User breakpoint management method, device, equipment and computer program product

The spark streaming technology obtains and processes user behavior data, accurately determines the time of user breakpoints, solves the problems of inaccurate detection of breakpoints in the existing technology and is limited in data processing in concurrent scenarios, and realizes accurate breakpoints in high concurrent scenarios Test.

CN113342614BActive Publication Date: 2024-11-08WEBANK (CHINA)
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110730473.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-06-29
Publication Date
2024-11-08
Estimated Expiration
2041-06-29

AI Technical Summary

Technical Problem

In the high concurrency real-time data processing scenario, existing user breakpoint detection technology cannot accurately determine the time point of the user breakpoint, and in the high concurrency scenario, the time window length is limited, and a large amount of data flow cannot be effectively processed.

Method used

The user behavior data is obtained through the first spark streaming, determine the latest behavior time and current summary behavior data of each user, and put it into the breakpoint detection queue, and transfer it to the delay queue after the breakpoint duration, and then pass the second spark streaming obtains past summary behavior data. If the latest occurrence time is the same as the latest behavior time, it is determined that the user has a breakpoint.

Benefits of technology

It realizes the accurate determination of user breakpoint time in high concurrency real-time data processing scenarios, improves the accuracy of breakpoint detection, and avoids the problem of time window length limitation due to memory limitations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113342614B_ABST
    Figure CN113342614B_ABST
Patent Text Reader

Abstract

The present invention discloses a user breakpoint management method, device, equipment and computer program product, the method comprising: obtaining a user behavior data set through a first spark streaming, determining the latest occurrence time of each user's recent behavior and the current summary behavior data according to the user behavior data set, so that the current summary behavior data can continue to be in the breakpoint detection queue after the breakpoint duration; using a second spark streaming to obtain the past summary behavior data in the breakpoint detection queue, if the latest occurrence time in the past summary behavior data is the same as the latest occurrence time of the corresponding user, the user generates a breakpoint and the generation time is the latest occurrence time. The present invention detects user behavior data by using spark streaming data flow real-time computing technology, and can accurately determine the time when the user generates a breakpoint in a high-concurrency real-time data processing scenario.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of financial technology (Fintech), and in particular to a user breakpoint management method, device, equipment and computer program product. Background Art

[0002] With the development of computer technology, more and more technologies (big data, distributed, blockchain, artificial intelligence, etc.) are applied in the financial field. The traditional financial industry is gradually transforming to financial technology (Fintech). However, due to the security and real-time requirements of the financial industry, higher requirements are also placed on user breakpoint detection technology.

[0003] More and more business services are being realized online through the Internet, which facilitates people's lives. When the business system provides various business services to users, business breakpoints may occur due to various reasons. Therefore, it is necessary to accurately determine whether the user has a breakpoint.

[0004] At present, in the process of user breakpoint management, the common breakpoint detection method is: using Spark Streaming time window sliding, setting a time window length, and detecting users who have breakpoints during this period by counting the users who slide out of the window. However, one of the defects of this method is that the user who slides out of the window may have a breakpoint at any time point within the time window length. This solution can only determine that the user has a breakpoint within the window time period, but cannot accurately determine the time point when the breakpoint occurs; the defects of this method also include: the Spark Streaming time window sliding solution needs to read all data streams in the window into the memory. Since the system memory cannot be expanded infinitely, the length of the time window is limited in high concurrency scenarios. Summary of the invention

[0005] The main purpose of the present invention is to propose a user breakpoint management method, device, equipment and computer program product, aiming to solve the technical problem of how to improve the accuracy of detecting user breakpoints in high-concurrency real-time data processing scenarios.

[0006] To achieve the above object, the present invention provides a user breakpoint management method, which comprises the following steps:

[0007] Acquire a user behavior data set in a preprocessing queue at a first preset frequency through a first spark streaming, and determine the latest occurrence time of each user's most recent behavior and current aggregated behavior data of each user according to the user behavior data set;

[0008] Put the current aggregated behavior data of each user into a breakpoint detection queue, and transfer the current aggregated behavior data of each user put into the breakpoint detection queue to a delay queue, which starts when the current aggregated behavior data of each user is received and transfers the current aggregated behavior data of each user to the breakpoint detection queue after the breakpoint duration.

[0009] Obtaining the breakpoint detection queue at a second preset frequency through a second spark streaming, and putting past summary behavior data before the breakpoint duration at the current moment;

[0010] If the latest occurrence time in the past aggregated behavior data is the same as the latest occurrence time of the recent behavior of the user corresponding to the past aggregated behavior data, then a breakpoint of the user corresponding to the past aggregated behavior data is determined, and the latest occurrence time in the past aggregated behavior data is used as the time when the breakpoint of the user occurs.

[0011] Optionally, the step of determining the latest occurrence time of each user's most recent behavior and each user's current aggregated behavior data according to the user behavior data set includes:

[0012] Sort the user behavior data of each user in the user behavior data set according to the occurrence time, and determine the latest occurrence time of the most recent behavior of each user;

[0013] The total behavior information of each user is generated according to the behavior information in the user behavior data of each user, and the current summary behavior data of each user is obtained according to the total behavior information of each user and the latest occurrence time of the latest behavior of each user.

[0014] Optionally, after the step of determining a breakpoint of the user corresponding to the past aggregated behavior data, the method further includes:

[0015] When a user recall request is detected, determining the number of recall messages sent to the user within the preset configured number of days closest to the current moment;

[0016] If the number is less than a preset threshold, a recall message is sent to the user.

[0017] Optionally, the step of determining the number of recall messages sent to the user within the preset configured number of days closest to the current moment includes:

[0018] Determine the real-time sending index of the current date and the summary sending index of each date before the current date, and determine the historical sending summary according to the summary sending index of each date before the current date and the preset configuration days;

[0019] The number of recall messages sent to the user within the preset configured number of days closest to the current moment is determined according to the historical sending summary and the real-time sending index of the current date.

[0020] Optionally, before the step of determining the aggregated sending index of each date before the current date, the step further includes:

[0021] Obtain a data array with a preset length, obtain the real-time sending index of the day before the current date and the batch sending index of the day before the current date;

[0022] Determine the summary sending index of the day before the current date according to the real-time sending index of the day before the current date and the batch sending index;

[0023] Determine, according to the date of the day before the current date and the preset value, a corresponding data node in the data array of the summary sending indicator of the day before the current date;

[0024] Put the summary sending index of the day before the current date into the corresponding data node in the data array;

[0025] The step of determining the summary sending index of each date before the current date and determining the historical sending summary according to the summary sending index of each date before the current date includes:

[0026] The summary sending index of each date before the current date is determined according to the data array, and the historical sending summary is determined according to the summary sending index of each date before the current date and the preset configuration days.

