A data processing method, device, computer equipment and storage medium

Abstract

The embodiment of the present application provides a data processing method, device, computer equipment and storage medium, which relates to the field of data processing technology. The method includes: when a target event is detected, real-time feature information associated with the target event is obtained; offline feature information associated with the target event in the most recent offline feature update cycle is obtained; based on the obtained real-time feature information and offline feature information, a feature snapshot of the target event is generated; based on the feature snapshot of the target event, the associated matters of the target event are processed. In the embodiment of the present application, feature information is extracted based on feature snapshots, and rich feature information can be extracted from feature snapshots, which reduces the cost of collecting feature information during the processing process, shortens the collection cycle of feature information, and improves data processing efficiency.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a data processing method, apparatus, computer equipment and storage medium. Background Art

[0002] With the rapid development of the Internet, more and more Internet products are being used by people, such as online car-hailing platforms. Among them, passengers can request car-hailing services by operating online car-hailing platforms, and drivers can provide car-hailing services through online car-hailing platforms.

[0003] In practice, when passengers operate online ride-hailing platforms to request ride-hailing services or when drivers provide ride-hailing services through online ride-hailing platforms, operations on online ride-hailing platforms usually generate corresponding events, and online ride-hailing platforms need to process the generated events to provide better services to passengers or drivers. At present, online ride-hailing platforms usually handle events as follows: collect feature information corresponding to the same event within a period of time, and then process the corresponding event based on the collected feature information corresponding to the same event.

[0004] However, in the above processing method, the collected feature information is limited, the collection cost of the feature information is high, and the collection cycle is long, which leads to low event processing efficiency. Summary of the invention

[0005] In view of this, the purpose of the embodiments of the present application is to provide a data processing method, apparatus, computer equipment and storage medium, which can obtain rich feature information, reduce the cost of collecting feature information, shorten the collection cycle of feature information, and improve data processing efficiency.

[0006] In a first aspect, an embodiment of the present application provides a data processing method, including:

[0007] When a target event is detected, real-time feature information associated with the target event is acquired;

[0008] Obtain offline feature information associated with the target event within a recent offline feature update cycle;

[0009] Generate a feature snapshot of the target event based on the acquired real-time feature information and the offline feature information;

[0010] Based on the characteristic snapshot of the target event, related matters of the target event are processed.

[0011] In combination with the first aspect, an embodiment of the present application provides a first possible implementation of the first aspect, wherein the processing of related matters of the target event based on the feature snapshot of the target event includes:

[0012] Based on the feature snapshot of the target event, a target model for processing the target event is trained or updated.

[0013] In combination with the first possible implementation of the first aspect, the embodiment of the present application provides a second possible implementation of the first aspect, wherein the training or updating of the target model for processing the target event based on the feature snapshot of the target event includes:

[0014] After predicting target feature information that affects the target model used to process the target event, selecting historical target feature information corresponding to the target event in a first historical time period from the feature snapshot corresponding to the target event;

[0015] Based on the selected historical target feature information, a target model for processing the target event is trained or updated to obtain a new target model.

[0016] In combination with the first aspect, the embodiment of the present application provides a third possible implementation of the first aspect, wherein the processing of the related matters of the target event based on the feature snapshot of the target event includes:

[0017] Acquire real-time feature information and offline feature information of the target event in a second historical time period from the feature snapshot of the target event;

[0018] Based on the acquired real-time feature information and the offline feature information, abnormal information corresponding to the target event is determined.

[0019] In combination with the first aspect, the embodiment of the present application provides a fourth possible implementation of the first aspect, wherein the processing of the associated matters of the target event based on the feature snapshot of the target event includes:

[0020] Acquire real-time feature information and offline feature information of the target event in a third historical time period from the feature snapshot of the target event;

[0021] For each indicator associated with the target event, statistics are collected on the acquired real-time feature information and the offline feature information according to a preset statistical period to obtain statistical data of the indicator in each statistical period;

[0022] According to the statistical data of the indicator in each statistical period, a change trend graph of the target event under the indicator is generated.

[0023] In combination with the second possible implementation of the first aspect, the embodiment of the present application provides a fifth possible implementation of the first aspect, wherein, after obtaining the new target model, the method further includes:

[0024] Evaluating the performance indicators of the new target model;

[0025] If the performance indicator of the new target model meets the preset conditions, then when a new target event is detected, the new target event is processed using the new target model.

[0026] In combination with the fifth possible implementation of the first aspect, the embodiment of the present application provides a sixth possible implementation of the first aspect, wherein the target event includes a bubble event; the target model for processing the bubble event includes one or more of the following models: a queuing model, a time estimation model, a dynamic price adjustment model, and a dynamic discount model;

[0027] The processing of the new target event by using the new target model includes:

[0028] Use the new queuing model to predict the queuing information of the new bubbling event, use the new time estimation model to predict the destination arrival time of the new bubbling event, use the new dynamic pricing model to predict the price information of the new bubbling event, and use the new dynamic discount model to predict the discount information of the new bubbling event.

[0029] In combination with the fifth possible implementation of the first aspect, the embodiment of the present application provides a seventh possible implementation of the first aspect, wherein the processing of the new target event by using the new target model includes:

[0030] Use the new accountability model to judge and handle new complaint incidents.

[0031] In combination with the fifth possible implementation of the first aspect, the embodiment of the present application provides an eighth possible implementation of the first aspect, wherein the processing of the new target event by using the new target model includes:

[0032] Use the new order splitting model to split new order splitting events.

