element sampling probability is adjusted for practical needs.

[0032] Step S14: Determine a target platform to which the event to be transferred is to be transferred based at least on the first description data, the target set and the reference event library.

[0033] In one implementation scenario, the reference event library may contain several reference events, and the reference events are also accompanied by marked event categories. Exemplarily, the reference event library may cover various event categories as comprehensively as possible, and the total number of reference events under each event category may be no less than a preset number (e.g., 10, 15, etc.). In addition, the reference events may be obtained in advance from various preset platforms, or may be automatically generated through generative artificial models such as large language models. The specific source of the reference events is not limited here. For the specific content of the reference events, please refer to the aforementioned description of the events to be circulated, and no examples will be given here one by one.

[0034] In an implementation scenario, the labeled event category of a reference event may at least include a responsible department, a responsibility item, and a specific classification. Among them, the responsible department is responsible for handling the reference event, the responsibility item represents a responsibility among the various responsibilities of the responsible department that corresponds to the reference event, and the specific classification represents a sub-category of the responsibility that matches the reference event. For ease of understanding, taking the responsible department "traffic management department" as an example, it may have multiple responsibility items such as "violation handling", "document handling", and "vehicle inspection". For the responsibility item "document handling", it may include multiple specific classifications such as "driving license processing" and "vehicle license processing". It should be noted that the above example is only a possible example in the actual application process, and other possible situations will not be given one by one here.

[0035] In an implementation scenario, after obtaining the first description data, the target set and the reference event library, the reference event can be selected as a candidate event based on the similarity between the first description data and the second description data of each reference event. It should be noted that the specific meaning of the second description data of the reference event can refer to the relevant description of the aforementioned first description data, which will not be repeated here. On this basis, based on the target set and the labeled event categories of each candidate event, the labeled event category of the candidate event can be selected as the target event category of the event to be circulated, and then based on the target event category and the first element value related to the location in the target set, the preset platform is selected as the target platform, and the configuration information of any preset platform at least includes the case category range and the acceptance geographical range of the preset platform. It should be noted that the preset platform is a disposal platform for receiving events to be circulated. In the actual application process, multiple preset platforms can be set according to the principles of geographical area, work responsibilities, etc., so that the events to be circulated can be selectively circulated according to the actual situation, and meet the needs of cross-domain circulation of events. The above method first selects reference events as candidate events by describing the similarity between data, then determines the target event category of the event to be transferred through the labeled event category of the candidate event and the target set determined by the element extraction of the event to be transferred, and then determines the target platform based on this. This can improve the accuracy of determining the target platform based on the reference event.

[0036] In a specific implementation scenario, as a possible example, the second description data of each reference event can be separately feature extracted in advance to obtain its feature representation. For example, the feature representation can be represented by a vector and stored in a vector database. After obtaining the first description data, the first description data can be further feature extracted to obtain its feature representation, which can also be represented by a vector. On this basis, the reference event can be selected as a candidate event based on the feature similarity (such as cosine similarity) between the feature representation of the first description data represented by a vector and the feature representation of each second description data represented by a vector. For example, the reference events can be sorted in order of feature similarity from high to low, and the reference events in the front preset order (such as the top 3, etc.) can be selected as candidate events.