[0027] Optionally, after the step of determining the real-time sending index within the current date, the method further includes:

[0028] Determine whether the summary sending index of the day before the current date has been obtained within the current date;

[0029] If yes, the steps are executed: determining the summary sending index of each date before the current date, and determining the historical sending summary according to the summary sending index of each date before the current date;

[0030] If not, determining the historical sending summary based on the summary sending index of each date before the day before the current date, and determining the historical sending summary based on the summary sending index of each date before the day before the current date;

[0031] The number of recall messages sent to the user within the preset configured number of days from the current moment is determined according to the historical sending summary, the real-time sending index within the current date and the real-time sending index of the day before the current date.

[0032] Optionally, the step of determining the real-time sending index of the current date includes:

[0033] Obtaining the node of the current day and the node of the previous day, and when the first recall message is successfully sent to the user within the current date, using the data in the node of the current day to overwrite the data in the node of the previous day;

[0034] Clear the data in the current day node, and record the real-time sending index of the current date in the current day node;

[0035] When the recall information is successfully sent to the user again within the current date, the real-time sending indicator of the current date is updated.

[0036] In addition, to achieve the above object, the present invention further provides a user breakpoint management device, the user breakpoint management device comprising:

[0037] A first monitoring module is used to obtain a user behavior data set in a preprocessing queue at a first preset frequency through a first spark streaming, and determine the latest occurrence time of each user's most recent behavior and current aggregated behavior data of each user according to the user behavior data set;

[0038] A data delay module, used to put the current aggregated behavior data of each user into a breakpoint detection queue, and transfer the current aggregated behavior data of each user put into the breakpoint detection queue to a delay queue, wherein the delay queue starts from when the current aggregated behavior data of each user is received, and transfers the current aggregated behavior data of each user to the breakpoint detection queue after the breakpoint duration;

[0039] A second monitoring module is used to obtain the past summary behavior data before the breakpoint duration at the current moment from the breakpoint detection queue through a second spark streaming at a second preset frequency;

[0040] A breakpoint judgment module is used to determine that a breakpoint has occurred for the user corresponding to the past aggregated behavior data if the latest occurrence time in the past aggregated behavior data is the same as the latest occurrence time of the most recent behavior of the user corresponding to the past aggregated behavior data, and to use the latest occurrence time in the past aggregated behavior data as the time when the breakpoint has occurred for the user.

[0041] In addition, to achieve the above-mentioned purpose, the present invention also provides a user breakpoint management device, which includes: a memory, a processor, and a user breakpoint management program stored in the memory and executable on the processor, and the user breakpoint management program implements the steps of the user breakpoint management method described above when executed by the processor.

[0042] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer storage medium, on which a user breakpoint management program is stored, and when the user breakpoint management program is executed by a processor, the steps of the user breakpoint management method described above are implemented.

[0043] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer program product, which includes a user breakpoint management program, and when the user breakpoint management program is executed by a processor, the steps of the user breakpoint management method described above are implemented.

[0044] The present invention obtains a user behavior data set through the first spark streaming, determines the latest occurrence time of each user's recent behavior and the current summary behavior data based on the user behavior data set, so that the current summary behavior data can continue to be in the breakpoint detection queue after the breakpoint duration; the past summary behavior data in the breakpoint detection queue is obtained by the second spark streaming. If the latest occurrence time in the past summary behavior data is the same as the latest occurrence time of the corresponding user, the user generates a breakpoint and the generation time is the latest occurrence time. The present invention detects user behavior data by using spark streaming data flow real-time computing technology, and can accurately determine the time when the user generates a breakpoint in a high-concurrency real-time data processing scenario. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the structure of a user breakpoint management device in a hardware operating environment involved in an embodiment of the present invention;

[0046] Figure 2 A schematic diagram of a flow chart of a first embodiment of a user breakpoint management method of the present invention;

[0047] Figure 3 A schematic diagram of a scenario of a data array in the user breakpoint management method of the present invention;

[0048] Figure 4 A schematic diagram of a scenario in which the data node of the array data is used to store the summary sending index of each date in the user breakpoint management method of the present invention;

[0049] Figure 5 A schematic diagram of a scenario of the current day node and the previous day node in the user breakpoint management method of the present invention;

[0050] Figure 6 It is a schematic diagram of a scenario in which a node of the current day and a node of the previous day are flipped when a recall instruction is successfully sent to a user for the first time within the current date in the user breakpoint management method of the present invention;

[0051] Figure 7 It is a module schematic diagram of the user breakpoint management device of the present invention.

[0052] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0053] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0054] like Figure 1 As shown, Figure 1 It is a schematic diagram of the structure of a user breakpoint management device in a hardware operating environment involved in an embodiment of the present invention.

[0055] The user breakpoint management device in the embodiment of the present invention may be a PC or a server device, on which a virtual machine runs.

[0056] like Figure 1 As shown, the user breakpoint management device may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or it may be a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0057] Those skilled in the art will understand that Figure 1 The user breakpoint management device structure shown in the figure does not constitute a limitation of the device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0058] like Figure 1 As shown, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a user breakpoint management program.

[0059] exist Figure 1 In the user breakpoint management device shown, the network interface 1004 is mainly used to connect to the background server and communicate data with the background server; the user interface 1003 is mainly used to connect to the client (user end) and communicate data with the client; and the processor 1001 can be used to call the user breakpoint management program stored in the memory 1005 and execute the operations in the following user breakpoint management method.

[0060] Based on the above hardware structure, an embodiment of the user breakpoint management method of the present invention is proposed.

[0061] Reference Figure 2 , Figure 2 This is a flow chart of a first embodiment of a user breakpoint management method of the present invention, the method comprising:

[0062] Step S10, obtaining a user behavior data set in a preprocessing queue at a first preset frequency through a first spark streaming, and determining the latest occurrence time of each user's most recent behavior and current aggregated behavior data of each user according to the user behavior data set;

[0063] The user breakpoint management method of this embodiment is applied to a user breakpoint management device, which may be a terminal, a robot, or a PC device.

[0064] In the prior art, in the process of user breakpoint management, a common breakpoint detection method is: using Spark Streaming time window sliding, setting a time window length, and detecting users who have breakpoints during this period by counting the users who slide out of the window. However, one of the defects of this method is that the user who slides out of the window may have a breakpoint at any time point within the time window length. This solution can only determine that the user has a breakpoint within the window time period, and cannot accurately determine the time point when the breakpoint occurs; the defects of this method also include: the Spark Streaming time window sliding solution needs to read all data streams in the window into the memory. Since the system memory cannot be expanded infinitely, the length of the time window is limited in high concurrency scenarios.