[0033] In combination with the first aspect, the embodiment of the present application provides a ninth possible implementation of the first aspect, wherein the generating a feature snapshot of the target event based on the acquired real-time feature information and the offline feature information specifically includes:

[0034] Acquire basic event information corresponding to the target event; wherein the basic event information includes target event identification information, area identification information corresponding to the target event, target event occurrence time information; and at least one of the following information: service requester identification information, service provider identification information;

[0035] Acquire the real-time feature information corresponding to each piece of identification information in the basic information of the event and the offline feature information in the most recent offline feature update cycle;

[0036] A feature snapshot of the event is generated based on the target event occurrence time information, the acquired real-time feature information, and the acquired offline feature information.

[0037] In combination with the sixth possible implementation of the first aspect, the embodiment of the present application provides a tenth possible implementation of the first aspect, wherein the bubbling event includes: an event start point and an event end point;

[0038] The real-time feature information corresponding to the bubble event includes one or more of the following feature information: current area information where the bubble event occurs, current area information and current environment information of the event starting point, current area information and current environment information of the event end point, service provider information around the area where the event starting point is located, service requester information around the area where the event starting point is located, service provider information around the area where the event end point is located, and service requester information around the area where the event end point is located; wherein the environment information includes weather information and road condition information.

[0039] In combination with the seventh possible implementation of the first aspect, an embodiment of the present application provides an eleventh possible implementation of the first aspect, wherein the real-time characteristic information corresponding to the complaint event includes one or more of the following characteristic information: complaint work order information, complaint order information, service provider information, service requester information, detour information, drunk information, abnormal stop information, abnormal deviation information, and high-risk driving information.

[0040] In a second aspect, an embodiment of the present application further provides a data processing device, including:

[0041] A first acquisition module, used for acquiring real-time feature information associated with a target event when a target event is detected;

[0042] A second acquisition module is used to acquire offline feature information associated with the target event within a recent offline feature update cycle;

[0043] A generating module, configured to generate a feature snapshot of the target event based on the acquired real-time feature information and the offline feature information;

[0044] A processing module is used to process related matters of the target event based on the characteristic snapshot of the target event.

[0045] In combination with the second aspect, the embodiment of the present application provides a first possible implementation of the second aspect, wherein the processing module is specifically used to:

[0046] Based on the feature snapshot of the target event, a target model for processing the target event is trained or updated.

[0047] In combination with the first possible implementation of the second aspect, the embodiment of the present application provides a second possible implementation of the second aspect, wherein the processing module is specifically used to:

[0048] After predicting target feature information that affects the target model used to process the target event, selecting historical target feature information corresponding to the target event in a first historical time period from the feature snapshot corresponding to the target event;

[0049] Based on the selected historical target feature information, a target model for processing the target event is trained or updated to obtain a new target model.

[0050] In combination with the second aspect, the embodiment of the present application provides a third possible implementation of the second aspect, wherein the processing module is specifically used to:

[0051] Acquire real-time feature information and offline feature information of the target event in a second historical time period from the feature snapshot of the target event;

[0052] Based on the acquired real-time feature information and the offline feature information, abnormal information corresponding to the target event is determined.

[0053] In combination with the second aspect, the embodiment of the present application provides a fourth possible implementation of the second aspect, wherein the processing module is specifically used to:

[0054] Acquire real-time feature information and offline feature information of the target event in a third historical time period from the feature snapshot of the target event;

[0055] For each indicator associated with the target event, statistics are collected on the acquired real-time feature information and the offline feature information according to a preset statistical period to obtain statistical data of the indicator in each statistical period;

[0056] According to the statistical data of the indicator in each statistical period, a change trend graph of the target event under the indicator is generated.

[0057] In combination with the second possible implementation of the second aspect, the embodiment of the present application provides a fifth possible implementation of the second aspect, wherein the data processing device further includes:

[0058] An evaluation module, used for evaluating the performance indicators of the new target model;

[0059] The processing module is further configured to process the new target event using the new target model when a new target event is detected to occur, if the performance indicator of the new target model meets a preset condition.

[0060] In combination with the fifth possible implementation of the second aspect, the embodiment of the present application provides a sixth possible implementation of the second aspect, wherein the target event includes a bubble event; the target model for processing the bubble event includes one or more of the following models: a queuing model, a time estimation model, a dynamic price adjustment model, and a dynamic discount model;

[0061] The processing module is specifically used to use a new queuing model to predict the queuing information of a new bubble event, use a new time estimation model to predict the destination arrival time of a new bubble event, use a new dynamic pricing model to predict the price information of a new bubble event, and use a new dynamic discount model to predict the discount information of a new bubble event.

[0062] In combination with the fifth possible implementation of the second aspect, the embodiment of the present application provides a seventh possible implementation of the second aspect, wherein:

[0063] The processing module is specifically used to use the new accountability model to conduct accountability processing on new complaint events.

[0064] In combination with the fifth possible implementation of the second aspect, the embodiment of the present application provides an eighth possible implementation of the second aspect, wherein:

[0065] The processing module is specifically used to process new order splitting events using a new order splitting model.

[0066] In combination with the second aspect, the embodiment of the present application provides a ninth possible implementation of the second aspect, wherein the generating module is specifically used to:

[0067] Acquire basic event information corresponding to the target event; wherein the basic event information includes target event identification information, area identification information corresponding to the target event, target event occurrence time information; and at least one of the following information: service requester identification information, service provider identification information;

[0068] Acquire the real-time feature information corresponding to each piece of identification information in the basic information of the event and the offline feature information in the most recent offline feature update cycle;

[0069] A feature snapshot of the event is generated based on the target event occurrence time information, the acquired real-time feature information, and the acquired offline feature information.

[0070] In combination with the sixth possible implementation of the second aspect, the embodiment of the present application provides a tenth possible implementation of the second aspect, wherein the bubbling event includes: an event start point and an event end point;

[0071] The real-time feature information corresponding to the bubble event includes one or more of the following feature information: current area information where the bubble event occurs, current area information and current environment information of the event starting point, current area information and current environment information of the event end point, service provider information around the area where the event starting point is located, service requester information around the area where the event starting point is located, service provider information around the area where the event end point is located, and service requester information around the area where the event end point is located; wherein the environment information includes weather information and road condition information.