[0037] In a specific implementation scenario, after selecting a candidate event, it is possible to continue to detect whether there is an element value in the target set that matches any of the responsibility departments, responsibility items, and specific classifications in the labeled event category of the candidate event, so that in response to the existence of an element value that meets the above conditions, the labeled event category of the candidate event can be selected as the target event category. For example, still taking the event to be transferred "X reported at XX: XX: I temporarily parked in XX section of road, and now I have received an on-site penalty from the traffic police X of the XX Traffic Police Brigade, and was fined XX points and XX yuan for driving license. How do I deal with it if I am dissatisfied?" as an example, its target set may include {element label "Place of event occurrence": "XX section"; element label "Time of event occurrence": "XX: XX: XX"; element label "Event key label": "On-site penalty", "I am dissatisfied"; element label "Event related person": "Traffic police X"; element label "Event organization": "XX Traffic Police Brigade"; element label "Event reporter": "X"}, if there is In the labeled event classification of a candidate event, the responsible department is "traffic management department", the responsibility item is "traffic violation handling", and the specific classification is "traffic violation review", then the element label "event organization": "XX traffic police brigade" in the target set matches the responsible department "traffic management department", the element label "event key label": "on-site punishment" in the target set matches the responsibility item "traffic violation handling", and the element label "event key label": "I am dissatisfied" in the target set matches the specific classification "traffic violation review", so the labeled event category of the reference event (that is, the responsible department is "traffic management department", the responsibility item is "traffic violation handling", and the specific classification is "traffic violation review") can be used as the target event category of the event to be transferred. Of course, the above example is only a possible example in the actual application process, and the determination process of the target event category will not be given one by one here. For example, when there are multiple candidate events that meet the above requirements, the labeled event category of the candidate event that matches the most element values ​​with the target set can be selected as the target event category of the event to be transferred.

[0038] In a specific implementation scenario, after determining the target event category, the preset platform can be selected as the target platform in combination with the location-related element value in the target set. Specifically, for each preset platform, it can be detected whether the location-related element value in the target set is located in the case acceptance geographical range of the preset platform and whether the target event category is within the case acceptance category range of the preset platform. If both are judged to be yes, the preset platform can be selected as the target platform. Still taking the aforementioned pending transfer event as an example, if the case acceptance category range of a preset platform includes "traffic violation processing" and "certificate processing", and the case acceptance geographical range includes "XX District", since the location-related element value "XX section" in the target set is located in the case acceptance geographical range "XX District" of the preset platform, and the target event category is also within the case acceptance category range of the preset platform, the preset platform can be selected as the target platform.

[0039] In a specific implementation scenario, the configuration information of any preset platform may also include the event receiving method of the preset platform, and the event to be transferred flows to the target platform according to the event receiving method defined in the configuration information of the target platform. Exemplarily, the event receiving method may include but is not limited to: encrypted transmission with a certain agreed encryption algorithm, unencrypted transmission, sending to a designated mailbox, packaging into a certain frame format and sending to a specific port, etc. Of course, the above examples can be selectively combined to form a new event receiving method, such as "encrypted transmission with a certain agreed encryption algorithm and sent to a designated mailbox", "unencrypted transmission and sent to a designated mailbox", "encrypted transmission with a certain agreed encryption algorithm and packaged into a certain frame format and sent to a specific port", etc. The above examples are only a few possible examples of event receiving methods, and other possible situations will not be given one by one here.

[0040] In another implementation scenario, different from the specific method of determining the target platform mentioned above, as another possible implementation method, after obtaining a reference event library with labeled event categories, the large language model can be used to learn the second description data of the reference event library and its labeled event categories. On this basis, after obtaining the first description data and the target set, a prompt instruction can be constructed, and the prompt instruction is used to instruct the large language model to refer to the first description data and the target set to determine the target event category to which the event to be transferred belongs. Still taking the aforementioned event to be transferred "X reported at XX:00 on XX: XX: I temporarily parked on XX section of road, and now I have received an on-site penalty from the traffic police X of the XX Traffic Police Brigade, and was fined XX points and XX yuan for my driver's license. How do I deal with it if I don't accept it" as an example, the following prompt instructions can be constructed, including but not limited to: You are an event classification expert. The current event description is: X reported at XX:XX on XX:XX in XX City: A person temporarily parked his car on the XX road section and was punished by the traffic police X of the XX traffic police brigade. He was fined XX points and fined XX yuan. What to do if I am not satisfied with the punishment?