[0065] In this context, in order to solve the technical problem of how to improve the accuracy of detecting user breakpoints in high-concurrency real-time data processing scenarios, this embodiment improves a user breakpoint management solution, obtains a user behavior data set through the first spark streaming, determines the latest occurrence time of each user's recent behavior and the current summary behavior data based on the user behavior data set, so that the current summary behavior data can continue to be in the breakpoint detection queue after the breakpoint duration; uses the second spark streaming to obtain the past summary behavior data in the breakpoint detection queue, if the latest occurrence time in the past summary behavior data is the same as the latest occurrence time of the corresponding user, the user generates a breakpoint and the generation time is the latest occurrence time. The present invention detects user behavior data by using spark streaming data flow real-time computing technology, and can accurately determine the time when the user generates a breakpoint in a high-concurrency real-time data processing scenario.

[0066] In this embodiment, the user's behavior of accessing the product front end generates a large amount of behavior data. In some marketing scenarios, it is necessary to pay attention to when the user stops using the product. Therefore, it is necessary to determine when the user has a breakpoint through the behavior data generated by the user on the product front end, so as to reach the user in time for recall. However, the user should not be reached too frequently to avoid disturbing the user's peaceful life.

[0067] When a user uses a product, if the user does not perform any operation on the product within a specified time, the user is defined as having a breakpoint. The corresponding user is defined as the user who has generated a breakpoint.

[0068] When users visit the product front end, a large amount of behavior data will be generated. These behavior data will be written into the pre-processing queue through the data collection system. The organization of user behavior data (MESSAGE_DATA) is as follows:

[0069] {"USER_ID":"xx","BUSSINESS_DATA":"xx","CREATED_DATETIME":"xx"}.

[0070] Among them, "USER_ID" is the user ID. Different user IDs are used to represent different users. "BUSSINESS_DATA" is a certain behavior information of the user, indicating the specific information of the user's behavior, including the port information of the user to access the product, the specific operation information of the user on the product front end, etc. "CREATED_DATETIME" indicates the time when the behavior occurs, that is, the time when the user performs the behavior.

[0071] After the behavior data is written into the pre-processing queue, there may be multiple user behavior data with the same user ID but different behavior information and behavior occurrence time in the pre-processing queue.

[0072] Afterwards, the user behavior data set in the preprocessing queue is obtained at a first preset frequency through the first spark streaming, that is, the preprocessing queue is monitored by the first spark streaming application, and the first spark streaming application pulls the message in the preprocessing queue at a time interval, that is, the user behavior data set is obtained. The user behavior data set includes one or more user behavior data of one or more users.

[0073] Spark is a distributed computing framework similar to MapReduce. Its core is a resilient distributed dataset. It provides a richer model than MapReduce and can quickly iterate datasets in memory multiple times to support complex data mining algorithms and graph computing algorithms. Spark streaming is a real-time computing framework built on Spark that extends Spark's ability to process large-scale streaming data.

[0074] Determine all user behavior data of each user, and determine the user behavior data with the latest behavior occurrence time among all user behavior data of each user based on the behavior occurrence time in the user behavior data. The user behaviors represented by these user behavior data are the most recent behaviors of each user before the current moment. These user behaviors are the most recent behaviors of each user, and the occurrence time of each user's most recent behavior is the latest occurrence time of each user, that is, the latest occurrence time of each user's most recent behavior.

[0075] At the same time, the user behavior data belonging to the same user ID in the user behavior data set obtained by the first spark streaming are combined into a summary behavior data, that is, all user behavior data belonging to the same user ID in the user behavior data set are integrated into a current summary behavior data of the user. The current summary behavior data of a user includes the user ID of the user and the latest occurrence time of the user's latest behavior.

[0076] Furthermore, in the above step S10, the step of determining the latest occurrence time of each user's recent behavior and each user's current aggregated behavior data according to the user behavior data set includes:

[0077] Step S101, sorting the user behavior data of each user in the user behavior data set according to the occurrence time, and determining the latest occurrence time of the latest behavior of each user;

[0078] Step S102, generating total behavior information of each user according to the behavior information in the user behavior data of each user, and obtaining current summary behavior data of each user according to the total behavior information of each user and the latest occurrence time of the most recent behavior of each user.

[0079] In this embodiment, all user behavior data of each user in the user behavior data set are determined, and the user behavior data of each user is sorted in reverse order of the behavior occurrence time, that is, the behavior represented by the user behavior data sorted first is the latest behavior, that is, the behavior is the most recent behavior to the current moment. The behavior occurrence time in the user behavior data ranked first is taken, and the behavior occurrence time is the latest occurrence time of the user's recent behavior.

[0080] It should be noted that the user behavior data of each user can also be sorted in the positive order of the time when the behavior occurred, that is, the behavior represented by the user behavior data sorted at the end is the latest, that is, the behavior is the behavior closest to the current moment. The behavior occurrence time in the last user behavior data is taken, and the behavior occurrence time is the latest occurrence time of the user's recent behavior.

[0081] At the same time, the MESSAGE_DATA (user behavior data) in the acquired user behavior data set is merged into the first spark streaming data organization unit RDD according to the user ID.

[0082] Specifically, the behavior information in all user behavior data corresponding to each user is integrated, and the behavior information "BUSSINESS_DATA" in the user behavior data is concatenated and merged to form a total behavior information "BUSSINESS_DATA_LIST". The current summary behavior data of each user is determined based on the total behavior information of each user and the latest occurrence time of the latest behavior, and the current summary behavior data is stored in the first spark streaming data organization unit RDD. The organization form of the current summary behavior data in RDD is as follows:

[0083] {"USER_ID":"xx","BUSSINESS_DATA_LIST":"xx","LASTEST_OCCUR_DATETIME":"xx"}.

[0084] Among them, "USER_ID" indicates the ID of a user, which is used to distinguish users, "BUSSINESS_DATA_LIST" indicates the total behavior information of the user, and "LASTEST_OCCUR_DATETIME": "xx" indicates the latest occurrence time of the user's most recent behavior. At this time, each user corresponds to a current summary behavior data.

[0085] Step S20, putting the current aggregated behavior data of each user into a breakpoint detection queue, transferring the current aggregated behavior data of each user put into the breakpoint detection queue to a delay queue, the delay queue starts from when the current aggregated behavior data of each user is received, and transfers the current aggregated behavior data of each user to the breakpoint detection queue after the breakpoint duration;

[0086] In this embodiment, the obtained summary behavior data of each user and the latest occurrence time are put into the write breakpoint detection queue, and at the same time, the user ID and the latest occurrence time (USER_ID, LASTEST_OCCUR_DATETIME) in the summary behavior data are recorded in redis, which is a database.

[0087] When the broker of the breakpoint detection queue detects the current summary behavior data being written, it transfers the current summary behavior data written in the breakpoint detection queue to the delay queue. At this time, the current summary behavior data written this time does not exist in the breakpoint detection queue, but there may be past summary behavior data written previously. The broker is a message storage server that can transmit messages from the sender to the receiver.