[0072] In combination with the seventh possible implementation of the second aspect, an embodiment of the present application provides an eleventh possible implementation of the second aspect, wherein the real-time feature information corresponding to the complaint event includes one or more of the following feature information: complaint work order information, complaint order information, service provider information, service requester information, detour information, drunk information, abnormal stop information, abnormal deviation information, and high-risk driving information.

[0073] In a third aspect, an embodiment of the present application further provides a computer device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor communicates with the memory through the bus, and the processor executes the machine-readable instructions to perform the steps of the data processing method as described in any one of the first aspects.

[0074] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the data processing method as described in any one of the first aspects are executed.

[0075] The embodiment of the present application provides a data processing method, device, computer equipment and storage medium. When a target event occurs, by obtaining the real-time feature information of the target event and the offline feature information associated with the target event in the most recent offline feature update cycle, a feature snapshot of the target event is produced, and related matters of the target event are processed based on the produced feature snapshot. In the embodiment of the present application, feature information is extracted based on the feature snapshot, and rich feature information can be extracted from the feature snapshot, which reduces the cost of collecting feature information during the processing process, shortens the collection cycle of feature information, and improves data processing efficiency.

[0076] Furthermore, when the associated item of the target event is a model, if the model needs to be updated, the prior art needs to collect relevant features based on the running model, which requires the model to be frequently online and offline, resulting in poor stability of the model; while in the embodiment of the present application, a feature snapshot method is adopted, which can directly obtain the features required for updating the model from the feature snapshot of the target event, and update the model based on the obtained features. In this way, the model only needs to be online and offline when updated, which reduces the number of times the model is online and offline, thereby ensuring the stability of the model.

[0077] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0079] Figure 1 A flow chart of a data processing method provided in an embodiment of the present application is shown.

[0080] Figure 2 A flow chart of another data processing method provided in an embodiment of the present application is shown.

[0081] Figure 3 A schematic diagram of the structure of a snapshot platform provided in an embodiment of the present application is shown.

[0082] Figure 4 A flow chart of another data processing method provided in an embodiment of the present application is shown.

[0083] Figure 5 A flow chart of another data processing method provided in an embodiment of the present application is shown.

[0084] Figure 6 A flow chart of another data processing method provided in an embodiment of the present application is shown.

[0085] Figure 7 A flow chart of another data processing method provided in an embodiment of the present application is shown.

[0086] Figure 8 A schematic diagram of the structure of a data processing device provided in an embodiment of the present application is shown.

[0087] Fig. 9A schematic diagram of the structure of a computer device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0088] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0089] Taking into account the problem that the online ride-hailing platform's method of handling events has the disadvantages of limited feature information collected, high feature information collection costs and long collection cycles, which in turn leads to low event processing efficiency, the embodiments of the present application improve a data processing method, apparatus, computer equipment and storage medium, which can obtain rich feature information, reduce the feature information collection costs, shorten the feature information collection cycle, and improve data processing efficiency.

[0090] The data processing method and device provided in the embodiments of the present application are described below in conjunction with the field of online car-hailing.

[0091] like Figure 1 As shown, a data processing method is provided for an embodiment of the present application, and the method can be applied to a server, and the method specifically includes:

[0092] S101. When a target event is detected to occur, real-time feature information associated with the target event is acquired.

[0093] In the embodiment of the present application, when the service requester triggers a target event at the first user terminal, the first user terminal will send target event information to the server, and after receiving the target event information from the first user terminal, the server detects the occurrence of the corresponding target event. Alternatively, when the service provider triggers an event at the second user terminal, the second user terminal will send target event information to the server, and after receiving the target event information from the second user terminal, the server detects the occurrence of the corresponding target event. The above-mentioned service requester is a passenger, and the corresponding first user terminal is a passenger terminal; the above-mentioned service provider is a driver, and the corresponding second user terminal is a driver terminal.

[0094] As an implementation method, the above-mentioned target event can be a bubble event, an order splitting event, a complaint event, or a security event; here, the security event includes but is not limited to: abnormal stay events, high-risk order events, abnormal relationship events, and abnormal preference events.

[0095] In the field of online car-hailing, bubble events: When passengers use online car-hailing, they need to enter the starting and ending locations of the online car-hailing. After the passenger enters the end location of the online car-hailing on the passenger side, the passenger side will send the end location of the online car-hailing to the server. At this time, the bubble event is triggered. Order splitting event: When the passenger confirms the order on the passenger side, the passenger side generates the corresponding order information and sends it to the server. At this time, the order splitting event is triggered. Complaint event: When the passenger triggers a complaint about the target order on the passenger side, the passenger side generates the complaint information of the target order and sends it to the server. At this time, the complaint event is triggered; or, after the driver triggers a complaint about the target order on the driver side, the driver side generates the complaint information of the target order and sends it to the server. At this time, the complaint event is triggered.

[0096] In the embodiment of the present application, the above-mentioned target event has real-time feature information; wherein different target events correspond to different real-time feature information.

[0097] S102: Obtain offline feature information associated with the target event within a recent offline feature update cycle.

[0098] In the embodiment of the present application, the storage system corresponding to the server pre-stores the offline feature information of the target event. As an implementation method, the storage system can be a storage medium in the server; or it can be an independent storage system, and the server can access the independent storage system and obtain the offline feature information in the storage system.

[0099] The server will periodically update the offline feature information associated with the target event in the storage system; for example, the server updates the offline feature information associated with the target event every day, or the server updates the offline feature information associated with the target event every week. Correspondingly, when the target event occurs, the server obtains the offline feature information associated with the target event within the most recent offline feature update cycle.