[0041] In addition, the elements extracted from this event are as follows: Element tag "place where the incident occurred": "XX road section"; Element tag "Event occurrence time": "XX month XX day XX time"; Element label "Event key label": "On-site punishment", "I am dissatisfied"; Element tag "person related to the incident": "traffic policeman X"; Element tag "Event Organization": "XX Traffic Police Brigade"; Element tag "Person who reported the incident": "X".

[0042] Please combine the event description and the above elements to output the event category to which the event belongs.

[0043] It should be noted that the above examples are only several possible examples of determining the target event category in actual application, and the specific methods of determining the target event category will not be given one by one here.

[0044] Step S15: In response to the return instruction of the target platform to the event to be transferred, the element sampling probability is reduced, and the new element sampling probability at least samples the elements of the new event to be transferred.

[0045] In one implementation scenario, if a return instruction is received from the target platform for the event to be transferred, it can be determined that the event to be transferred has a transfer error. In this case, as a possible example, the extraction of elements from non-common sets can be appropriately reduced, that is, the element sampling probability can be reduced to appropriately reduce the extraction of non-common elements from the perspective of element sampling probability, and adaptively exclude non-common elements that may interfere. For example, the element sampling probability can be subtracted from a fixed value (e.g., any value in the range of 0 to 0.1). In addition, when there are multiple element sampling probabilities (such as the aforementioned first sampling probability, second sampling probability, etc.), at least one of them can be selected and reduced.

[0046] In another implementation scenario, as another possible example, if a return instruction is received from the target platform for a circulation event, several element labels can also be re-determined based on the return reason parsed from the return instruction. For example, taking the return reason "Please go to the responsible department with jurisdiction over the reporting location for processing" as an example, it can be re-determined: add the element label "location of event reporting". Of course, the above example is only a possible example in the actual application process, and other possible situations are not limited here. The above method, after receiving the return instruction from the target platform, re-determines several element labels based on the return reason parsed from the return instruction. From the perspective of element labels, it can be possible to adjust the element labels in a targeted manner in combination with the return reason, which helps to improve the accuracy of subsequent event classification, platform determination and other operations.

[0047] In an implementation scenario, different from the aforementioned situation where a return instruction is received, in actual application, if a completion instruction for a circulation event is received from the target platform, the event to be circulated can be taken as a new reference event and included in the reference event library. Of course, when it is included in the reference event library, its target event category can be stored with it as a labeled event category, and its first description data can be stored with it as the second description data. In addition, as mentioned above, when feature similarity is used for similarity measurement, its feature representation can be simultaneously included in the vector database for direct retrieval and use during subsequent event circulation. In the above method, when a completion instruction for a circulation event is received from the target platform, the event to be circulated is included in the reference event library as a new reference event, and the reference event library can be continuously updated with completed events, which helps to improve the accuracy of determining the target platform.