[0088] The function of the delay queue is that when the delay queue receives the current summary behavior data transferred from the breakpoint detection queue, the delay queue starts a scheduled task. After the delay queue receives the current summary behavior data and the breakpoint duration interval has passed, the scheduled task takes the current summary behavior data out of the delay queue and writes it back to the original breakpoint detection queue. In this way, the current summary behavior data in the breakpoint detection queue can be consumed only after the breakpoint duration interval has passed, and the current summary behavior data becomes the past summary behavior data after being written back from the delay queue to the breakpoint detection queue.

[0089] Among them, the breakpoint duration is used to determine whether the user has generated a breakpoint. If the user has not performed any operation on the product within the specified time, it is defined that the user has generated a breakpoint. The length of this specified time is the breakpoint duration.

[0090] It should be noted that this embodiment does not limit the length of the breakpoint duration, and the breakpoint duration may be 0.5h, 24h, or a longer or shorter duration.

[0091] Step S30, obtaining the breakpoint detection queue at a second preset frequency through the second spark streaming, and putting the past summary behavior data before the breakpoint time from the current moment.

[0092] In this embodiment, since the current summary behavior data in the breakpoint detection queue can be consumed only after the breakpoint duration interval, the past summary behavior data before the breakpoint duration interval is pulled by the second spark streaming at a second preset frequency, that is, the past summary behavior data in the breakpoint detection queue is written before the breakpoint duration at the current moment by the second spark streaming at a second preset frequency. The second preset frequency may be the same as or different from the first preset frequency.

[0093] It should be noted that when the breakpoint detection queue receives the current summary behavior data, the current summary behavior data needs to be transferred to the delay queue, and can only be transferred from the delay queue to the breakpoint detection queue after the breakpoint duration has passed. Therefore, the summary behavior data in the breakpoint detection queue is generally written before the breakpoint duration from the current moment, and is generally past summary behavior data.

[0094] For example, if the breakpoint duration is 0.5h, data 1 was placed 1h before the current time, data 2 was placed 0.6h before the current time, and data 3 was placed 0.2h before the current time, then data 1 and data 2 have both been transferred from the breakpoint detection queue to the delay queue, and then from the delay queue to the breakpoint detection queue, and data 3 is in the delay queue at the current moment, so the past summary behavior data is data 1 and data 2.

[0095] In the above process, the flow order of each user's aggregated behavior data is:

[0096] Breakpoint detection queue push -> delay queue -> breakpoint detection queue pop → second spark streaming.

[0097] When the current summary behavior data of each user just enters the breakpoint detection queue, the current summary behavior data in the breakpoint detection queue is immediately transferred to the delay queue. At this time, there is no current summary behavior data in the breakpoint detection queue. After the breakpoint duration, the current summary behavior data in the delay queue is transferred back to the breakpoint detection queue. At this time, the current summary behavior data becomes the past summary behavior data. The second spark streaming reads the past summary behavior data in the breakpoint detection queue.

[0098] Step S40, if the latest occurrence time in the past aggregated behavior data is the same as the latest occurrence time of the recent behavior of the user corresponding to the past aggregated behavior data, then determine that a breakpoint has occurred for the user corresponding to the past aggregated behavior data, and use the latest occurrence time in the past aggregated behavior data as the time when the breakpoint has occurred for the user.

[0099] In this embodiment, in the acquired past aggregated behavior data, if the time point in a certain past aggregated behavior data, that is, the latest occurrence time in the past aggregated behavior data, is the same as the LASTEST_OCCUR_DATETIME (latest occurrence time) corresponding to the USER_ID (user ID) of the past aggregated behavior data in redis, it indicates the interval of the breakpoint duration, and the user has no new behavior on the product front end and is written to redis. Otherwise, the latest occurrence time of the behavior data corresponding to the user in redis will be updated. Therefore, it can be determined that the user corresponding to the past aggregated behavior data has a breakpoint. At this time, the time when the breakpoint occurred for the user is obtained, and the latest occurrence time in the past aggregated behavior data is used as the time when the breakpoint occurred for the user, that is, the latest occurrence time in the past aggregated behavior data is the time when the user has a breakpoint. After determining that the user has a breakpoint, a marketing recall can be conducted on the user.

[0100] By writing the user's real-time data into two message queues, and then using a spark streaming application to listen to the preprocessing queue, receiving the user behavior data, extracting the fields related to the breakpoint marketing rules in the data, and writing the latest timestamp of the user's behavior into the redis database, the received user behavior data is then sent to the breakpoint detection queue. After the breakpoint detection queue is delayed by a scheduled task, another spark streaming queue listens to the breakpoint detection queue, obtains the timestamp of the user behavior data from the breakpoint detection queue, and compares this timestamp with the timestamp recorded by the user in redis. If the two timestamps are consistent, the user has a breakpoint, and the time when the user's breakpoint occurs is obtained, that is, the two identical timestamps are the time when the user's breakpoint occurs.

[0101] This embodiment obtains a user behavior data set through the first spark streaming, determines the latest occurrence time of each user's recent behavior and the current summary behavior data based on the obtained user behavior data set, and puts the current summary behavior data into the breakpoint detection queue, and then transfers the current summary behavior data in the breakpoint detection queue to the delay queue, so that the current summary behavior data can be put back into the breakpoint detection queue after the breakpoint duration; uses the second spark streaming to obtain the past summary behavior data in the breakpoint detection queue, if the latest occurrence time in the past summary behavior data is the same as the latest occurrence time of the user corresponding to the past summary behavior data, it is determined that the user has a breakpoint, and the latest occurrence time in the past summary behavior data is used as the time when the user breaks. The present invention detects user behavior data by using the spark streaming data flow real-time computing technology, and can accurately determine the time when the user generates a breakpoint in a high-concurrency real-time data processing scenario.

[0102] Furthermore, based on the above embodiments, a second embodiment of the user breakpoint management method of the present invention is proposed.

[0103] In the above step S40, after the step of determining the breakpoint of the user corresponding to the past aggregated behavior data, the following step is further included:

[0104] Step S501, when a user recall request is detected, determining the number of recall messages sent to the user within the preset configured number of days closest to the current moment;

[0105] Step S502: If the number is less than a preset threshold, a recall message is sent to the user.

[0106] In this embodiment, after determining that a breakpoint occurs in a certain user, a marketing recall action will be performed on the user, and a recall message will be sent to the user.

[0107] When marketing to users, users should not be disturbed frequently. Therefore, a frequency control scheme is proposed as follows: First, the number of messages sent to a single user on the previous day (T-1 day) can be batch calculated at midnight on the current date (T day). Secondly, the number of messages sent to a single user on the current date (T day) is accumulated in real time on T day. Adding these two types of indicators according to the configured number of days is the total number of messages sent to the user within the configured number of days. Finally, compare this sum with the configured threshold. If this sum is greater than or equal to the configured threshold, it means that the frequency of sending messages within the configured number of days has reached the upper limit. The request to send messages will be intercepted in the system, thereby achieving the purpose of limiting the frequency of sending messages. Otherwise, it will not be intercepted, and a recall message will be sent to the user, and the number of messages sent on T day will be accumulated in real time for this message.