[0100] For example, the server updates the offline feature information associated with the target event at 24:00 every day. Assuming that the target event occurs at 10:00 am on January 2, 2019, the server obtains the offline feature information associated with the target event updated at 24:00 on January 1, 2019.

[0101] In the embodiment of the present application, the offline characteristic information associated with the target event includes one or more of the following information associated with the target event: offline characteristic information of the service requester, offline characteristic information of the service provider, economic conditions of the target area, categories of points of interest (PoI) in the target area, and prosperity of the target area. Here, the offline characteristic information of the service requester includes: the name of the passenger, the age of the passenger, the age and family situation of the passenger, etc.; the offline characteristic information of the service provider includes: the name of the driver, the age of the driver, the family situation of the driver, the work and rest time of the driver, the length of service information of the driver, etc.

[0102] S103: Generate a feature snapshot of the target event based on the acquired real-time feature information and the offline feature information.

[0103] In an embodiment of the present application, after obtaining the real-time feature information and offline feature information associated with the target event, the server writes the real-time feature information and offline feature information associated with the target event into a corresponding storage system as a feature snapshot of the target event.

[0104] S104: Processing related matters of the target event based on the feature snapshot of the target event.

[0105] In an embodiment of the present application, processing matters related to the target event may specifically include: training or updating a target model used to process the target event, replaying scenes of the target event in any historical time period, constructing a simulation environment based on specific feature information corresponding to the target event, and changing trends of indicators related to the target event.

[0106] The data processing method provided in the embodiment of the present application extracts feature information based on feature snapshots, can extract rich feature information from feature snapshots, reduce the cost of collecting feature information during the processing, shorten the feature information collection cycle, and improve data processing efficiency.

[0107] Furthermore, when the associated item of the target event is a model, if the model needs to be updated, the prior art needs to collect relevant features based on the running model, which requires the model to be frequently online and offline, resulting in poor stability of the model; while in the embodiment of the present application, a feature snapshot method is adopted, which can directly obtain the features required for updating the model from the feature snapshot of the target event, and update the model based on the obtained features. In this way, the model only needs to be online and offline when updated, which reduces the number of times the model is online and offline, thereby ensuring the stability of the model.

[0108] In the embodiment of the present application, different target events correspond to different real-time feature information. The real-time feature information corresponding to the target event is described below in conjunction with some specific target events:

[0109] 1. For bubbling events:

[0110] A bubble event includes: an event start point and an event end point; the real-time feature information corresponding to the bubble event includes one or more of the following feature information: current area information where the bubble event occurs, current area information and current environment information of the event start point, current area information and current environment information of the event end point, service provider information around the area where the event start point is located, service requester information around the area where the event start point is located, service provider information around the area where the event end point is located, and service requester information around the area where the event end point is located; wherein the environment information includes weather information and road condition information.

[0111] Among them, the current area information where the bubble event occurs includes: the current area identification number (identification, ID) where the bubble event occurs, whether there are large-scale activities in the current area, the road conditions information of the current area, the weather information of the current area, the drivers who are in the waiting state within 500 meters around, the passengers who are in the issuing state within 500 meters around, etc.; the current area information and current environmental information of the starting point of the event, namely: whether there are large-scale activities in the current area, the road conditions information of the current area, the weather information of the current area, etc.; the current area information and current environmental information of the end point of the event, namely: whether there are large-scale activities in the current area, the road conditions information of the current area, the weather information of the current area, etc. The service provider information around the area where the starting point of the event is located includes: the number of drivers who are waiting for orders within 500 meters around the area where the starting point of the event is located, the total number of drivers within 500 meters around the area where the starting point of the event is located, etc.; the service requester information around the area where the starting point of the event is located is: the number of passengers who are issuing orders within 500 meters around the area where the starting point of the event is located, and the total number of passengers within 500 meters around the area where the starting point of the event is located; the service provider information around the area where the end point of the event is located is: the number of drivers who are waiting for orders within 500 meters around the end point of the event is located, and the total number of drivers within 500 meters around the end point of the event; the service requester information around the area where the end point of the event is located is: the number of passengers who are issuing orders within 500 meters around the end point of the event is located, and the total number of passengers within 500 meters around the end point of the event.

[0112] 2. Regarding the order issuance event:

[0113] The real-time characteristic information corresponding to the above-mentioned order issuance event includes one or more of the following characteristic information: current area information where the order issuance event occurs, order information, environmental information of the order starting point, environmental information of the order ending point, service provider information around the area where the order starting point is located, service requester information around the area where the order starting point is located, service provider information around the area where the order ending point is located, service requester information around the area where the order ending point is located, and order issuance reason information.

[0114] 3. Regarding complaints:

[0115] The real-time characteristic information corresponding to the above-mentioned complaint events includes one or more of the following characteristic information: complaint work order information, complaint order information, service provider information, service requester information, detour information, drunk information, abnormal stop information, abnormal deviation information, and high-risk driving information.

[0116] Further, such as Figure 2 As shown, in the data processing method provided in the embodiment of the present application, based on the acquired real-time feature information and the offline feature information, a feature snapshot of the target event is generated, which specifically includes:

[0117] S201. Obtain basic event information corresponding to the target event; wherein the basic event information includes target event identification information, area identification information corresponding to the target event, target event occurrence time information; and at least one of the following information: service requester identification information, service provider identification information.

[0118] In an embodiment of the present application, the above-mentioned target event includes: target event identification information, area identification information corresponding to the target event, target event occurrence time information, and service requester identification information; or, the above-mentioned target event includes: target event identification information, area identification information corresponding to the target event, target event occurrence time information, and service provider identification information; or, the above-mentioned target event includes: target event identification information, area identification information corresponding to the target event, target event occurrence time information, service requester identification information, and service provider identification information.

