WO2019196495A1 - 基于时间窗口的业务流量分发 - Google Patents

基于时间窗口的业务流量分发 Download PDF

Info

Publication number
WO2019196495A1
WO2019196495A1 PCT/CN2018/121940 CN2018121940W WO2019196495A1 WO 2019196495 A1 WO2019196495 A1 WO 2019196495A1 CN 2018121940 W CN2018121940 W CN 2018121940W WO 2019196495 A1 WO2019196495 A1 WO 2019196495A1
Authority
WO
WIPO (PCT)
Prior art keywords
time window
user
service traffic
client
service
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2018/121940
Other languages
English (en)
French (fr)
Inventor
杨逸帆
白帆
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Sankuai Online Technology Co Ltd
Original Assignee
Beijing Sankuai Online Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Sankuai Online Technology Co Ltd filed Critical Beijing Sankuai Online Technology Co Ltd
Priority to US17/047,371 priority Critical patent/US20210119927A1/en
Publication of WO2019196495A1 publication Critical patent/WO2019196495A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/28—Flow control; Congestion control in relation to timing considerations
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/70—Admission control; Resource allocation
    • H04L47/80—Actions related to the user profile or the type of traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/70—Admission control; Resource allocation
    • H04L47/80—Actions related to the user profile or the type of traffic
    • H04L47/808—User-type aware