[0108] Specifically, when a user recall request is detected, the number of recall messages sent to the user with the breakpoint within the preset number of days closest to the current moment is determined. The current date of the current moment is also counted as one day. For example, if the current moment is 9:00 on the 23rd, and the preset number of days is 3 days, the number of recall messages sent to the user on the entire day of the 21st, the entire day of the 22nd, and between 0:00 and 9:00 on the 23rd will be determined.

[0109] The preset configuration days are used to control the frequency of sending recall information to users. The frequency of sending recall information to users is controlled by setting a certain threshold for the number of recall information sent to users within the preset configuration days.

[0110] If the number of recall messages sent to the user within the preset configured number of days closest to the current moment is less than the preset threshold, the user's recall request will not be intercepted, and the recall message will be sent to the user.

[0111] If the number of recall messages sent to the user within the preset configured number of days closest to the current moment is greater than or equal to the preset threshold, the user's recall request is intercepted and no recall message is sent to the user.

[0112] When a user recall request is detected, this embodiment will count the frequency of recall information sent to the user within the last preset configuration days. If too many recall messages are sent, the user recall request will be blocked to avoid frequent disturbing of the user, thereby improving the user experience.

[0113] Furthermore, based on the above embodiments, a third embodiment of the user breakpoint management method of the present invention is proposed.

[0114] In the above step S501, the step of determining the number of recall messages sent to the user within the preset configured number of days closest to the current moment includes:

[0115] Step S5011, determining the real-time sending index of the current date and the summary sending index of each date before the current date, and determining the historical sending summary according to the summary sending index of each date before the current date and the preset configuration days;

[0116] Step S5012: determining the number of recall messages sent to the user within the preset configured number of days closest to the current moment according to the historical sending summary and the real-time sending index of the current date.

[0117] In this embodiment, the step of determining the number of recall messages sent to the user within the preset configuration days closest to the current moment is specifically to determine the real-time sending index within the current date, where the current date refers to the natural day where the current moment is located, and the real-time sending index is the number of recall messages that have been sent to the user from the current date to the current moment within the current date. At the beginning of the current date, the real-time sending index is 0, and each time a recall message is successfully sent to the user, the value of the real-time sending index increases by 1 until the end of the current date, that is, the real-time sending index of the current date may change with the specific time of the current date.

[0118] At the same time, the summary sending index of each date before the current date is determined. The method for determining the summary sending index of each date before the current date can be to determine the summary sending index of the day before the current date within the current date. Since the real-time sending index of each natural day is determined within each natural day, after the real-time sending index of the whole day of the current day is counted, the summary sending index of the previous day will be determined on the next day based on the real-time sending index of the previous day and the batch sending index of the previous day. The batch sending index refers to the number of recall information sent in batches to a large number of users. The batch sending index is generally not counted in the real-time sending index, so it is necessary to count the summary sending index of the previous day on the next day based on the real-time sending index and the batch sending index of the previous day. The previous day and the next day are two consecutive days, such as yesterday and today.

[0119] Through the above method, the summary sending index of each date before the current date can be determined. After the preset configuration days A is determined, the historical sending summary can be determined according to the preset configuration days A and the summary sending index of each date before the current date. The method can be to determine the summary sending index of each date A-1 days before the current date, and add these summary sending indexes to obtain the historical sending summary.

[0120] After obtaining the historical sending summary, the historical sending summary is added to the real-time sending index of the current date to obtain the number of recall messages sent to the user within the preset configuration days closest to the current date.

[0121] That is, if the preset configuration days are A, then the number of recall messages sent to users within the preset configuration days closest to the current moment = the real-time sending index of the current date + the historical sending summary = the real-time sending index of the current date + the sum of the summary sending index of each day within (A-1) days before the current date. For example, if the current date is the 23rd, and if the preset configuration days are 4 days, then the number of recall messages sent to users within the preset configuration days closest to the current moment is the real-time sending index of the 23rd + the summary sending index of the 22nd + the summary sending index of the 21st + the summary sending index of the 20th, where the summary sending index of the 22nd + the summary sending index of the 21st + the summary sending index of the 20th is the historical sending summary.

[0122] This embodiment can quickly and accurately determine the number of recall messages sent to users within the preset configured number of days closest to the current moment by determining the real-time sending index of recall messages that have been sent on the current date, as well as the aggregated sending index of recall messages sent throughout the days on each date before the current date. This can then accurately determine whether the user's recall request should be intercepted to avoid frequent disturbing of the user.

[0123] Furthermore, based on the above embodiments, a fourth embodiment of the user breakpoint management method of the present invention is proposed.

[0124] In the above step S5011, the step of determining the summary sending index of each date before the current date includes:

[0125] Step S50111, obtaining a data array with a length of a preset value, obtaining a real-time sending index of the day before the current date and a batch sending index of the day before the current date;

[0126] Step S50112, determining a summary sending index for the day before the current date according to the real-time sending index for the day before the current date and the batch sending index;

[0127] In this embodiment, a data data with a length of a preset value is obtained, referring to Figure 3 , Figure 3 This is a schematic diagram of a scenario of a data array. The preset value is the maximum number of days in the frequency control interval, and the frequency control interval is the maximum interval for recording data. For example, if the frequency control interval is 30 days, the preset value is 30, which means that the maximum number of recall information sent to users per day within 30 days is recorded.

[0128] Assuming that the maximum frequency control interval is N days, an array of length = N will be generated. Each node in the array represents the summary data on a natural day, including two attributes: date (date) and total number of sends (count). The date is the date of the natural day, including year, month, and day; the total number of sends is the summary sending index of recall information sent to the user within that date.

[0129] Get the real-time sending index and batch sending index of the day before the current date. Since the data information of the previous day can be fully determined within the current date, that is, the real-time sending index and batch sending index of the previous day can be fully determined, the summary sending index of the day before the current date can be obtained based on the real-time sending index and batch sending index of the previous day.

[0130] Step S50113, determining the corresponding data node of the summary sending index of the day before the current date in the data array according to the date of the day before the current date and the preset value;

[0131] Step S50114, putting the summary sending index of the day before the current date into the corresponding data node in the data array;

[0132] Since the length of the array is limited and the number of natural days is infinite, the mapping between natural days and array indices uses a hash algorithm to make adjacent natural day data fall on adjacent nodes. Figure 4, Figure 4 This is a schematic diagram of a scenario where the data node of the array data is used to store the summary and transmission indicators of each date. The method is as follows:

[0133] M=[unix_timestamp(current_date) / 86400]%N;

[0134] Among them, `N` represents the array length, `Array` represents the array, `M` represents the array subscript, the `unix_timestamp()` function means converting the date in the "yyyy-MM-dd" format into a timestamp, `[]` represents rounding, and `%` represents the remainder. The mapping function between natural days and array subscripts is called a hash.