[0119] As an implementation method, the target event is a bubbling event, and the bubbling event includes a starting position and an end position. Accordingly, the bubbling event includes: a bubbling event ID, a starting position ID, an end position, an occurrence time information of the bubbling event, and passenger identification information.

[0120] S202: Acquire real-time feature information corresponding to each piece of identification information in the basic information of the event and offline feature information in a recent offline feature update cycle.

[0121] In the embodiment of the present application, taking the bubble event as an example, the real-time feature information corresponding to the starting position ID includes: the current traffic information of the starting position, weather information, whether there are large-scale activities, the number of passengers in the order-issuing state around the starting position, the total number of passengers around the starting position, the number of drivers in the order-listening state around the starting position, and the total number of drivers around the starting position; the information of passengers in the order-issuing state around the terminal position, the total number of passengers around the terminal position, the information of drivers in the order-listening state around the terminal position, and the total number of drivers around. Among them, the above-mentioned area around the starting position refers to the preset range of the starting position, such as within 500 meters around the starting position; correspondingly, the above-mentioned area around the terminal position refers to the preset range of the terminal position, such as within 500 meters around the terminal position.

[0122] S203: Generate a feature snapshot of the event based on the target event occurrence time information, the acquired real-time feature information, and the acquired offline feature information.

[0123] In the embodiment of the present application, the server pre-establishes the association between the target event identification information, the target event occurrence time information and the area identification information corresponding to the target event, and then writes the real-time feature information and offline feature information corresponding to each of the acquired identification information into the corresponding storage system to obtain the feature snapshot of the target event. Figure 3 A schematic diagram of the structure of the snapshot platform provided in the embodiments of the present application is shown.

[0124] The following describes the feature snapshot of the target event generated in the embodiment of the present application in conjunction with a specific application scenario.

[0125] Application scenario 1

[0126] With the rapid development of the Internet, artificial intelligence technology has also developed rapidly, and machine learning, as an important research field in artificial intelligence, has also developed rapidly. Models trained by machine learning are widely used in various scenarios, such as chatbots, reading comprehension models, speech recognition models, text recognition models, dynamic pricing models, etc.; and the above-mentioned various scenarios are also widely used in online car-hailing platforms. In an embodiment of the present application, the server can train or update the target model used to process the above-mentioned target event based on the feature snapshot of the target event.

[0127] like Figure 4 As shown, based on the feature snapshot of the target event, training or updating the target model for processing the target event includes:

[0128] S401. After predicting target feature information that affects a target model used to process the target event, selecting historical target feature information corresponding to the target event in a first historical time period from feature snapshots corresponding to the target event.

[0129] As an implementation method, after receiving the target feature information input by the staff that can affect the target model used to process the target event, the server obtains a feature snapshot corresponding to the target event, and selects the historical target feature information corresponding to the above target event within the first historical time period from the feature snapshot.

[0130] For example, for a bubble event, the server predicts that the target feature of "the number of drivers waiting for orders within 500 meters around the starting location of the bubble event" is beneficial to the dynamic pricing model used to process bubble events. Accordingly, the server obtains the feature snapshot corresponding to the bubble event, and selects multiple historical target feature information within six months before the current time from the obtained feature snapshot; the above multiple historical target feature information are: the number of drivers waiting for orders within 500 meters around the starting location of each bubble event.

[0131] S402: Based on the selected historical target feature information, a target model for processing the target event is trained or updated to obtain a new target model.

[0132] Here, after selecting the number of drivers in the waiting state within 500 meters around the starting position of each bubble event within six months before the current time, the server trains the original model for processing bubble events based on the selected historical target feature information to obtain a trained dynamic pricing model; or, the server updates the dynamic pricing model for processing bubble events based on the selected historical target feature information to obtain an updated dynamic pricing model. The above new target model is the trained dynamic pricing model, or the updated dynamic pricing model.

[0133] At present, when updating a model, it is necessary to put the model online and collect feature information corresponding to the target event within a period of time based on the online model. After the feature information is collected, the model is taken offline and updated based on the collected features. If the performance indicators of the updated model meet the preset conditions, the updated model is put online; if the performance indicators of the updated model do not meet the preset conditions, the original model is put online and the feature or other features are continued to be collected to update the model. However, this method requires the model to be constantly online and offline, resulting in poor model stability.

[0134] In the embodiment of the present application, a feature snapshot method is adopted, which can directly obtain the feature information required for updating the model from the feature snapshot of the target event, and update the model based on the obtained feature information. In this way, the model only needs to be online and offline when updated, which reduces the number of times the model is online and offline, thereby ensuring the stability of the model.

[0135] like Figure 5 As shown, in the embodiment of the present application, after obtaining the new target model, the method further includes:

[0136] S501: Evaluate the performance indicators of the new target model.

[0137] In the embodiment of the present application, the target model corresponding to the target event is usually a regression model. For example, the dynamic pricing model corresponding to the bubble event is a regression model. The performance indicators of the regression model include mean absolute error (MAE), mean square error (MSE), etc.

[0138] S502: If the performance indicator of the new target model meets the preset condition, when a new target event is detected, the new target event is processed using the new target model.

[0139] As an implementation mode, when the server determines that the mean absolute error (MAE) of the trained dynamic pricing model is less than the error threshold and the mean square error (MSE) is less than the variance threshold, it determines that the performance indicators of the trained dynamic pricing model meet the preset conditions; accordingly, the server puts the trained dynamic pricing model online to process the new target event through the dynamic pricing model.