[0048] In an implementation scenario, please refer to Figure 2a and Figure 2b , Figure 2a This is a process diagram of an embodiment of the event flow method of the present application. Figure 2b This is a schematic diagram of another embodiment of the event transfer method of the present application. Figure 2a and Figure 2bA possible example of the event flow method of the present application is described as a whole. Users can report events through terminals such as mobile phones. For example, when reporting events through an APP, the reported information will be transmitted to source platform 1. When reporting events through a web page, the reported information will be transmitted to source platform 2, ..., and so on. When reporting events through other channels, the reported information will be transmitted to source platform n. In addition, the reported information can mainly include event description text (such as text input text, voice recognition text, etc.), and incidentally, it can also include relevant information at the time of reporting, such as reporting time, reporter, reporting location, etc. Each event is separately used as an event to be circulated, and the above information of each event is integrated and qualified after quality verification, which can be used as its first description data for subsequent event element extraction operations. Specifically, a variety of element extraction strategies can be used to extract the element values ​​of several event tags (such as the place where the event occurred, the time of the event, the key tags of the event, the event-related persons, the event organization, the event reporter, etc.) of the first description data of the event to be circulated, so as to obtain the element sets corresponding to the various element extraction strategies, and then the commonality set of these element sets and the sampling set obtained by sampling the non-commonality set of these element sets using the element sampling probability can be combined to obtain the target set. On this basis, the reference event can be further selected as the candidate event according to the similarity between the first description data of the event to be circulated and the second description data of each reference event, and based on the above target set and the annotated event category of each candidate event, the annotated event category of the candidate event is selected as the target event category of the event to be circulated, so as to complete the event classification of the event to be circulated, and then the platform determination can be continued to determine several preset platforms (such as preset platform 1, preset platform 2, ..., preset platform n) for handling the target platform of the event to be circulated, and the platform address corresponding to each area maintained and provided by the platform library can be combined to realize the flow of the event to the target platform. If a return instruction is received from the target platform for the event to be transferred (that is, the target platform explicitly does not accept the event to be transferred), the above platform can be returned to the decision-making platform, and the event with the error of assignment can be used as the basis for the decision area of ​​similar events. For example, the element can be re-determined according to the reason for the return; and / or, the sampling probability of the element can be reduced, etc., which are not limited here. On the contrary, if the target platform accepts the event to be transferred, the process can be terminated. Of course, the reference event library can also be updated with the event to be transferred as a new reference event. In addition, the above process description is only the overall process in the actual application process. For details, please refer to the above related description, which will not be repeated here.

[0049] It should be noted that after the element label is re-determined or the element sampling probability is readjusted, the new element label or new element sampling probability can take effect on the next event flow. The next event flow can include both the returned events to be transferred and the newly acquired events to be transferred, which are not limited here.

[0050] The above scheme extracts elements from the first description data of the event to be circulated based on different element extraction strategies, obtains several element sets, obtains common sets and non-common sets of the several element sets, and then obtains the target set based on the common set and the sampling set obtained by sampling elements of the non-common set using the element sampling probability, thereby determining the target platform to which the event to be circulated is to be directed based on at least the first description data, the target set and the reference event library, and then responds to the return instruction of the target platform for the event to be circulated, reduces the element sampling probability, and the new element sampling probability at least samples the elements of the new event to be circulated, so on the one hand Compared with manual event transfer, event transfer can be realized through a series of operations such as feature extraction, feature sampling, and platform determination, which can improve the efficiency of event transfer. On the other hand, since the feedback instructions of the target platform for the transfer event are monitored in real time during the whole process of event transfer, and the feature sampling probability is timely reduced when the return instruction is received to reduce the sampling of features of non-common sets, the feature sampling probability can be adaptively updated and adjusted under the indirect intervention of the target platform, which helps to extract features from non-common sets as much as possible, so as to improve the accuracy of determining the target platform as much as possible, that is, to improve the accuracy of event transfer. Therefore, the accuracy and efficiency of event transfer can be improved.

[0051] See also Figure 3 , Figure 3 It is a schematic diagram of the framework of an embodiment of the event flow device of the present application. The event flow device 30 includes: an element extraction module 31, a set processing module 32, an element sampling module 33, a platform determination module 34 and a probability update module 35. The element extraction module 31 is used to extract elements from the first description data of the flow event based on different element extraction strategies to obtain several element sets; the set processing module 32 is used to obtain a common set and at least one non-common set of several element sets; the element sampling module 33 is used to obtain a target set based on the common set and the sampling set obtained by sampling the elements of the non-common set using the element sampling probability; the platform determination module 34 is used to determine the target platform to which the flow event is to be transferred based on at least the first description data, the target set and the reference event library; the probability update module 35 is used to reduce the element sampling probability in response to the return instruction of the flow event by the target platform, and the new element sampling probability is used to perform element sampling on at least the new flow event.