Definitions

  • Embodiments of the present application relate to the field of computer technology, and in particular, to time window based traffic distribution.
  • Real-time user behavior operations respond to real-time user behavior flows collected by clients. For example, when the user completes a certain action in the application, the certain marketing action is executed in real time.
  • users can browse and purchase multiple services in the app to express mixed intentions. For example, on the client of the Meituan app (app), users can purchase food, or buy a hotel or movie.
  • each business module wants to make a marketing touch to the user, and how to assign the user's behavior to the most suitable business module if the frequency of use of the user access channel is limited. To improve the operational efficiency of business traffic is a problem to be solved.
  • each service module operates independently and uses preemptive service traffic distribution, that is, which service module is triggered by the user first, and the service module has the priority of using the user access channel, and does not consider the service flow utilization efficiency of the overall application, thereby Lead to inefficient distribution of business traffic.
  • the embodiment of the present application provides a time window-based service traffic distribution method, which performs traffic distribution according to user behavior data in the same length time window, and can comprehensively consider the overall traffic utilization efficiency of the application, and helps improve service traffic distribution. effectiveness.
  • the embodiment of the present application provides a time window-based service traffic distribution method, including: generating a time window based on a preset rule; determining, according to an operation behavior of the client in the generated time window, determining that the distribution needs to be distributed. Business traffic and distribution of the determined traffic.
  • the embodiment of the present application provides a time window-based service flow distribution apparatus, including: a time window generating module, configured to generate a time window based on a preset rule; and a service flow distribution module, configured to generate Describe the service traffic of the user to the client in the time window, determine the traffic that needs to be distributed, and distribute the determined traffic.
  • an embodiment of the present application further discloses an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program
  • the time window based service traffic distribution method described in the embodiment of the present application is implemented.
  • the embodiment of the present application provides a computer readable storage medium, where a computer program is stored, and the program is executed by a processor to implement a time window based service traffic distribution method disclosed in the embodiment of the present application.
  • the service traffic distribution method disclosed in the embodiment of the present application generates a time window by using a preset rule; determining, according to the operation behavior of the client in the generated time window, the service traffic that needs to be distributed, and performing the determined service traffic. distribution.
  • the service traffic distribution method disclosed in the embodiment of the present application helps to improve the service traffic distribution efficiency by comprehensively considering the traffic utilization efficiency of the application according to the behavior of the user in the same length time window.
  • FIG. 1 is a flowchart of a service traffic distribution method according to Embodiment 1 of the present application.
  • FIG. 3 is a schematic diagram of another time window in a service traffic distribution method according to Embodiment 1 of the present application.
  • FIG. 4 is a schematic structural diagram of a service flow distribution apparatus according to Embodiment 2 of the present application.
  • FIG. 5 is a second schematic structural diagram of a service flow distribution apparatus according to Embodiment 2 of the present application.
  • a service traffic distribution method disclosed in this embodiment is shown in FIG. 1 , and the method includes: step 110 and step 120.
  • Step 110 Generate a time window based on a preset rule.
  • the time window in the embodiment of the present application refers to a certain certain time period.
  • the time window is used as a time scale for generating an arbitration point, and by obtaining the context in which the user is using the client in the time window, the inconsistency of the intent recognition on the time scale can be eliminated.
  • the preset rule may refer to using a running period of the client as a time window (referred to as a first time window for convenience of distinction).
  • generating the time window based on the preset rule may specifically include: when the client exits, generating a first time window, wherein the first time window is a time period from the start to the exit of the client.
  • the client runtime will increase. If the client runs for 5 minutes and then exits at 18:05, the client exits.
  • the first time window is generated, and the first time window is a time period from the start of the client to the exit, that is, a time period from 18:00 to 18:05.
  • the first time window can be generated immediately upon detecting that the client exits. Alternatively, when the client exits, the first time window is not generated immediately, but the exit timing is started; when the timeout period reaches the preset duration and the client is not restarted, the first time window is generated. Specifically, when the user leaves the client of the app, the time window is not generated immediately, but an exit delay window is created to time the time the user leaves the app. When the user does not enter the client again within the preset time, the user is confirmed to exit the client, and a first time window is generated, where the first time window is a time period from the client to the exit. Based on the generated first time window, the user's intent can be better determined.
  • the preset rule may refer to periodically generating a time window (referred to as a second time window for convenience of distinction) when the client is running, where the start time of the second time window It is the start time of the cycle, and the end time is the end time of the cycle.
  • a time window referred to as a second time window for convenience of distinction
  • the current period may also be adjusted according to user behavior habits and/or the type of service currently operated by the user in the client. For example, for a user who has a long decision period according to historical statistical information of the user, a longer period, such as 1 minute, is set for the user. Or, according to the characteristics of the product category currently viewed by the user in the client, etc., if the browsing for the wedding category is longer than the browsing for the take-away category, the user will set the browsing when the user browses the wedding product. For a longer period, such as 5 minutes, when the user switches to browse the US food category, set a shorter period, such as 30 seconds. By dynamically adjusting the current period, the user's complete intent is obtained as much as possible within one decision period, and the number of decisions can be reduced, thereby improving the efficiency of business traffic distribution.
  • the second time window may be periodically generated in the second manner while the client is running, and a first time window may be generated in the first manner when the client exits.
  • the monitored items that are reported to the server form the real-time data stream of the user. For example, when the user opens the client of the application, that is, when the client starts, the "app start” monitoring item is triggered; when the user clicks on a certain hotel, the "click poi" monitoring item is triggered, and the monitoring item is simultaneously acquired. The clicked hotel information is reached; when the client exits, the "app exit” monitoring item is triggered.
  • the server can determine the operation behavior of the client on the client within the corresponding time window.
  • the period of generating the second time window is 30 seconds.
  • the second time is generated.
  • an arbitration point is generated.
  • the server determines the user's operation behavior on the client according to the real-time data flow reported by the client in the time period corresponding to the second time window. For example, an advertisement is displayed on the page of client 1. After the user A starts the client 1, the server receives the "app start" monitoring item through the buried point set by the client.
  • the user A continues to use the client 1, and performs a series of operations such as searching for a hotel, clicking a hotel list, browsing a hotel, etc. through the client 1, and the server receives the various operations triggered by the user through the buried point set by the client. Monitor items and record them as real-time behavioral data streams for users.
  • the length of the running time gradually increases.
  • the client runs for the current period length, such as 30 seconds
  • the first second time window is generated, and an arbitration point is generated, and the server collects the user A.
  • the real-time behavior data in the 30 seconds and based on the collected real-time behavior data, determine the operating behavior of the client.
  • User A performs a search for a hotel, clicks on a hotel list, and browses a hotel operation within 30 seconds after the client 1 is started.
  • the server will perform the above operation according to User A within 30 seconds after the client 1 is started, as shown in FIG.
  • the time window is displayed, including search, click, and browse, to determine the behavior of user A in the first second time window.
  • user A continues to use client 1, and performs operations such as clicking on a hotel and booking a hotel.
  • the length of the client's running time gradually increases, assuming that the length of time increases to a preset period of integers. Times, such as 90 seconds, generates a third arbitration point by generating a second time window.
  • the server will collect the real-time behavior data of user A in this 90 seconds, as shown in the time window in Figure 3, including: behavior data of search, click, browse, click, and scheduled operation, and determine that user A is in the third The behavior of the operation within the second time window.
  • the server can also collect real-time behavior data of user A within 60 seconds to 90 seconds after startup, as the operation behavior of user A.
  • Manner 1 The intent score of the user for a plurality of preset business activities is determined according to the operation behavior; and the service traffic corresponding to a preset business activity with the highest intention score is used as the service traffic to be distributed.
  • the intent described in the embodiments of the present application is used to indicate whether the user is interested in a certain product or item in the client.
  • the intent can be described by some pre-configured user behavior pattern or by an intent calculated by the algorithm. For example, a user behavior pattern that clicks on 3 hotels and does not purchase within 10 minutes corresponds to one intent, or the order probability that the intent recognition system recognizes is a preset value corresponding to one intent.
  • the intent recognition system performs comprehensive scoring based on the collected real-time behavior data of the user, such as clicking, browsing, collecting, and positioning data.
  • the intent recognition system can offlinely count the preference scores of items such as the user's category preference based on the user data. For example, according to the history of the user, the browsing behavior is purchased, and the user's preference for the specific business and the specific category is scored offline (for example, the preference value of the takeaway crayfish is m, and the preference value of the takeaway yellow pheasant is n), and according to the business.
  • the demand and/or user historical behavior data gives weights to the user's behavior, such as the weight of the click is x, the weight of the collection is y, the weight of the unpaid order is z, and so on.
  • the intention identification system integrates the time of the user to stay at the client, the geographical position in the real-time data stream, the user behavior and the like, and can be calculated.
  • the behavior of the user within the time period corresponding to the time window is an intentional score for the preset business activity or a certain category or an item.
  • the client's operator pre-configures two business activities: business activity 1: configured action sequence a->b->c, business activity 2: intent recognition system configured with take-out service needs to score higher than p.
  • business activity 1 configured action sequence a->b->c
  • business activity 2 intent recognition system configured with take-out service needs to score higher than p.
  • the server matches the preset business activity by the behavior action of the user 12345 during this time, and obtains the activity 1.