[0140] As another implementation, when the server determines that the mean absolute error (MAE) of the updated dynamic pricing model is less than the error threshold and the mean variance (MSE) is less than the variance threshold, it determines that the performance indicators of the updated dynamic pricing model meet the preset conditions; accordingly, the server takes offline the dynamic pricing model before the update and puts online the updated dynamic pricing model. After the server puts online the updated dynamic pricing model, when a new bubble event is detected, the server uses the online updated dynamic pricing model to predict the price information of the new bubble event.

[0141] Furthermore, in the data processing method provided in the embodiment of the present application, the target event includes a bubble event; the target model for processing the bubble event includes one or more of the following models: a queuing model, a time estimation model, a dynamic price adjustment model, and a dynamic discount model;

[0142] Processing the new target event using the new target model includes:

[0143] Use the new queuing model to predict the queuing information of the new bubbling event, use the new time estimation model to predict the destination arrival time of the new bubbling event, use the new dynamic pricing model to predict the price information of the new bubbling event, and use the new dynamic discount model to predict the discount information of the new bubbling event.

[0144] In an embodiment of the present application, after a passenger triggers a new bubble event on the passenger side, the passenger side generates corresponding bubble event information and inputs the bubble event information into the queuing model, time estimation model, dynamic price adjustment model and dynamic discount model respectively. Accordingly, the above models output the queuing information, destination arrival time, price information and discount information of the bubble event in turn.

[0145] The above queue information is the current queue position of the new bubble event, for example, the current queue position of the new bubble event is 16th. The above destination arrival time is the estimated destination arrival time of the new bubble event, for example, the occurrence time of the new bubble event is 10:00 am on January 2, 2019, and the corresponding destination arrival time is 11:00 am on January 2, 2019. The above price information is the estimated price corresponding to the new bubble event, for example, 52 yuan. The above discount information is the estimated discount price corresponding to the new bubble event, for example, 5 yuan.

[0146] Furthermore, in the data processing method provided in the embodiment of the present application, the new target event is processed using the new target model, including:

[0147] Use the new accountability model to judge and handle new complaint incidents.

[0148] In the embodiment of the present application, the above-mentioned target event includes a complaint event, and the target model for processing the complaint event includes a judgment model. As an implementation method, if a passenger complains about a target order on the passenger side, the passenger side generates complaint information of the target order and sends it to the server, and the server inputs the complaint information of the target order into the judgment model and outputs the judgment result.

[0149] Furthermore, in the data processing method provided in the embodiment of the present application, the new target event is processed using the new target model, including:

[0150] Use the new order splitting model to split new order splitting events.

[0151] In the embodiment of the present application, the target event includes a split order event, and the target model for processing the split order event includes a split order model. As an implementation method, if the passenger confirms the order issuance at the passenger end, the passenger end generates corresponding order issuance information and sends it to the server, and the server inputs the received order issuance information into the split order model and outputs the split order result.

[0152] Application scenario 2

[0153] like Figure 6 As shown, in the data processing method provided in the embodiment of the present application, based on the characteristic snapshot of the target event, the associated matters of the target event are processed, including:

[0154] S601: Acquire real-time feature information and offline feature information of the target event in a second historical time period from a feature snapshot of the target event.

[0155] S602: Determine abnormal information corresponding to the target event based on the acquired real-time feature information and the offline feature information.

[0156] In combination with step 601 and step 602, as an implementation method, if a passenger makes a complaint regarding a target event, the server obtains a feature snapshot of the target event, and obtains real-time feature information and offline feature information of the feature snapshot, and determines abnormal information from the real-time feature information and offline feature information, so as to determine a solution to the current complaint based on the abnormal information.

[0157] The first historical time period and the second historical time period may be the same or different.

[0158] Application scenario three

[0159] like Figure 7 As shown, in the data processing method provided in the embodiment of the present application, based on the characteristic snapshot of the target event, the associated matters of the target event are processed, including:

[0160] S701: Acquire real-time feature information and offline feature information of the target event in a third historical time period from the feature snapshot of the target event.

[0161] In the embodiment of the present application, the first historical time period, the second historical time period and the third historical time period may be the same or different.

[0162] S702: For each indicator associated with the target event, statistics are collected on the acquired real-time feature information and the offline feature information according to a preset statistical period to obtain statistical data of the indicator in each statistical period.

[0163] As an implementation method, the server counts the number of drivers in the waiting state around area A every hour. The server counts the real-time feature information and the offline feature information within six months before the current time to obtain the number of drivers in the waiting state around area A every hour within six months before the current time.

[0164] S703: Generate a change trend graph of the target event under the indicator according to the statistical data of the indicator in each statistical period.

[0165] As an implementation method, the server generates a changing trend graph of the target event under the above-mentioned indicator based on the number of drivers in the waiting state around area A in the six months before the current time and the corresponding time information, so that the staff can observe the changing trend of the bubbling event under the indicator.

[0166] Among them, the indicators associated with the target events can be used for composition analysis, trend analysis, construction and optimization of target models for processing target events, indicator evaluation and root cause analysis, etc.

[0167] A data processing method provided in an embodiment of the present application extracts feature information based on feature snapshots, can extract rich feature information from feature snapshots, reduce the cost of collecting feature information during the processing process, shorten the feature information collection cycle, and improve data processing efficiency.

[0168] Furthermore, when the associated item of the target event is a model, if the model needs to be updated, the prior art needs to collect relevant features based on the running model, which requires the model to be frequently online and offline, resulting in poor stability of the model; while in the embodiment of the present application, a feature snapshot method is adopted, which can directly obtain the features required for updating the model from the feature snapshot of the target event, and update the model based on the obtained features. In this way, the model only needs to be online and offline when updated, which reduces the number of times the model is online and offline, thereby ensuring the stability of the model.

[0169] like Figure 8 As shown, an embodiment of the present application provides a data processing device for processing the above data processing method, the device comprising:

[0170] The first acquisition module 801 is used to acquire real-time feature information associated with the target event when the target event is detected.