[0052] In the above scheme, the event flow device 30 extracts elements from the first description data of the event to be circulated based on different element extraction strategies, obtains several element sets, obtains common sets and non-common sets of the several element sets, and then obtains the target set based on the common set and the sampling set obtained by sampling elements of the non-common set using the element sampling probability, thereby determining the target platform to which the event to be circulated is to be directed based on at least the first description data, the target set and the reference event library, and then responds to the return instruction of the target platform to the event to be circulated, reduces the element sampling probability, and the new element sampling probability at least samples the elements of the new event to be circulated, so On the one hand, compared with manually realizing event flow, realizing event flow through a series of operations such as feature extraction, feature sampling, and platform determination can improve the efficiency of event flow. On the other hand, since the feedback instructions of the target platform for the flow event are monitored in real time during the whole process of event flow, and the feature sampling probability is timely reduced when the return instruction is received to reduce the sampling of features of non-common sets, the feature sampling probability can be adaptively updated and adjusted under the indirect intervention of the target platform, which helps to extract features from non-common sets as much as possible, so as to improve the accuracy of determining the target platform as much as possible, that is, to improve the accuracy of event flow. Therefore, the accuracy and efficiency of event flow can be improved.

[0053] In some disclosed embodiments, several element sets include a first element set extracted based on a first extraction strategy and a second element set extracted based on a second extraction strategy, and the set processing module 32 is specifically used to obtain a common set between the first element set and the second element set, and to obtain a first non-common set formed by element values ​​that exist in the first element set but not in the second element set, and to obtain a second non-common set formed by element values ​​that do not exist in the first element set but exist in the second element set; the element sampling module 33 is specifically used to obtain a target set based on the common set, using a first sampling set obtained by sampling elements of the first non-common set using a first sampling probability, and using a second sampling set obtained by sampling elements of the second non-common set using a second sampling probability.

[0054] In some disclosed embodiments, the feature set includes feature values ​​regarding several feature tags in the first description data, and the event flow device 30 also includes a set detection module for detecting whether the feature sets extracted by each feature extraction strategy respectively completely cover several feature tags; the event flow device 30 also includes a probability increase module for increasing the feature sampling probability corresponding to the feature extraction strategy in response to the feature set extracted by any feature extraction strategy completely covering several feature tags.

[0055] In some disclosed embodiments, the probability increase module includes an amplitude determination submodule, which is used to obtain the probability increase amplitude based on the current average daily number of events; wherein the probability increase amplitude is negatively correlated with the current average daily number of events; the probability increase module includes an amplitude increase submodule, which is used to add the probability increase amplitude based on the current element sampling probability of the element extraction strategy.

[0056] In some disclosed embodiments, any element set includes element values ​​regarding several element tags in the first description data, and the event flow device 30 also includes an element determination module for re-determining several element tags based on the return reason parsed from the return instruction when a return instruction is received.

[0057] In some disclosed embodiments, the reference event library includes a number of reference events, and the platform determination module 34 includes a similarity measurement submodule for selecting reference events as candidate events based on the similarity between the first description data and the second description data of each reference event; the platform determination module 34 includes an event classification submodule for selecting the labeled event category of the candidate event as the target event category of the event to be circulated based on the target set and the labeled event category of each candidate event; the platform determination module 34 includes a platform selection submodule for selecting a preset platform as the target platform based on the target event category and the first element value related to the location in the target set; wherein the configuration information of any preset platform includes at least the case category range and the case geographical range of the preset platform.

[0058] In some disclosed embodiments, the configuration information of any preset platform also includes the event receiving method of the preset platform, and the event to be transferred flows to the target platform according to the event receiving method defined in the configuration information of the target platform.

[0059] In some disclosed embodiments, the labeled event category includes at least a responsibility department, a responsibility item, and a specific classification. The event classification submodule includes a matching detection unit for detecting whether there is a second element value in the target set that matches any of the responsibility department, the responsibility item, and the specific classification in the labeled event category of the candidate event; the event classification submodule includes a category selection unit for selecting the labeled event category of the candidate event as the target event category in response to the existence of the second element value.