[0135] For example, if the current date is "2021-04-23", and the maximum number of days N for configuring the threshold value of the frequency control interval is 31, then according to

[0136] M=[unix_timestamp(current_date) / 86400]%N;

[0137] M=unix_timestamp(2020-04-23) / 86400%31;

[0138] M = 15;

[0139] Then the data of sending the recall message to the user on "2021-04-23" will be recorded in the node with the array index "15" (Array

[15] ).

[0140] If the current date is T, the summary sending index of the T-1 day can be determined in the early morning of the T day, and the array subscript M can be obtained by addressing according to the hash algorithm, and the data of the T-1 day can be written into the corresponding node. When writing data into the array node, the date will be written into the date (date) attribute of the corresponding subscript node of the array, and the summary sending index within the date will be written into the total number of transmissions (count) attribute of the corresponding subscript node of the array. Similarly, the summary data of the T day is generated on the T+1 day, and the date of the next day will be on the right side of the node.

[0141] Therefore, through the above method, each node in the data array stores the summary sending indicators of each date before the current date.

[0142] In the above step S5011, the step of determining the summary sending index of each date before the current date, and determining the historical sending summary according to the summary sending index of each date before the current date and the preset configuration days includes:

[0143] Step S50115, determining the summary sending index of each date before the current date according to the data array, and determining the historical sending summary according to the summary sending index of each date before the current date and the preset configuration days.

[0144] Since the data array stores the summary sending index of each date before the current date, the summary sending index of each date before the current date can be determined based on the data array.

[0145] At this time, the historical sending summary can be determined based on the summary sending indicators of each date before the current date and the preset configuration days, and then the number of recall messages sent to the user within the preset configuration days from the current moment can be determined based on the historical sending summary and the real-time sending indicator of the current date:

[0146]

[0147] in, Represents the historical sending summary, which is the sum of the aggregated sending indicators of all dates within A-1 date from the current date T.

[0148] R T Represents the real-time sending indicator of the current date, Array represents an array, Array[i] represents the summary sending indicator of a certain date, M represents the subscript of the current date in the array, A represents the preset configuration days, and Total represents the number of recall messages sent to users within the preset configuration days from the current moment.

[0149] For example, the current date is 2021-04-23, the preset configuration days are 5 days, and the array subscript of the current date is:

[0150] M=[unix_timestamp(2021-04-23) / 86400]%N=15;

[0151] The historical sending summary is:

[0152]

[0153] Then the number of recall messages sent to the user within the preset number of days closest to the current time is:

[0154]

[0155] Total=R T +Array

[11] +Array

[12] +Array

[13] +Array

[14] ;

[0156] Among them, R Trepresents the real-time sending indicators of 2021-04-23, Array

[14] represents the aggregated sending indicators of 2021-04-22, Array

[13] represents the aggregated sending indicators of 2021-04-21, Array

[12] represents the aggregated sending indicators of 2021-04-20, and Array

[11] represents the aggregated sending indicators of 2021-04-19.

[0157] This embodiment uses an array to continuously store the summary sending indicators within consecutive dates, so that the summary sending indicators of each date before the current date can be quickly and accurately obtained, thereby quickly and accurately determining the historical sending summary.

[0158] Furthermore, based on the above embodiments, a fifth embodiment of the user breakpoint management method of the present invention is proposed.

[0159] In the above step S5011, after the step of determining the real-time sending index of the current date, the method further includes:

[0160] Step S601, determining whether the summary sending index of the day before the current date has been obtained within the current date;

[0161] Step S602: If yes, then execute the steps: determining the summary sending index of each date before the current date, and determining the historical sending summary according to the summary sending index of each date before the current date;

[0162] In this embodiment, it is determined whether the aggregate sending index of the day before the current date has been obtained at the current moment within the current date. Since the aggregate sending index of the previous day is calculated based on the real-time sending index of the previous day and the batch sending index of the previous day within the current date, when determining the number of recall messages sent to the user within the preset configured number of days closest to the current moment, the aggregate sending index of the day before the current date may not be obtained within the current date.

[0163] If the summary sending indicators of the day before the current date have been obtained within the current date, the historical sending summary is determined based on the summary sending indicators of each date before the current date. Based on the historical sending summary and the real-time sending indicators of the current date, the number of recall messages sent to users within the preset configuration days closest to the current moment is determined.

[0164] Step S603, if not, determining the historical sending summary according to the summary sending index of each date before the day before the current date;

[0165] Step S604, determining the number of recall messages sent to the user within the preset configured number of days from the current moment according to the historical sending summary, the real-time sending index within the current date and the real-time sending index of the day before the current date.

[0166] If the summary sending index of the day before the current date is not obtained at the current time within the current date, it means that only the real-time sending index of the day before the current date can be obtained at this time. Therefore, the historical sending summary is determined according to the summary sending index of each date before the day before the current date, and then the number of recall messages sent to the user within the preset configuration days closest to the current time is determined according to the historical sending summary, the real-time sending index of the day before the current date and the real-time sending index of the current date:

[0167]

[0168] in, It represents the historical sending summary, which is the sum of the summary sending indicators of all dates from T-1 day before the current date to A-2 days before.

[0169] R T Indicates the real-time sending indicator of the current date, R T-1 Represents the real-time sending index of the day before the current date, Array represents an array, Array[i] represents the summary sending index of a certain date, M represents the subscript of the current date in the array, A represents the preset configuration days, and Total represents the number of recall messages sent to users within the preset configuration days from the current moment.

[0170] For example, assuming that the current date is 2021-04-23, the preset configuration day number A is 5, and the number of recall messages currently sent to users is:

[0171] Array index of the current date:

[0172] M=[unix_timestamp(2021-04-23) / 86400]%N=15;

[0173] Historical sending summary:

[0174]

[0175] The number of recall messages sent to users within the preset configured number of days closest to the current time:

[0176] Total=R T +R T-1 +Array

[11] +Array

[12] +Array

[13] .

[0177] Among them, R T Indicates the real-time sending indicator on 2021-04-23, R T-1 It represents the real-time sending indicators of 2021-04-22, Array

[13] represents the aggregated sending indicators of 2021-04-21, Array

[12] represents the aggregated sending indicators of 2021-04-20, and Array

[11] represents the aggregated sending indicators of 2021-04-19.

[0178] If the Total is less than the preset threshold, the user's recall request will not be intercepted and a recall message will be sent to the user.

[0179] This embodiment improves the accuracy of counting the number of recall messages by determining the number of recall messages sent to users within a preset number of days based on the real-time sending indicator of the current date, the real-time sending indicator of the previous day, and the historical sending summary when the aggregate sending indicator of the previous day cannot be counted within the current date. The frequency of sending recall messages to users can be controlled more accurately.

[0180] Furthermore, based on the above embodiments, a sixth embodiment of the user breakpoint management method of the present invention is proposed.