[0171] The second acquisition module 802 is used to acquire offline feature information associated with the target event within the most recent offline feature update cycle;

[0172] A generating module 803, configured to generate a feature snapshot of the target event based on the acquired real-time feature information and the offline feature information;

[0173] The processing module 804 is used to process the related matters of the target event based on the characteristic snapshot of the target event.

[0174] Furthermore, in the data processing device provided in the embodiment of the present application, the processing module 804 is specifically used for:

[0175] Based on the feature snapshot of the target event, a target model for processing the target event is trained or updated.

[0176] Furthermore, in the data processing device provided in the embodiment of the present application, the processing module 804 is specifically used for:

[0177] After predicting target feature information that affects the target model used to process the target event, selecting historical target feature information corresponding to the target event in a first historical time period from the feature snapshot corresponding to the target event;

[0178] Based on the selected historical target feature information, a target model for processing the target event is trained or updated to obtain a new target model.

[0179] Furthermore, in the data processing device provided in the embodiment of the present application, the processing module 804 is specifically used for:

[0180] Acquire real-time feature information and offline feature information of the target event in a second historical time period from the feature snapshot of the target event;

[0181] Based on the acquired real-time feature information and the offline feature information, abnormal information corresponding to the target event is determined.

[0182] Furthermore, in the data processing device provided in the embodiment of the present application, the processing module 804 is specifically used for:

[0183] Acquire real-time feature information and offline feature information of the target event in a third historical time period from the feature snapshot of the target event;

[0184] For each indicator associated with the target event, statistics are collected on the acquired real-time feature information and the offline feature information according to a preset statistical period to obtain statistical data of the indicator in each statistical period;

[0185] According to the statistical data of the indicator in each statistical period, a change trend graph of the target event under the indicator is generated.

[0186] Furthermore, the data processing device provided in the embodiment of the present application also includes an evaluation module; wherein,

[0187] An evaluation module, used for evaluating the performance indicators of the new target model;

[0188] The processing module 804 is further configured to process the new target event using the new target model when a new target event is detected to occur, if the performance indicator of the new target model meets a preset condition.

[0189] Further, in the data processing device provided in the embodiment of the present application, the target event includes a bubble event; the target model for processing the bubble event includes one or more of the following models: a queuing model, a time estimation model, a dynamic price adjustment model, and a dynamic discount model;

[0190] Processing module 804 is specifically used to use the new queuing model to predict the queuing information of the new bubble event, use the new time estimation model to predict the destination arrival time of the new bubble event, use the new dynamic pricing model to predict the price information of the new bubble event, and use the new dynamic discount model to predict the discount information of the new bubble event.

[0191] Furthermore, in the data processing device provided in the embodiment of the present application, the processing module 804 is specifically used to use the new accountability model to perform accountability processing on new complaint events.

[0192] Furthermore, in the data processing device provided in the embodiment of the present application, the processing module 804 is specifically used to perform order splitting processing on new order splitting events using a new order splitting model.

[0193] Furthermore, in the data processing device provided in the embodiment of the present application, the generating module 803 is specifically used for:

[0194] Acquire basic event information corresponding to the target event; wherein the basic event information includes target event identification information, area identification information corresponding to the target event, target event occurrence time information; and at least one of the following information: service requester identification information, service provider identification information;

[0195] Acquire the real-time feature information corresponding to each piece of identification information in the basic information of the event and the offline feature information in the most recent offline feature update cycle;

[0196] A feature snapshot of the event is generated based on the target event occurrence time information, the acquired real-time feature information, and the acquired offline feature information.

[0197] Further, in the data processing device provided in the embodiment of the present application, the bubbling event includes: an event start point and an event end point;

[0198] The real-time feature information corresponding to the bubble event includes one or more of the following feature information: current area information where the bubble event occurs, current area information and current environment information of the event starting point, current area information and current environment information of the event end point, service provider information around the area where the event starting point is located, service requester information around the area where the event starting point is located, service provider information around the area where the event end point is located, and service requester information around the area where the event end point is located; wherein the environment information includes weather information and road condition information.

[0199] Furthermore, in the data processing device provided in the embodiment of the present application, the real-time characteristic information corresponding to the complaint event includes one or more of the following characteristic information: complaint work order information, complaint order information, service provider information, service requester information, detour information, drunk information, abnormal stop information, abnormal deviation information, and high-risk driving information.

[0200] The data processing device provided in the embodiment of the present application, when a target event occurs, makes a feature snapshot of the target event by acquiring the real-time feature information of the target event and the offline feature information associated with the target event in the most recent offline feature update cycle, and processes the associated matters of the target event based on the produced feature snapshot. In the embodiment of the present application, feature information is extracted based on the feature snapshot, and rich feature information can be extracted from the feature snapshot, which reduces the cost of collecting feature information during the processing process, shortens the collection cycle of feature information, and improves the processing efficiency of data.

[0201] Furthermore, when the associated item of the target event is a model, if the model needs to be updated, the prior art needs to collect relevant features based on the running model, which requires the model to be frequently online and offline, resulting in poor stability of the model; while in the embodiment of the present application, a feature snapshot method is adopted, which can directly obtain the features required for updating the model from the feature snapshot of the target event, and update the model based on the obtained features. In this way, the model only needs to be online and offline when updated, which reduces the number of times the model is online and offline, thereby ensuring the stability of the model.

[0202] like Fig. 9 As shown, a computer device 90 provided in an embodiment of the present application includes: a processor 901, a memory 902 and a bus, wherein the memory 902 stores machine-readable instructions executable by the processor 901. When the computer device is running, the processor 901 communicates with the memory 902 through the bus, and the processor 901 executes the machine-readable instructions to perform the steps of the above-mentioned data processing method.