  • the server invokes the intent recognition system through the behavior actions of the user 12345 and the user 12345 during the period of time, and the intent score of the take-out activity is q, q>p, that is, the user 12345 also matches the activity 2.
  • the intention score r of the user 12345 for the preset business activity 1 is determined.
  • the user 12345 compares the intention score r of the preset business activity 1 with the intention score q of the preset business activity 2 to obtain a preset business activity with the highest intention score.
  • activity 2 is the target business activity for traffic distribution. That is, the service flow occupying channel corresponding to the activity 2 performs a preset marketing activity on the user 12345.
  • the server matches the preset business activity by the behavior action of the user 12345 during this time, and the action sequence matching can be performed in an exact matching manner, and then the user's intention recognition score for the activity 1 is obtained by the intention recognition system.
  • a, b, and c respectively represent a behavioral action of the user
  • x, y, z, m, n, p, and q are positive numbers between [0, 1] for measuring the intent size.
  • the user may express the intention of mixing in a time window, when the specific intent is identified, a plurality of preset business activities that the user matches may be identified, and by further comparing the intention scores, multiple presets matched by the user are selected. For a business activity with the highest intention score in the business activity, the business traffic corresponding to the business activity with the highest intention score can use the channel to make the user reach, thereby improving the utilization of the business traffic.
  • determining the one of the preset service activities with the highest intention score as the target business activity of the service flow distribution may specifically include: determining one of the pre-scores with the highest intention score and the intention score higher than the set value. Set the business activity, the business traffic corresponding to the business activity as the business traffic that needs to be distributed. In this way, it is possible to avoid the distribution of traffic corresponding to the target business activity whose intention score is too low.
  • the “app exit" monitoring item is triggered.
  • the triggered monitoring item is reported to the background server.
  • the first time window is generated, and an arbitration point is generated.
  • the server obtains the time period from the time when the client starts to the time when the client exits, and the client reports the time.
  • a real-time data stream determining a preset business activity that the user matches, and an intent score of the preset business activity. Then, according to the intent score, one of the business activities that matches the user is selected, and the user is reached through the user access channel.
  • Method 2 Determine the service traffic hit by the operation behavior according to the hit rule of each service activity and the obtained operation behavior, and select the service traffic corresponding to the service activity with the highest priority among the service activities that are hit as the service traffic to be distributed. .
  • the hit rule may be a configuration condition of a business activity. For example, a user clicking on a takeaway hot pot category for more than 3 times may be a business activity configuration condition and is a hit rule for the business activity.
  • the priority of each business activity can be set in advance.
  • the operation behavior of the user in a time window hits a plurality of business activities
  • the one with the highest priority among the plurality of business activities is selected, and the service traffic corresponding to the business activity is regarded as the service traffic to be distributed.
  • the user access channel is a channel for the application to push information to the user, including at least: short message, push, red envelope/voucher, station letter, and advertisement inside and outside the station.
  • the short message further includes: functional short message and marketing short message, and the marketing short message has frequency limitation;
  • the push further includes: functional information and marketing information, and the marketing information has frequency limitation.
  • User access channels with frequency restrictions are exclusive channels, and user access channels without frequency restrictions are non-exclusive channels.
  • Exclusive channels usually include marketing SMS and marketing push. In order to avoid excessive interruption to users, there are strict restrictions on the number of touches per day. For example, users can only receive one marketing message per day.
  • Non-exclusive channels are usually reusable user access channels, such as resource-bit advertisements, app-site letters, coupons, etc., which can be accessed multiple times a day.
  • the user selects a preset business activity with the highest intention score as the target business activity of the traffic distribution, allowing it to occupy a non-exclusive channel to improve the utilization of the traffic.
  • select a preset business activity with the highest intention score select a preset business activity with the highest intention score, and allow it to occupy the exclusive channel as the target business activity of traffic distribution to improve the utilization of traffic.
  • a time window is generated by using a preset rule; the service traffic that needs to be distributed is determined according to the operation behavior of the user on the client in the generated time window, and the determined service traffic is distributed. Due to the business traffic distribution method disclosed in the present application, the matching degree of user behavior with all business activities in the same time period is comprehensively considered, and the business activity with the highest matching degree is selected to conduct marketing access to the user, and the application is comprehensively considered from the application level. The overall traffic utilization efficiency effectively improves the efficiency of business traffic distribution.
  • the user's built-in user access channel can be fully utilized to perform marketing activities for the user, for example, sending a message to the user, publishing an advertisement in the station, etc.; when the user exits the application, the selection is preferred.
  • Business activities, occupying exclusive channels, and marketing activities to users, such as sending marketing messages through third-party platforms, and rationally utilizing user access channels, can help improve service traffic distribution efficiency and improve resource utilization.
  • a time window based service traffic distribution device disclosed in this embodiment as shown in FIG. 4, the device includes:
  • the time window generating module 410 is configured to generate a time window based on the preset rule.
  • the service traffic distribution module 420 is configured to determine, according to the operation behavior of the user in the generated time window, the service traffic that needs to be distributed, and distribute the determined service traffic.
  • the time window generating module 410 includes:
  • the first time window generating sub-module 4101 is configured to generate a first time window when the client exits, wherein the first time window is a time period from the start to the exit of the client.
  • the time window generating module 410 includes:
  • a second time window generating sub-module 4102 configured to generate a second time window according to a period when the client is running, where a start time of the second time window is a start time of the cycle, and an end time is the The end of the cycle.
  • the device further includes:
  • the period adjustment module 430 is configured to adjust the duration of the period according to the behavior habits of the user and/or the type of service that the user currently operates in the client.
  • a longer period such as 1 minute
  • the current period is adjusted. If the browsing of the wedding category is longer than the browsing of the take-away category, when the user browses the wedding products, a longer period is set, such as 5 minutes, when the user switches to browse the US food category, Then set a shorter period, such as 30 seconds.
  • the service traffic distribution module 420 is further configured to:
  • the service traffic corresponding to the preset service activity with the highest intention score is used as the service traffic that needs to be distributed, including:
  • the service traffic corresponding to one of the preset service activities with the highest intention score and the intention score higher than the set value is used as the traffic flow that needs to be distributed.
  • the determined service traffic is distributed, including:
  • the exclusive channel For determining the traffic that needs to be distributed according to the operation behavior of the user to the client in the generated first time window, using the exclusive channel to distribute the determined service traffic, wherein the exclusive channel is a frequency access restricted user access channel .
  • the determined service traffic is distributed, including:
  • the non-exclusive channel Determining, by the non-exclusive channel, the determined service traffic, according to the service traffic that needs to be distributed, according to the user's operation behavior on the client in the generated second time window, wherein the non-exclusive channel is a user that has no frequency restriction Access to the channel.
  • This embodiment is an embodiment of the device corresponding to the first embodiment.
  • the specific implementations of the modules and sub-modules of the device are the same as the corresponding method steps, and are not described in this embodiment.
  • the service traffic distribution apparatus disclosed in the embodiment of the present application generates a time window based on a preset rule; determines, according to the operation behavior of the client in the generated time window, the service traffic that needs to be distributed, and performs the determined service traffic. distribution. Because the service flow distribution device disclosed in the present application comprehensively considers the matching degree of user behavior with all business activities in the same time period, and selects the business activity with the highest matching degree to conduct marketing access to the user, comprehensively considers the application from the application level. The overall traffic utilization efficiency helps to improve the efficiency of business traffic distribution.
  • the real-time data of the user is matched with the preset activity occupying the non-exclusive channel, and when the user withdraws from using the client, according to the user Real-time data is matched with preset activities occupying exclusive channels, which greatly improves the utilization of non-exclusive channels and exclusive channels.
  • the time window is not used to manage the actual distribution of service traffic.
  • the matching of the business activity begins, and it is impossible to predict whether the user may hit other business activities in the future, and according to the probability of hitting and the business activity.
  • Real-time effects, etc. are estimated to abandon the current hits and wait for other hits, or directly respond to this hit. Therefore, the prior art business activity matching algorithm cannot accurately guess the user's future intentions and Variety.
  • the service traffic distribution method disclosed in the present application allows the user to express the intent completely by using the time window, so as to perform accurate intent identification and business activity matching, thereby further improving the accuracy of service traffic distribution, thereby effectively improving the efficiency of service traffic distribution. Improve resource utilization.
  • the service flow distribution device disclosed in the embodiment of the present application may be used to implement the time window-based service traffic distribution method provided in the first embodiment.
  • the service flow distribution device disclosed in the embodiment of the present application may be used to implement the time window-based service traffic distribution method provided in the first embodiment.
  • the present application also discloses an electronic device including a memory, a processor, and a computer program stored on the memory and operable on the processor, the processor executing the computer program, such as The time window based service traffic distribution method according to the first embodiment of the present application.
  • the electronic device can be a PC, a mobile terminal, a personal digital assistant, a tablet, or the like.
  • the present application also discloses a computer readable storage medium having stored thereon a computer program, which is executed by a processor to implement a time window based service traffic distribution method according to the first embodiment of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