[0060] In some disclosed embodiments, the event transfer apparatus 30 further includes a reference update module for incorporating the event to be transferred as a new reference event into the reference event library in response to a completion instruction of the target platform for the event to be transferred.

[0061] See also Figure 4 , Figure 4: is a schematic diagram of the framework of an embodiment of an electronic device of the present application. The electronic device 40 includes at least a memory 41 and a processor 42 coupled to each other, the memory 41 stores at least program instructions, and the processor 42 is used to execute the program instructions to implement the steps in any of the above-mentioned event flow method embodiments. For details, please refer to the aforementioned disclosed embodiments, which will not be repeated here. As a possible example, the electronic device 40 may include but is not limited to a server, etc., and the specific type of the electronic device 40 is not limited here.

[0062] Specifically, the processor 42 is used to control itself and the memory 41 to implement the steps in any of the above-mentioned event flow method embodiments. The processor 42 can also be called a CPU (Central Processing Unit). The processor 42 may be an integrated circuit chip with signal processing capabilities. The processor 42 can also be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field-programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. In addition, the processor 42 can be implemented by an integrated circuit chip.

[0063] In the above scheme, the electronic device 40 extracts elements from the first description data of the event to be circulated based on different element extraction strategies, obtains several element sets, obtains common sets and non-common sets of the several element sets, and then obtains the target set based on the common set and the sampling set obtained by sampling the elements of the non-common set using the element sampling probability, thereby determining the target platform to which the event to be circulated is to be directed based on at least the first description data, the target set and the reference event library, and then responds to the return instruction of the target platform to the event to be circulated, reduces the element sampling probability, and the new element sampling probability at least performs element sampling on the new event to be circulated, so that one Compared with manual event transfer, event transfer can be realized through a series of operations such as feature extraction, feature sampling, and platform determination, which can improve the efficiency of event transfer. On the other hand, since the feedback instructions of the target platform for the transfer event are monitored in real time during the whole process of event transfer, and the feature sampling probability is timely reduced when the return instruction is received to reduce the sampling of features of non-common sets, the feature sampling probability can be adaptively updated and adjusted under the indirect intervention of the target platform, which helps to extract features from non-common sets as much as possible, so as to improve the accuracy of determining the target platform as much as possible, that is, to improve the accuracy of event transfer. Therefore, the accuracy and efficiency of event transfer can be improved.

[0064] See also Figure 5 , Figure 5 It is a schematic diagram of the framework of an embodiment of the event flow system of the present application. The event flow system 50 includes: a system front desk 51, a system backend 52, a convergence platform 53 and several preset platforms 54 connected in communication. The system front desk 51 is used to receive events reported by user terminals (not shown), and the system backend 52 is used to store events and forward events to the convergence platform 53. The convergence platform 53 at least includes the electronic device in the above-mentioned electronic device embodiment. As a possible example, the system front desk 51 can be provided with one; or multiple, such as a system front desk 51 for receiving events reported by APP channels and a system front desk 51 for receiving events reported by web channels, which are not limited here. The system backend 52 can be provided with one; or multiple, such as multiple system backends 52 can be provided in order to achieve load sharing. The convergence platform 53 can be provided with one; or multiple, such as multiple convergence platforms 53 can be provided in order to achieve load sharing. Of course, the above examples are only several possible situations in the actual application process, and they are not given one by one here.