[0181] In the above step S5011, the step of determining the real-time sending index of the current date includes:

[0182] Step S701, obtaining a node of the current day and a node of the previous day, and when the first recall message is successfully sent to the user within the current date, using the data in the node of the current day to overwrite the data in the node of the previous day;

[0183] Step S702, clearing the data in the current day node, and recording the real-time sending index of the current date in the current day node;

[0184] Step S703: when the recall information is successfully sent to the user again within the current date, the real-time sending indicator of the current date is updated.

[0185] In this embodiment, the node T of the current day and the node T-1 of the previous day are obtained. Figure 5 , Figure 5 The diagram is a scene diagram of the current day node and the previous day node. The current day node and the previous day node are data nodes used to store data. Among them, the current day node T is used to store the real-time sending index R within the current date. T , the previous day node T-1 is used to store the real-time sending index R of the day before the current date T-1 The node of the current day and the node of the previous day may be data nodes on an array for storing the summary sending indicators of each date.

[0186] When the recall information is successfully sent to the user for the first time on the current date, the data in the current day node is the real-time sending indicator of the day before the current date. Therefore, it is necessary to flip the current day node and the previous day node. Figure 6 , Figure 6 This is a schematic diagram of the scenario of flipping the node of the current day and the node of the previous day when the recall instruction is successfully sent to the user for the first time within the current date. First, the node of the previous day is reassigned with the data of the node of the current day, and then the node of the current day is reassigned with the data of the current date.

[0187] The specific steps are as follows: first, all the data in the node of the current day, such as the date attribute data and the real-time sending index (count) attribute data, are used to cover the data in the node of the previous day. At this time, the data in the node T of the current day has been transferred to the node T-1 of the previous day.

[0188] Then, initialize the node T of the day, clear the data in the node T of the day, write the date of the current date into the date (date) attribute of the node T of the day, and write the real-time sending index of the current date into the real-time sending index (count) attribute of the node T of the day. Since it is the first time that the recall information is successfully sent to the user within the current date, the real-time sending index of the current date in the node of the day is 1 at this time.

[0189] When a recall message is successfully sent to the user again within the current date, the real-time sending index of the current date within the node T of the day is updated, that is, when a recall message is successfully sent to the user again, the value of the real-time sending index within the node of the day is increased by 1; if n recall messages are successfully sent to the user again, the value of the real-time sending index within the node of the day is increased by n.

[0190] Through the above method, the real-time sending indicator of the current date is obtained.

[0191] This embodiment uses the current day node to record the real-time sending index of the current date. When the recall information is successfully sent to the user, the real-time sending index in the current day node is updated, so that the real-time sending index of the current date can be accurately determined.

[0192] The present invention also provides a user breakpoint management device, referring to Figure 7 , the user breakpoint management device comprises:

[0193] A first monitoring module 10 is used to obtain a user behavior data set in a preprocessing queue at a first preset frequency through a first spark streaming, and determine the latest occurrence time of each user's most recent behavior and current aggregated behavior data of each user according to the user behavior data set;

[0194] The data delay module 20 is used to put the current summary behavior data of each user into a breakpoint detection queue, and transfer the current summary behavior data of each user put into the breakpoint detection queue to a delay queue, and the delay queue starts from when the current summary behavior data of each user is received, and transfers the current summary behavior data of each user to the breakpoint detection queue after the breakpoint duration.

[0195] The second monitoring module 30 is used to obtain the past summary behavior data before the breakpoint duration at the current moment in the breakpoint detection queue through the second spark streaming at a second preset frequency;

[0196] The breakpoint judgment module 40 is used to determine that a breakpoint has occurred for the user corresponding to the past aggregated behavior data if the latest occurrence time in the past aggregated behavior data is the same as the latest occurrence time of the most recent behavior of the user corresponding to the past aggregated behavior data, and use the latest occurrence time in the past aggregated behavior data as the time when the breakpoint has occurred for the user.

[0197] Optionally, the first monitoring module is further used to:

[0198] Sort the user behavior data of each user in the user behavior data set according to the occurrence time, and determine the latest occurrence time of the most recent behavior of each user;

[0199] The total behavior information of each user is generated according to the behavior information in the user behavior data of each user, and the current summary behavior data of each user is obtained according to the total behavior information of each user and the latest occurrence time of the latest behavior of each user.

[0200] Optionally, the breakpoint determination module is further used to:

[0201] When a user recall request is detected, determining the number of recall messages sent to the user within the preset configured number of days closest to the current moment;

[0202] If the number is less than a preset threshold, a recall message is sent to the user.

[0203] Optionally, the breakpoint determination module is further used to:

[0204] Determine the real-time sending index of the current date and the summary sending index of each date before the current date, and determine the historical sending summary according to the summary sending index of each date before the current date and the preset configuration days;

[0205] The number of recall messages sent to the user within the preset configured number of days closest to the current moment is determined according to the historical sending summary and the real-time sending index of the current date.

[0206] Optionally, the breakpoint determination module is further used to:

[0207] Obtain a data array with a preset length, obtain the real-time sending index of the day before the current date and the batch sending index of the day before the current date;

[0208] Determine the summary sending index of the day before the current date according to the real-time sending index of the day before the current date and the batch sending index;

[0209] Determine, according to the date of the day before the current date and the preset value, a corresponding data node in the data array of the summary sending indicator of the day before the current date;

[0210] Put the summary sending index of the day before the current date into the corresponding data node in the data array;

[0211] The summary sending index of each date before the current date is determined according to the data array, and the historical sending summary is determined according to the summary sending index of each date before the current date and the preset configuration days.

[0212] Optionally, the breakpoint determination module is further used to:

[0213] Determine whether the summary sending index of the day before the current date has been obtained within the current date;

[0214] If yes, the steps are executed: determining the summary sending index of each date before the current date, and determining the historical sending summary according to the summary sending index of each date before the current date;

[0215] If not, determining the historical sending summary based on the summary sending index of each date before the day before the current date, and determining the historical sending summary based on the summary sending index of each date before the day before the current date;

[0216] The number of recall messages sent to the user within the preset configured number of days from the current moment is determined according to the historical sending summary, the real-time sending index within the current date and the real-time sending index of the day before the current date.

[0217] Optionally, the breakpoint determination module is further used to:

[0218] Obtaining the node of the current day and the node of the previous day, and when the first recall message is successfully sent to the user within the current date, using the data in the node of the current day to overwrite the data in the node of the previous day;

[0219] Clear the data in the current day node, and record the real-time sending index of the current date in the current day node;

[0220] When the recall information is successfully sent to the user again within the current date, the real-time sending indicator of the current date is updated.

[0221] The methods executed by the above-mentioned program units can refer to the various embodiments of the user breakpoint management method of the present invention, which will not be described in detail here.

[0222] The present invention also provides a user breakpoint management device, which includes: a memory, a processor, and a user breakpoint management program stored in the memory and executable on the processor. The method implemented when the user breakpoint management program is executed by the processor can refer to the various embodiments of the user breakpoint management method of the present invention, which will not be repeated here.