本申请的实施例公开的一种基于时间窗口的业务流量分发方法、装置、电子设备及计算机可读存储介质，所述方法包括：基于预设规则生成时间窗口；根据在生成的时间窗口内用户对客户端的操作行为确定需要分发的业务流量，并对所确定的业务流量进行分发。

Description

基于时间窗口的业务流量分发
相关申请的交叉引用
本专利申请要求于2018年04月11日提交的、申请号为201810322980.5、发明名称为“基于时间窗口的业务流量分发方法及装置，电子设备”的中国专利申请的优先权，该申请的全文以引用的方式并入本文中。
技术领域
本申请的实施例涉及计算机技术领域，特别是涉及基于时间窗口的业务流量分发。
背景技术
实时用户行为运营，就是对客户端收集到的实时用户行为流进行响应。例如，当用户在应用内完成某某动作后，实时执行某某营销动作。随着互联网企业业务种类愈加丰富，用户可以在应用内浏览购买使用多种服务，表达混合的意图。例如在美团应用(app)的客户端上用户可以购买美食，也可以购买酒店或者电影等。当用户在应用内表达混合的意图时，各个业务模块都想对用户进行营销触达，在用户触达通道使用频次受限的情况下，如何将用户的行为分配给最合适的业务模块。以提高业务流量的运营效率是一个待解决的问题。目前，各个业务模块独立运营，采用抢占式业务流量分发，即用户先触发了哪个业务模块，那个业务模块就拥有使用用户触达通道的优先权，并没有考虑整体应用的业务流量利用效率，从而导致业务流量分发效率低下。
发明内容
本申请的实施例提供一种基于时间窗口的业务流量分发方法，根据相同长度时间窗口内用户的行为数据，进行业务流量分发，可以综合考虑应用整体的流量利用效率，有助于提高业务流量分发效率。
第一方面，本申请实施例提供了一种基于时间窗口的业务流量分发方法包括：基于预设规则生成时间窗口；根据在生成的所述时间窗口内用户对客户端的操作行为，确定需要分发的业务流量，并对所确定的所述业务流量进行分发。
第二方面，本申请实施例提供了一种基于时间窗口的业务流量分发装置，包括：时 间窗口生成模块，用于基于预设规则生成时间窗口；业务流量分发模块，用于根据在生成的所述时间窗口内用户对客户端的操作行为，确定需要分发的业务流量，并对所确定的所述业务流量进行分发。
第三方面，本申请实施例还公开了一种电子设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现本申请实施例所述的基于时间窗口的业务流量分发方法。
第四方面，本申请实施例提供了一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行以实现本申请实施例公开的基于时间窗口的业务流量分发方法。
本申请实施例公开的业务流量分发方法，通过基于预设规则生成时间窗口；根据在生成的时间窗口内用户对所述客户端的操作行为确定需要分发的业务流量，并对所确定的业务流量进行分发。本申请的实施例公开的业务流量分发方法通过根据相同长度时间窗口内用户的行为，进行业务流量分发，综合考虑应用整体的流量利用效率，有助于提高业务流量分发效率。
附图说明
为了更清楚地说明本申请实施例的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1是本申请实施例一的业务流量分发方法流程图；
图2是本申请实施例一的业务流量分发方法中一个时间窗口示意图；
图3是本申请实施例一的业务流量分发方法中另一个时间窗口示意图；
图4是本申请实施例二的业务流量分发装置的结构示意图之一；
图5是本申请实施例二的业务流量分发装置的结构示意图之二。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本 申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
实施例一
本实施例公开的一种业务流量分发方法，如图1所示，该方法包括：步骤110和步骤120。
步骤110，基于预设规则生成时间窗口。
本申请实施例中的时间窗口是指特定的某一时间段。在本申请实施例中，时间窗口被用于作为产生仲裁点的时间尺度，通过获取在所述时间窗口用户在使用客户端过程中的上下文，可以消除意图识别在时间尺度上不一致。
所述预设规则可以根据实际需要进行设置。下面对本申请实施例提供的两种设置方式进行说明。
在第一种方式中，预设规则可以是指，将客户端的一次运行期间作为一个时间窗口(为了便于区分，这里称为第一时间窗口)。此时，基于预设规则生成时间窗口，可以具体包括：在客户端退出时，生成第一时间窗口，其中，所述第一时间窗口为客户端从启动到退出的时间段。
例如，客户端在18时00分启动后，随着用户继续使用该客户端，客户端运行时间会增长，假设客户端运行了5分钟后，在18时05分退出，则在客户端退出时，生成第一时间窗口，则所述第一时间窗口为客户端启动到退出的时间段，即18时00分至18时05分的时间段。
可以在检测到客户端退出时，立即生成第一时间窗口。或者，也可以在客户端退出时，不立即生成第一时间窗口，而是启动退出计时；当退出计时的时长达到预设时长且所述客户端未重新启动时，生成第一时间窗口。具体来说，当用户离开app的客户端后，不会立即生成时间窗口，而是，创建一个退出延时窗口对用户离开app的时间进行计时。当在预设时间内用户没有再次进入所述客户端，则确认用户退出所述客户端，并生成第一时间窗口，所述第一时间窗口为客户端启动到退出的时间段。基于所生成的第一时间窗口，可以更好的确定用户的意图。
在第二种方式中，预设规则可以是指在客户端运行时，周期性的产生时间窗口(为了便于区分，称为第二时间窗口)，其中，所述第二时间窗口的起始时刻为所述周期的开始时刻，结束时刻为所述周期的结束时刻。
下面以在客户端运行时，30秒为周期产生第二时间窗口举例，说明第二时间窗口的生成过程。例如，客户端在18时00分00秒启动后，随着用户继续使用该客户端，客户端运行时间会增长，客户端运行了30秒钟后，则生成一个第二时间窗口，该第二时间窗口对应的时间段为第一个周期的开始时刻18时00分00秒至第一个周期的结束时刻18时00分30秒。随着用户继续使用该客户端，客户端运行时间会继续增长，在客户端继续运行30秒钟后，再一次生成一个第二时间窗口，该第二时间窗口对应的时间段为第二个周期的起始时刻18时00分30秒至第二个周期的结束时刻18时01分00秒。
可选的，还可以根据用户行为习惯和/或用户当前在客户端内操作的业务类型调整当前周期。例如，针对根据用户的历史统计信息，对于决策周期较长的用户，会为其设置较长的周期，如1分钟。或者，根据用户当前在客户端内查看的商品品类特性等调整当前周期，如对于婚庆类品类的浏览相比于外卖品类的浏览会使用更长，则当用户浏览婚庆类品商品时，会设置较长的周期，如5分钟，当用户切换至浏览美食品类时，则设置较短的周期，如30秒。通过动态调整当前周期，使得在一个决策周期内尽量获取用户的完整意图，并可以减少决策次数，从而提升业务流量分发的效率。
以上分别介绍了本公开实施例提供的两种生成时间窗口的方式。需要指出的是，在具体实施时，上述两种生成时间窗口的方式并不是相互排斥的。在一些实施例中，可以在客户端运行时，按照第二种方式周期性的产生第二时间窗口，并在客户端退出时，按照第一种方式产生一个第一时间窗口。
步骤120，根据在生成的时间窗口内用户对客户端的操作行为，确定需要分发的业务流量，并对所确定的业务流量进行分发。
首先介绍“获取生成的时间窗口内用户对所述客户端的操作行为”的实现方式。
可以通过在应用程序的客户端软件中设置监控项，获取用户操作客户端的实时数据。当监控项被触发后，被触发的监控项会会上报给后台服务器，这些被上报到服务器的一条条的监控项，就形成了所述用户的实时数据流。例如，当用户打开应用的客户端，即客户端启动时，就会触发“app启动”这个监控项；当用户点击的某个酒店，就会触发“点击poi”这个监控项，监控项同时获取到了点击的酒店信息；当客户端退出时，就会触发“app退出”这个监控项。服务器端通过客户端上报的实时数据，可以确定客户端在相应的时间窗口内，用户对所述客户端的操作行为。