[0065] In the above scheme, the event flow system 50 includes: a system front end 51, a system back end 52, a convergence platform 53 and several preset platforms 54 connected in communication, the system front end 51 is used to receive events reported by user terminals, the system back end 52 is used to store events and forward events to the convergence platform 53, and the convergence platform 53 at least includes the electronic device in the above electronic device embodiment. Therefore, compared with the manual realization of event flow, the event flow is realized through a series of operations such as element extraction, element sampling, and platform determination, which can improve the efficiency of event flow. On the other hand, since the feedback instructions of the target platform for the flow event are monitored in real time during the whole process of event flow, and the element sampling probability is timely reduced when the return instruction is received to reduce the element sampling of the non-commonality set, the element sampling probability can be adaptively updated and adjusted under the indirect intervention of the target platform, which helps to extract elements in the non-commonality set as much as possible, so as to improve the accuracy of determining the target platform as much as possible, that is, to improve the accuracy of event flow. Therefore, the accuracy and efficiency of event flow can be improved.

[0066] See also Figure 6 , Figure 6 Schematic diagram of a framework of an embodiment of a computer-readable storage medium 60 of the present application. The computer-readable storage medium 60 stores program instructions 61 that can be executed by a processor, and the program instructions 61 are used to implement the steps in any of the above event flow method embodiments.

[0067] In the above scheme, the computer-readable storage medium 60 extracts elements from the first description data of the event to be circulated based on different element extraction strategies, obtains several element sets, obtains common sets and non-common sets of the several element sets, and then obtains the target set based on the common set and the sampling set obtained by sampling the elements of the non-common set using the element sampling probability, thereby determining the target platform to which the event to be circulated is to be directed based on at least the first description data, the target set and the reference event library, and then responds to the return instruction of the target platform for the event to be circulated, reduces the element sampling probability, and the new element sampling probability at least samples the elements of the new event to be circulated. Therefore, compared with the manual realization of event flow, event flow can be realized through a series of operations such as feature extraction, feature sampling, and platform determination, which can improve the efficiency of event flow. On the other hand, since the feedback instructions of the target platform for the flow event are monitored in real time during the whole process of event flow, and the feature sampling probability is timely reduced when the return instruction is received to reduce the sampling of features of non-common sets, the feature sampling probability can be adaptively updated and adjusted under the indirect intervention of the target platform, which helps to extract features in non-common sets as much as possible, so as to improve the accuracy of determining the target platform as much as possible, that is, to improve the accuracy of event flow. Therefore, the accuracy and efficiency of event flow can be improved.

[0068] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the method described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0069] The above description of various embodiments tends to emphasize the differences between the various embodiments. The same or similar aspects can be referenced to each other, and for the sake of brevity, they will not be repeated herein.

[0070] In the several embodiments provided in the present application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device implementation described above is only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0071] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0072] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0073] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of each implementation method of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

[0074] If the technical solution of this application involves personal information, the product using the technical solution of this application has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the technical solution of this application involves sensitive personal information, the product using the technical solution of this application has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, the personal information processing rules are notified by obvious signs / information, and the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

Claims

1. An event flow method, characterized in that: include: Based on different element extraction strategies, elements are extracted from the first description data of the flow event respectively to obtain a number of element sets; Acquire a commonality set and at least one non-commonality set of the plurality of element sets; Obtain a target set based on the commonality set and a sampling set obtained by sampling elements of the non-commonality set using element sampling probability; Determine a target platform to which the event to be transferred is to be transferred based at least on the first description data, the target set and the reference event library; In response to the return instruction of the target platform for the event to be transferred, the element sampling probability is reduced, and the new element sampling probability at least performs element sampling on the new event to be transferred.

2. The method according to claim 1, characterized in that: The plurality of element sets include a first element set extracted based on a first extraction strategy and a second element set extracted based on a second extraction strategy, and obtaining a commonality set and at least one non-commonality set of the plurality of element sets includes: Acquire a commonality set of the first element set and the second element set, acquire a first non-commonality set formed by element values ​​that exist in the first element set but do not exist in the second element set, and acquire a second non-commonality set formed by element values ​​that do not exist in the first element set but exist in the second element set; The method of obtaining a target set by sampling elements of the commonality set and the non-commonality set using element sampling probability based on the commonality set comprises: The target set is obtained based on the commonality set, a first sampling set obtained by sampling elements of the first non-commonality set using a first sampling probability, and a second sampling set obtained by sampling elements of the second non-commonality set using a second sampling probability.