[0223] The invention also provides a computer storage medium.

[0224] The computer storage medium of the present invention stores a user breakpoint management program, and when the user breakpoint management program is executed by a processor, the steps of the user breakpoint management method described above are implemented.

[0225] Among them, the method implemented when the user breakpoint management program running on the processor is executed can refer to the various embodiments of the user breakpoint management method of the present invention, and will not be repeated here.

[0226] The present invention also provides a computer program product.

[0227] The computer program product of the present invention comprises a user breakpoint management program, and when the user breakpoint management program is executed by a processor, the steps of the user breakpoint management method described above are implemented.

[0228] Among them, the method implemented when the user breakpoint management program running on the processor is executed can refer to the various embodiments of the user breakpoint management method of the present invention, and will not be repeated here.

[0229] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0230] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0231] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0232] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A user breakpoint management method, characterized in that: The user breakpoint management method comprises: Acquire a user behavior data set in a preprocessing queue at a first preset frequency through a first spark streaming, and determine the latest occurrence time of each user's most recent behavior and current aggregated behavior data of each user according to the user behavior data set; Put the current aggregated behavior data of each user into a breakpoint detection queue, and transfer the current aggregated behavior data of each user put into the breakpoint detection queue to a delay queue, which starts when the current aggregated behavior data of each user is received and transfers the current aggregated behavior data of each user to the breakpoint detection queue after the breakpoint duration. Obtaining the breakpoint detection queue at a second preset frequency through a second spark streaming, and putting past summary behavior data before the breakpoint duration at the current moment; If the latest occurrence time in the past aggregated behavior data is the same as the latest occurrence time of the most recent behavior of the user corresponding to the past aggregated behavior data, then determining that a breakpoint occurred for the user corresponding to the past aggregated behavior data, and using the latest occurrence time in the past aggregated behavior data as the time of the breakpoint of the user; The step of determining the latest occurrence time of each user's recent behavior and each user's current aggregated behavior data according to the user behavior data set includes: Sort the user behavior data of each user in the user behavior data set according to the occurrence time, and determine the latest occurrence time of the most recent behavior of each user; The total behavior information of each user is generated according to the behavior information in the user behavior data of each user, and the current summary behavior data of each user is obtained according to the total behavior information of each user and the latest occurrence time of the latest behavior of each user.

2. The user breakpoint management method according to claim 1, characterized in that: After the step of determining a breakpoint of the user corresponding to the past aggregated behavior data, the method further includes: When a user recall request is detected, determining the number of recall messages sent to the user within the preset configured number of days closest to the current moment; If the number is less than a preset threshold, a recall message is sent to the user.

3. The user breakpoint management method according to claim 2, characterized in that: The step of determining the number of recall messages sent to the user within the preset configured number of days closest to the current moment comprises: Determine the real-time sending index of the current date and the summary sending index of each date before the current date, and determine the historical sending summary according to the summary sending index of each date before the current date and the preset configuration days; The number of recall messages sent to the user within the preset configured number of days closest to the current moment is determined according to the historical sending summary and the real-time sending index of the current date.

4. The user breakpoint management method according to claim 3, characterized in that: Before the step of determining the summary sending index of each date before the current date, the method further includes: Obtain a data array with a preset length, obtain the real-time sending index of the day before the current date and the batch sending index of the day before the current date; Determine the summary sending index of the day before the current date according to the real-time sending index of the day before the current date and the batch sending index; Determine, according to the date of the day before the current date and the preset value, a corresponding data node in the data array of the summary sending indicator of the day before the current date; Put the summary sending index of the day before the current date into the corresponding data node in the data array; The step of determining the summary sending index of each date before the current date, and determining the historical sending summary according to the summary sending index of each date before the current date and the preset configuration days includes: The summary sending index of each date before the current date is determined according to the data array, and the historical sending summary is determined according to the summary sending index of each date before the current date and the preset configuration days.

5. The user breakpoint management method according to claim 4, characterized in that: After the step of determining the real-time transmission index of the current date, the method further includes: Determine whether the summary sending index of the day before the current date has been obtained within the current date; If yes, the steps are executed: determining the summary sending index of each date before the current date, and determining the historical sending summary according to the summary sending index of each date before the current date; If not, determining the historical sending summary based on the summary sending index of each date before the day before the current date, and determining the historical sending summary based on the summary sending index of each date before the day before the current date; The number of recall messages sent to the user within the preset configured number of days from the current moment is determined according to the historical sending summary, the real-time sending index within the current date and the real-time sending index of the day before the current date.

6. The user breakpoint management method according to claim 3, characterized in that: The step of determining the real-time sending index of the current date includes: Obtaining the node of the current day and the node of the previous day, and when the first recall message is successfully sent to the user within the current date, using the data in the node of the current day to overwrite the data in the node of the previous day; Clear the data in the current day node, and record the real-time sending index of the current date in the current day node; When the recall information is successfully sent to the user again within the current date, the real-time sending indicator of the current date is updated.

7. A user breakpoint management device, characterized in that: The user breakpoint management device comprises: A first monitoring module is used to obtain a user behavior data set in a preprocessing queue at a first preset frequency through a first spark streaming, and determine the latest occurrence time of each user's most recent behavior and current aggregated behavior data of each user according to the user behavior data set; A data delay module, used to put the current aggregated behavior data of each user into a breakpoint detection queue, and transfer the current aggregated behavior data of each user put into the breakpoint detection queue to a delay queue, wherein the delay queue starts from when the current aggregated behavior data of each user is received, and transfers the current aggregated behavior data of each user to the breakpoint detection queue after the breakpoint duration; A second monitoring module is used to obtain the past summary behavior data before the breakpoint duration at the current moment from the breakpoint detection queue through a second spark streaming at a second preset frequency; a breakpoint determination module, configured to determine that a breakpoint has occurred for the user corresponding to the past aggregated behavior data if the latest occurrence time in the past aggregated behavior data is the same as the latest occurrence time of the most recent behavior of the user corresponding to the past aggregated behavior data, and use the latest occurrence time in the past aggregated behavior data as the time when the breakpoint has occurred for the user; The first monitoring module is specifically used to sort the user behavior data of each user in the user behavior data set according to the occurrence time, and determine the latest occurrence time of the most recent behavior of each user; generate the total behavior information of each user according to the behavior information in the user behavior data of each user, and obtain the current summary behavior data of each user according to the total behavior information of each user and the latest occurrence time of the most recent behavior of each user.

8. A user breakpoint management device, characterized in that: The user breakpoint management device comprises: a memory, a processor, and a user breakpoint management program stored in the memory and executable on the processor. When the user breakpoint management program is executed by the processor, the steps of the user breakpoint management method according to any one of claims 1 to 6 are implemented.

9. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the user breakpoint management method according to any one of claims 1 to 6 are implemented.