以客户端运行过程中，生成第二时间窗口的周期为30秒举例，在用户使用所述客户 端的过程中，每当客户端运行时间长度等于30秒的整数倍时，即通过生成第二时间窗口时，产生一个仲裁点，此时，服务器端根据所述客户端在所述第二时间窗口对应的时间段内上报的实时数据流，确定用户对客户端操作行为。例如，在客户端1的页面显示广告。当用户A启动客户端1之后，服务器通过所述客户端设置的埋点，接收到“app启动”这个监控项。
然后，用户A继续使用客户端1，通过客户端1执行了搜索酒店、点击酒店列表、浏览酒店等一系列操作，服务器通过所述客户端设置的埋点，接收到用户的各种操作触发的监控项，并记录为用户的实时行为数据流。随着用户继续使用客户端，运行的时间长度逐渐增长，当客户端运行时长达到当前周期长度，如30秒时，生成第一个第二时间窗口，并产生一个仲裁点，服务器将搜集用户A在这30秒内的实时行为数据，并根据搜集到的所述实时行为数据，确定客户端的操作行为。假设用户A在客户端1启动后的30秒内执行了搜索酒店、点击酒店列表、浏览酒店操作，服务器将根据用户A在客户端1启动后的30秒内执行的上述操作，如图2中的时间窗口所示，包括搜索、点击、浏览，确定用户A在第一个第二时间窗口内的操作行为。假设用户A继续使用客户端1，又执行了点击一家酒店，以及预订酒店等的操作，随着用户继续使用客户端，客户端的运行时间的长度逐渐增长，假设时间长度增长到预设周期的整数倍，如90秒时，通过生成第二时间窗口，产生第三个仲裁点。服务器将搜集用户A在这90秒内的实时行为数据，如图3中的时间窗口所示，即包括：搜索、点击、浏览、点击、预定操作的行为数据，确定用户A在第三个第二时间窗口内的操作行为。当然，对于不同的业务场景，服务器也可以搜集用户A在启动后的60秒到90秒内的实时行为数据，作为用户A的操作行为。
以上介绍了“获取生成的时间窗口内用户对所述客户端的操作行为”的一些方式。下面介绍本申请的实施例提供的“根据获取到的操作行为确定需要分发的业务流量”的一些实施方式。
方式一：根据所述操作行为，确定所述用户对多个预设业务活动的意图得分；将所述意图得分最高的一个预设业务活动所对应的业务流量，作为需要分发的业务流量。
可以按照如下方式确定用户的意图得分：通过将所述实时数据流中的用户动作与预设业务活动配置的动作序列进行匹配，确定所述用户匹配的预设业务活动，然后，通过匹配的所述预设业务活动对应的意图识别系统确定用户对匹配的所述预设业务活动的意图得分。或者，通过预设意图识别系统，根据所述实时数据流中的用户动作，确定所 述用户匹配的预设业务活动，以及对匹配的所述预设业务活动的意图得分。
本申请实施例中所述的意图，用于表示用户对客户端中的某一产品或项目等是否感兴趣。意图可以通过预先配置的某种用户行为模式描述，或者由算法计算的意图打分描述。例如：预先配置点击3个酒店并在10分钟内没有购买的用户行为模式对应一个意图，或者，意图识别系统识别的下单概率为预设值对应一个意图。意图识别系统会根据搜集到的用户的实时行为数据，如点击、浏览、收藏、定位等数据进行综合打分。
意图识别系统可以根据用户数据离线统计用户的品类偏好等项目的偏好分。例如，根据用户的历史购买浏览行为，离线计算得到该用户对具体业务、具体品类的偏好打分(例如，外卖小龙虾的偏好值为m,外卖黄焖鸡的偏好值为n)，并根据业务需求和/或用户历史行为数据，对用户行为给出权值，例如点击的权值为x，收藏的权值为y，下单未支付的权值为z等等。在实际使用过程中，当对用户对某一预设业务活动进行意图打分时，意图识别系统综合用户在客户端各停留的时间、实时数据流中的地理位置、用户行为等数据，可以计算得到用户在所述时间窗口对应的时间段内的行为对所述预设业务活动或某一品类、某一项目的意图得分。
例如，客户端的运营人员预先配置了两个业务活动：业务活动1：配置了动作序列a->b->c，业务活动2：配置了外卖业务的意图识别系统打分需要高于p。当用户12345使用了客户端，并执行了动作序列a->b->c之后,在一个时间窗口内，服务器通过用户12345在这段时间内的行为动作匹配预设业务活动，得到活动1。同时，服务器通过用户12345，以及用户12345在这段时间内的行为动作调用意图识别系统，获得外卖活动的意图得分为q，q>p，即用户12345与活动2也匹配。然后，确定用户12345对预设业务活动1的意图得分r。将用户12345对预设业务活动1的意图得分r和预设业务活动2的意图得分q相比较，得到意图得分最高的一个预设业务活动。此例中，若r<q，则表示活动2的意图得分最高，所以将活动2作为流量分发的目标业务活动。即由活动2对应的业务流量占用通道对用户12345执行预设的营销活动。服务器通过用户12345在这段时间内的行为动作匹配预设业务活动，可以采用完全匹配的方式进行动作序列匹配，然后，再通过意图识别系统获得用户对活动1的意图识别得分。本申请实施例中的a、b、c分别代表用户的一个行为动作，x、y、z、m、n、p和q为[0,1]之间的正数，用于衡量意图大小。由于用户在一个时间窗口内可能表达了混合的意图，所以，在具体意图识别时，可能会识别到用户匹配的多个预设业务活动，通过进一步比较意图得分，选择用户匹配的多个预设业务活动中意图得分最高的一个业务活动，这个意图得分最高的 业务活动对应的业务流量可以使用通道，进行用户触达，从而提升业务流量的利用率。
此外，还可以判断所述意图得分是否高于设定值，当意图得分高于设定值时，才确定意图匹配成功。此时，确定所述意图得分最高的一个所述预设业务活动，作为业务流量分发的目标业务活动，可以具体包括：确定所述意图得分最高且意图得分高于设定值的一个所述预设业务活动，该业务活动所对应的业务流量作为需要分发的业务流量。通过这种方式，能够避免将意图得分过低的目标业务活动所对应的业务流量进行分发。
再以所述客户端退出时生成第一时间窗口为例，说明根据在生成的第一时间窗口内用户对所述客户端的操作行为，确定需要分发的业务流量的具体实施方案。首先，当客户端退出时，会触发“app退出”这个监控项。被触发的监控项上报给后台服务器，此时，生成第一时间窗口，并产生一个仲裁点，服务器获取自所述客户端启动时至所述客户端退出的时间段内，所述客户端上报的实时数据流，确定所述用户匹配的预设业务活动，以及所述预设业务活动的意图得分。然后，根据所述意图得分，选择与用户匹配的一个所述业务活动，通过用户触达通道，对用户进行营销触达。
方式二：根据各个业务活动的命中规则以及所获取到的操作行为确定操作行为所命中的业务流量，选择所命中的业务活动中优先级最高的业务活动所对应的业务流量作为需要分发的业务流量。
所述命中规则可以为一种业务活动的配置条件。例如，用户点击外卖火锅品类3次以上可以是一种业务活动配置条件，是该业务活动的命中规则。
可以预先设置各个业务活动的优先级。当用户在一个时间窗口内的操作行为命中多个业务活动时，选择多个业务活动中优先级最高的一个，将该业务活动对应的业务流量作为需要分发的业务流量。
业务流量分发的方式可以参考本领域技术人员熟知的技术。在本申请提供的一种实施方式中，当按照预设规则既产生第一时间窗口，又产生第二时间窗口时，对所确定的业务流量进行分发可以具体包括：对于根据在生成的第一时间窗口内用户对所述客户端的操作行为确定需要分发的业务流量，利用独占通道分发所确定的业务流量；对于根据在生成的第二时间窗口内用户对所述客户端的操作行为确定需要分发的业务流量，利用非独占通道分发所确定的业务流量。
用户触达通道是应用向用户推送信息的渠道，至少包括：短信、推送、红包/代金券等、站内信以及站内外广告等。其中，短信进一步包括：功能型短信和营销短信，营销 短信有频次限制；推送进一步包括：功能型信息和营销信息，营销信息有频次限制。有频次限制的用户触达通道属于独占通道，没有频次限制的用户触达通道属于非独占通道。独占类通道通常包括营销短信和营销推送，为了避免对用户的过度打扰，对每天触达次数有严格限制，例如：用户一天只能收到一条营销短信。非独占类通道通常为可以重复使用的用户触达通道，例如资源位广告，app站内信，优惠券等，一天可以多次触达用户。