3. The method according to claim 1 or 2, characterized in that: The element set includes element values ​​of several element tags in the first description data, and the method further includes: Detecting whether the element sets extracted by each of the element extraction strategies completely cover the several element labels; In response to the fact that the element set extracted by any of the element extraction strategies completely covers the several element labels, the element sampling probability corresponding to the element extraction strategy is increased.

4. The method according to claim 3, characterized in that The increasing the sampling probability of the element corresponding to the element extraction strategy includes: Based on the current average number of daily events, a probability increase amplitude is obtained; wherein the probability increase amplitude is negatively correlated with the current average number of daily events; The probability increase amplitude is added to the current sampling probability of the element in the element extraction strategy.

5. The method according to claim 1, characterized in that Any of the element sets includes element values ​​of several element tags in the first description data. In the case where the return instruction is received, the method further includes: The plurality of element tags are re-determined based on the return reason parsed from the return instruction.

6. The method according to claim 1, characterized in that The reference event library includes a plurality of reference events, and determining a target platform to which the event to be transferred is to be transferred based at least on the first description data, the target set and the reference event library, includes: Selecting the reference event as a candidate event based on the similarity between the first description data and the second description data of each reference event; Based on the target set and the labeled event categories of each of the candidate events, selecting the labeled event category of the candidate event as the target event category of the event to be transferred; Based on the target event category and the first element value related to the location in the target set, a preset platform is selected as the target platform; wherein the configuration information of any preset platform at least includes the case category range and the case geographical range of the preset platform.

7. The method according to claim 6, characterized in that The configuration information of any of the preset platforms also includes the event receiving method of the preset platform, and the event to be transferred flows to the target platform according to the event receiving method defined in the configuration information of the target platform.

8. The method according to claim 6, characterized in that The marked event category at least includes a responsible department, a responsibility item, and a specific classification. The selecting the marked event category of the candidate event as the target event category of the event to be transferred based on the target set and the marked event category of each candidate event includes: Detecting whether there is a second element value in the target set that matches any one of the responsible department, the responsible item, and the specific classification in the labeled event category of the candidate event; In response to the existence of the second factor value, the labeled event category of the candidate event is selected as the target event category.

9. The method according to claim 1, characterized in that: The method further comprises: In response to the target platform's instruction to complete the event to be transferred, the event to be transferred is incorporated into the reference event library as a new reference event.

10. An event transfer device, characterized in that: include: An element extraction module is used to extract elements from the first description data of the flow event based on different element extraction strategies to obtain a plurality of element sets; A set processing module, used for obtaining a common set and at least one non-common set of the plurality of element sets; An element sampling module, configured to obtain a target set based on the commonality set and a sampling set obtained by sampling elements of the non-commonality set using element sampling probability; A platform determination module, configured to determine a target platform to which the event to be transferred is to be transferred based at least on the first description data, the target set and the reference event library; The probability updating module is used to reduce the element sampling probability in response to the return instruction of the target platform to the event to be transferred, and the new element sampling probability at least performs element sampling on the new event to be transferred.

11. An electronic device, characterized in that: The method comprises at least a memory and a processor coupled to each other, wherein the memory at least stores program instructions, and the processor is used to execute the program instructions to implement the event flow method according to any one of claims 1 to 9.

12. An event flow system, characterized in that: It includes a system foreground, a system background, a convergence platform and several preset platforms that are communicatively connected, the system foreground is used to receive events reported by user terminals, the system background is used to store events and forward events to the convergence platform, and the convergence platform at least includes the electronic device as described in claim 11.

13. A computer-readable storage medium, characterized in that: Program instructions that can be executed by a processor are stored, and the program instructions are used to implement the event flow method described in any one of claims 1 to 9.

Images