[0203] Specifically, the above-mentioned memory 902 and processor 901 can be general-purpose memories and processors, which are not specifically limited here. When the processor 901 runs the computer program stored in the memory 902, the above-mentioned data processing method can be executed.

[0204] Corresponding to the above-mentioned data processing method, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned data processing method are executed.

[0205] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0206] In the embodiment of the present application, the terms "passenger" and "service requester" are used interchangeably to refer to a person, entity or tool that can request or order a service. In the embodiment of the present application, the terms "driver" and "service provider" are used interchangeably to refer to a person, entity or tool that can provide a service. In the embodiment of the present application, "first user end" and "passenger end" are used interchangeably, and "second user end" and "driver end" are used interchangeably.

[0207] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working process of the system and device described above can refer to the corresponding process in the method embodiment, and will not be repeated in this application. In the several embodiments provided in this application, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules 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 communication interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0208] The modules described as separate components may or may not be physically separated, and the components shown as modules 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 solution of this embodiment.

[0209] 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.

[0210] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various storage media that can store program codes, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks.

[0211] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

Claims

1. A data processing method, characterized in that: include: When a target event is detected, real-time feature information associated with the target event is acquired; Obtain offline feature information associated with the target event within a recent offline feature update cycle; Generate a feature snapshot of the target event based on the acquired real-time feature information and the offline feature information; Based on the characteristic snapshot of the target event, processing the related matters of the target event; The processing of the related items of the target event specifically includes: training or updating the target model used to process the target event, replaying the scene of the target event in any historical time period, constructing a simulation environment based on the specific feature information corresponding to the target event, and the change trend of the indicators related to the target event; The step of generating a feature snapshot of the target event based on the acquired real-time feature information and the offline feature information specifically includes: Acquire basic event information corresponding to the target event; wherein the basic event information includes target event identification information, area identification information corresponding to the target event, target event occurrence time information; and at least one of the following information: service requester identification information, service provider identification information; Acquire the real-time feature information corresponding to each piece of identification information in the basic information of the event and the offline feature information in the most recent offline feature update cycle; A feature snapshot of the event is generated based on the target event occurrence time information, the acquired real-time feature information, and the acquired offline feature information.

2. The data processing method according to claim 1, characterized in that: The training or updating of a target model for processing the target event based on the feature snapshot of the target event includes: After predicting target feature information that affects the target model used to process the target event, selecting historical target feature information corresponding to the target event in a first historical time period from the feature snapshot corresponding to the target event; Based on the selected historical target feature information, a target model for processing the target event is trained or updated to obtain a new target model.

3. The data processing method according to claim 1, characterized in that: The processing of the related matters of the target event based on the characteristic snapshot of the target event also includes: Acquire real-time feature information and offline feature information of the target event in a second historical time period from the feature snapshot of the target event; Based on the acquired real-time feature information and the offline feature information, abnormal information corresponding to the target event is determined.

4. The data processing method according to claim 1, characterized in that: The processing of the related matters of the target event based on the characteristic snapshot of the target event also includes: Acquire real-time feature information and offline feature information of the target event in a third historical time period from the feature snapshot of the target event; For each indicator associated with the target event, statistics are collected on the acquired real-time feature information and the offline feature information according to a preset statistical period to obtain statistical data of the indicator in each statistical period; According to the statistical data of the indicator in each statistical period, a change trend graph of the target event under the indicator is generated.

5. The data processing method according to claim 2, characterized in that: After obtaining the new target model, the method further includes: Evaluating the performance indicators of the new target model; If the performance indicator of the new target model meets the preset conditions, then when a new target event is detected, the new target event is processed using the new target model.

6. The data processing method according to claim 5, characterized in that: The target event includes one or more of the following events: bubbling event, complaint event, order splitting event; The target model corresponding to the bubbling event includes one or more of the following models: a queuing model, a time estimation model, a dynamic price adjustment model, and a dynamic discount model; The target model corresponding to the complaint event includes a responsibility judgment model; The target model corresponding to the order splitting event includes an order splitting model.

7. The data processing method according to claim 6, characterized in that: The bubbling event includes: an event start point and an event end point; The real-time feature information corresponding to the bubble event includes one or more of the following feature information: current area information where the bubble event occurs, current area information and current environment information of the event starting point, current area information and current environment information of the event end point, service provider information around the area where the event starting point is located, service requester information around the area where the event starting point is located, service provider information around the area where the event end point is located, and service requester information around the area where the event end point is located; wherein the environment information includes weather information and road condition information.

8. A data processing device, characterized in that: include: A first acquisition module, used for acquiring real-time feature information associated with a target event when a target event is detected; A second acquisition module is used to acquire offline feature information associated with the target event within a recent offline feature update cycle; A generating module, configured to generate a feature snapshot of the target event based on the acquired real-time feature information and the offline feature information; A processing module, used for processing related matters of the target event based on the characteristic snapshot of the target event; The processing module is specifically used to: train or update the target model used to process the target event, replay the scene of the target event in any historical time period, build a simulation environment based on the specific feature information corresponding to the target event, and the change trend of the target event-related indicators; Wherein, the generation module is specifically used for: Acquire basic event information corresponding to the target event; wherein the basic event information includes target event identification information, area identification information corresponding to the target event, target event occurrence time information; and at least one of the following information: service requester identification information, service provider identification information; Acquire the real-time feature information corresponding to each piece of identification information in the basic information of the event and the offline feature information in the most recent offline feature update cycle; A feature snapshot of the event is generated based on the target event occurrence time information, the acquired real-time feature information, and the acquired offline feature information.

9. A computer device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of the data processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the data processing method according to any one of claims 1 to 7 are executed.

Images