在用户使用客户端的过程中，选择意图得分最高的一个预设业务活动，作为流量分发的目标业务活动，允许其占用非独占通道，以提高业务流量的利用率。在用户退出客户端之后，选择意图得分最高的一个预设业务活动，允许其占用独占通道，作为流量分发的目标业务活动，以提高业务流量的利用率。
对于人为配置意图，由于在时间尺度上的不一致性，导致了哪个业务活动配置意图的门槛低(例如，仅一次点击)，哪个业务活动先占用用户触达通道，那么对于独占类通道(短信，推送等)，从而使得意图耗时长的活动将无法利用该通道触达用户。同时由于意图识别在时间尺度上的不一致性，导致意图识别系统无法介入仲裁、择优选择业务活动，用户触达通道的利用不合理。
本申请所提供的实施例中，通过基于预设规则生成时间窗口；根据在生成的时间窗口内用户对所述客户端的操作行为确定需要分发的业务流量，并对所确定的业务流量进行分发。由于本申请公开的业务流量分发方法，综合考虑了同一时间段内，用户行为与所有业务活动的匹配度，并选择匹配度最高的业务活动对用户进行营销触达，从应用层面，综合考虑应用整体的流量利用效率，有效提高业务流量分发效率。
同时，在使用应用程序的过程中，可以充分利用应用程序内置的用户触达通道，对用户进行营销活动，例如，给用户发站内信，发布站内广告等；当用户退出应用程序后，择优选择业务活动，占用独占通道，对用户进行营销活动，例如通过第三方平台发送营销短信，合理利用用户触达通道，有助于提升业务流量分发效率，提高资源利用率。
实施例二
本实施例公开的一种基于时间窗口的业务流量分发装置，如图4所示，所述装置包括：
时间窗口生成模块410，用于基于预设规则生成时间窗口。
业务流量分发模块420，用于根据在生成的所述时间窗口内用户对客户端的操作行 为，确定需要分发的业务流量，并对所确定的业务流量进行分发。
可选的，如图5所示，所述时间窗口生成模块410包括：
第一时间窗口生成子模块4101，用于在客户端退出时，生成第一时间窗口，其中，所述第一时间窗口为客户端从启动到退出的时间段。
可选的，如图5所示，所述时间窗口生成模块410包括：
第二时间窗口生成子模块4102，用于在客户端运行时，按照周期产生第二时间窗口，其中，所述第二时间窗口的起始时刻为所述周期的开始时刻，结束时刻为所述周期的结束时刻。
可选的，如图5所示，所述装置还包括：
周期调整模块430，用于根据用户的行为习惯和/或用户当前在客户端内操作的业务类型调整所述周期的时长。
例如，针对根据历史统计信息，对于决策周期较长的用户，会为其设置较长的周期，如1分钟；或者，根据用户当前在所述客户端内查看的商品品类特性等调整当前周期，如对于婚庆类品类的浏览相比于外卖品类的浏览会使用更长，则当用户浏览婚庆类品商品时，会设置较长的周期，如5分钟，当用户切换至浏览美食品类时，则设置较短的周期，如30秒。通过动态调整当前周期，使得在一个决策周期内尽量获取用户的完整意图，并可以减少决策次数，提升业务流量分发的效率。
可选的，如图5所示，所述业务流量分发模块420进一步用于：
根据所述操作行为，确定所述用户对多个预设业务活动的意图得分；将所述意图得分最高的一个预设业务活动所对应的业务流量，作为需要分发的业务流量。
可选的，将所述意图得分最高的一个所述预设业务活动所对应的业务流量，作为需要分发的业务流量，包括：
将所述意图得分最高且所述意图得分高于设定值的一个所述预设业务活动所对应的业务流量，作为需要分发的业务流量。
可选的，对所确定的业务流量进行分发，包括：
对于根据在生成的第一时间窗口内用户对所述客户端的操作行为确定需要分发的业务流量，利用独占通道分发所确定的业务流量，其中，所述独占通道为有频次限制的用户触达通道。
可选的，对所确定的业务流量进行分发，包括：
对于根据在生成的第二时间窗口内用户对所述客户端的操作行为确定需要分发的业务流量，利用非独占通道分发所确定的业务流量，其中，所述非独占通道为没有频次限制的用户触达通道。
本实施例是与实施例一对应的装置实施例，装置的各模块和子模块的具体实施方式与相应的方法步骤相同，本实施例中不再赘述。
本申请实施例公开的业务流量分发装置，通过基于预设规则生成时间窗口；根据在生成的时间窗口内用户对所述客户端的操作行为确定需要分发的业务流量，并对所确定的业务流量进行分发。由于本申请公开的业务流量分发装置，综合考虑了同一时间段内，用户行为与所有业务活动的匹配度，并选择匹配度最高的业务活动对用户进行营销触达，从应用层面，综合考虑应用整体的流量利用效率，有助于提高业务流量分发效率。
同时，在基于时间窗口进行业务流量分发的过程中，当用户正在使用客户端时，根据用户的实时数据与占用非独占通道的预设活动进行匹配，当用户退出使用客户端时，根据用户的实时数据与占用独占通道的预设活动进行匹配，充分提高了非独占通道和独占通道的利用率。当在使用应用程序的过程中，可以充分利用应用程序内置的用户触达通道，对用户进行营销活动，例如，给用户发站内信，发布站内广告等；当用户退出应用程序后，择优选择业务活动，占用独占通道，对用户进行营销活动，例如通过第三方平台发送营销短信，合理利用用户触达通道，有效提升业务流量分发效率，提高资源利用率。
现有技术中，不采用时间窗口管理对业务流量进行分发的实际，当某个意图发生时就开始匹配业务活动，无法预测用户未来是否有可能命中其他业务活动，并根据命中概率和业务活动的实时效果等，估算出是放弃当前已有的命中而等待可能的其他命中，还是直接响应本次命中，因此，现有技术中的业务活动匹配算法并不能非常准确的猜测用户未来意图的走向和变化。而本申请公开的业务流量分发方法，通过采用时间窗口的方式可以让用户完整表达意图，以执行准确的意图识别和业务活动匹配，进一步提升业务流量分发的准确性，从而有效提升业务流量分发效率，提高资源利用率。
本申请实施例公开的业务流量分发装置可以用以实施实施例一所提供的基于时间窗口的业务流量分发方法，相关概念及具体实施方式可以参考实施例一中的描述，这里不再赘述。
相应的，本申请还公开了一种电子设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现如本申请实施例一所述的基于时间窗口的业务流量分发方法。所述电子设备可以为PC机、移动终端、个人数字助理、平板电脑等。
本申请还公开了一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行以实现如本申请实施例一所述的基于时间窗口的业务流量分发方法。
本说明书中的各个实施例均采用递进的方式描述，每个实施例重点说明的都是与其他实施例的不同之处，各个实施例之间相同相似的部分互相参见即可。对于装置实施例而言，由于其与方法实施例基本相似，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。
以上对本申请提供的一种基于时间窗口的业务流量分发方法及装置进行了详细介绍，本文中应用了具体个例对本申请的原理及实施方式进行了阐述，以上实施例的说明只是用于帮助理解本申请的方法及其核心思想；同时，对于本领域的一般技术人员，依据本申请的思想，在具体实施方式及应用范围上均会有改变之处，综上所述，本说明书内容不应理解为对本申请的限制。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到各实施方式可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件实现。基于这样的理解，上述技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品可以存储在计算机可读存储介质中，如ROM/RAM、磁碟、光盘等，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行各个实施例或者实施例的某些部分所述的方法。

Claims (18)

  1. 一种基于时间窗口的业务流量分发方法，包括：
    基于预设规则生成时间窗口；
    根据在生成的所述时间窗口内用户对客户端的操作行为，确定需要分发的业务流量，并
    对所确定的所述业务流量进行分发。
  2. 根据权利要求1所述的方法，其特征在于，基于所述预设规则生成所述时间窗口，包括：
    在所述客户端退出时，生成第一时间窗口，其中，所述第一时间窗口为从所述客户端启动到退出的时间段。
  3. 根据权利要求1或2所述的方法，其特征在于，基于所述预设规则生成所述时间窗口，包括：
    在所述客户端运行时，按照周期产生第二时间窗口，其中，所述第二时间窗口的起始时刻为所述周期的开始时刻，结束时刻为所述周期的结束时刻。
  4. 根据权利要求3所述的方法，其特征在于，所述方法还包括：
    根据所述用户的行为习惯和/或所述用户当前在所述客户端内操作的业务类型调整所述周期的时长。
  5. 根据权利要求1所述的方法，其特征在于，根据在生成的所述时间窗口内所述用户对所述客户端的所述操作行为，确定需要分发的所述业务流量，包括：
    根据所述操作行为，确定所述用户对多个预设业务活动的意图得分；
    将所述意图得分最高的一个预设业务活动所对应的业务流量，作为需要分发的所述业务流量。
  6. 根据权利要求5所述的方法，其特征在于，将所述意图得分最高的一个所述预设业务活动所对应的所述业务流量，作为需要分发的所述业务流量，包括：
    将所述意图得分最高且所述意图得分高于设定值的一个所述预设业务活动所对应的所述业务流量，作为需要分发的所述业务流量。
  7. 根据权利要求2所述的方法，其特征在于，对所确定的所述业务流量进行分发，包括：
    对于根据在生成的所述第一时间窗口内所述用户对所述客户端的所述操作行为确定需要分发的所述业务流量，利用独占通道分发所确定的所述业务流量，
    其中，所述独占通道为有频次限制的用户触达通道。
  8. 根据权利要求3所述的方法，其特征在于，对所确定的所述业务流量进行分发，包括：
    对于根据在生成的所述第二时间窗口内所述用户对所述客户端的所述操作行为确定需要分发的所述业务流量，利用非独占通道分发所确定的所述业务流量，
    其中，所述非独占通道为没有频次限制的用户触达通道。
  9. 一种基于时间窗口的业务流量分发装置，包括：
    时间窗口生成模块，用于基于预设规则生成时间窗口；
    业务流量分发模块，用于根据在生成的所述时间窗口内用户对客户端的操作行为，确定需要分发的业务流量，并对所确定的业务流量进行分发。
  10. 根据权利要求9所述的装置，其特征在于，所述时间窗口生成模块包括：
    第一时间窗口生成子模块，用于在所述客户端退出时，生成第一时间窗口，其中，所述第一时间窗口为客户端从启动到退出的时间段。
  11. 根据权利要求9或10所述的装置，其特征在于，所述时间窗口生成模块包括：
    第二时间窗口生成子模块，用于在所述客户端运行时，按照周期产生第二时间窗口，其中，所述第二时间窗口的起始时刻为所述周期的开始时刻，结束时刻为所述周期的结束时刻。
  12. 根据权利要求11所述的装置，其特征在于，所述装置还包括：
    周期调整模块，用于根据所述用户的行为习惯和/或所述用户当前在所述客户端内操作的业务类型调整所述周期的时长。
  13. 根据权利要求9所述的装置，其特征在于，所述业务流量分发模块进一步用于：
    根据所述操作行为，确定所述用户对多个预设业务活动的意图得分；
    将所述意图得分最高的一个预设业务活动所对应的业务流量，作为需要分发的所述业务流量。
  14. 根据权利要求13所述的装置，其特征在于，将所述意图得分最高的一个所述预设业务活动所对应的业务流量，作为需要分发的所述业务流量，包括：
    将所述意图得分最高且所述意图得分高于设定值的一个所述预设业务活动所对应的所述业务流量，作为需要分发的所述业务流量。
  15. 根据权利要求10所述的装置，其特征在于，对所确定的所述业务流量进行分发，包括：
    对于根据在生成的所述第一时间窗口内所述用户对所述客户端的所述操作行为确定需要分发的所述业务流量，利用独占通道分发所确定的所述业务流量，
    其中，所述独占通道为有频次限制的用户触达通道。
  16. 根据权利要求11所述的装置，其特征在于，对所确定的所述业务流量进行分发，包括：
    对于根据在生成的所述第二时间窗口内所述用户对所述客户端的所述操作行为确定需要分发的所述业务流量，利用非独占通道分发所确定的所述业务流量，
    其中，所述非独占通道为没有频次限制的用户触达通道。
  17. 一种电子设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，其特征在于，所述处理器执行所述计算机程序时实现权利要求1至8任意一项所述的基于时间窗口的业务流量分发方法。
  18. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，该程序被处理器执行以实现权利要求1至8任意一项所述的基于时间窗口的业务流量分发方法。
PCT/CN2018/121940 2018-04-11 2018-12-19 基于时间窗口的业务流量分发 Ceased WO2019196495A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/047,371 US20210119927A1 (en) 2018-04-11 2018-12-19 Service traffic distribution based on time window

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810322980.5 2018-04-11
CN201810322980.5A CN108696457B (zh) 2018-04-11 2018-04-11 一种业务流量分发方法及装置、电子设备及存储介质

Publications (1)

Publication Number Publication Date
WO2019196495A1 true WO2019196495A1 (zh) 2019-10-17

Family

ID=63844929

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/121940 Ceased WO2019196495A1 (zh) 2018-04-11 2018-12-19 基于时间窗口的业务流量分发

Country Status (3)

Country Link
US (1) US20210119927A1 (zh)
CN (1) CN108696457B (zh)
WO (1) WO2019196495A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108696457B (zh) * 2018-04-11 2021-04-23 北京三快在线科技有限公司 一种业务流量分发方法及装置、电子设备及存储介质
CN114124819B (zh) * 2021-10-22 2024-02-09 北京乐我无限科技有限责任公司 流量分发控制方法、装置、存储介质及计算机设备
WO2023123377A1 (zh) * 2021-12-31 2023-07-06 华为技术有限公司 一种调度网络流量的方法、装置以及系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090183218A1 (en) * 2008-01-10 2009-07-16 At&T Knowledge Ventures, Lp Predictive Allocation of Multimedia Server Resources
CN105959374A (zh) * 2016-05-12 2016-09-21 腾讯科技（深圳）有限公司 一种数据推荐方法及其设备
CN107341690A (zh) * 2017-06-30 2017-11-10 上海精数信息科技有限公司 信息推送方法及装置
CN107517190A (zh) * 2016-06-17 2017-12-26 中国移动通信集团上海有限公司 一种业务推荐方法及装置
CN107800746A (zh) * 2016-09-07 2018-03-13 百度在线网络技术（北京）有限公司 消息推送方法及装置
CN108696457A (zh) * 2018-04-11 2018-10-23 北京三快在线科技有限公司 基于时间窗口的业务流量分发方法及装置，电子设备

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090183218A1 (en) * 2008-01-10 2009-07-16 At&T Knowledge Ventures, Lp Predictive Allocation of Multimedia Server Resources
CN105959374A (zh) * 2016-05-12 2016-09-21 腾讯科技（深圳）有限公司 一种数据推荐方法及其设备
CN107517190A (zh) * 2016-06-17 2017-12-26 中国移动通信集团上海有限公司 一种业务推荐方法及装置
CN107800746A (zh) * 2016-09-07 2018-03-13 百度在线网络技术（北京）有限公司 消息推送方法及装置
CN107341690A (zh) * 2017-06-30 2017-11-10 上海精数信息科技有限公司 信息推送方法及装置
CN108696457A (zh) * 2018-04-11 2018-10-23 北京三快在线科技有限公司 基于时间窗口的业务流量分发方法及装置，电子设备

Also Published As

Publication number Publication date
CN108696457B (zh) 2021-04-23
US20210119927A1 (en) 2021-04-22
CN108696457A (zh) 2018-10-23

Similar Documents

Publication Publication Date Title
US12380175B2 (en) Dynamic radius threshold selection
US20240320713A1 (en) Systems and methods for providing content items via a push marketing auction
US10817931B2 (en) Systems and methods for selecting third party content based on feedback
US10313462B2 (en) Systems and methods of determining a geographic location based conversion
JP6839234B2 (ja) データ送信のためのフィードバックコントローラ
KR101840420B1 (ko) 챗봇 플랫폼 제공 방법 및 장치
US9195743B2 (en) System and method for creating and applying predictive user click models to predict a target page associated with a search query
US10163130B2 (en) Methods and apparatus for identifying a cookie-less user
US10185975B2 (en) Predicting unsubscription of potential customers
CN119577231A (zh) 分析个性化框架
US10715612B2 (en) Identifying users&#39; identity through tracking common activity
US20150006279A1 (en) Systems and methods of generating a valid location cluster based on a location of a commercial entity
WO2014071169A1 (en) Providing content to linked devices associated with a user
CN104317804B (zh) 发布投票信息的方法和装置
US11087424B1 (en) Image recognition-based content item selection
WO2019196495A1 (zh) 基于时间窗口的业务流量分发
US10922722B2 (en) System and method for contextual video advertisement serving in guaranteed display advertising
US20150302483A1 (en) Customized landing page system and method
CN119991252A (zh) 一种基于大数据的电子商务数据推送方法及系统
CN103793427A (zh) 确定目标呈现关键词的优化呈现关键词的方法与设备
CN103365845B (zh) 一种微博中的搜索方法及系统
US20160098773A1 (en) Infusing product popularity insights into search results for product research queries
CN111224731A (zh) 基于语音广播的内容推送方法、装置及设备
OA16858A (en) Method, system and computer storage medium for searching in microblog

Legal Events

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

Ref document number: 18914352

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18914352

Country of ref document: EP

Kind code of ref document: A1