WO2015096742A1

WO2015096742A1 - Information processing method, device and system

Info

Publication number: WO2015096742A1
Application number: PCT/CN2014/094830
Authority: WO
Inventors: 习明昊
Original assignee: 腾讯科技（深圳）有限公司
Priority date: 2013-12-24
Filing date: 2014-12-24
Publication date: 2015-07-02
Also published as: CN104735117A

Abstract

Disclosed are an information processing method, device and system, the method comprising: receiving multiple alternative information sent by one or more deployed terminals; a server off-line calculates the weight of each alternative information according to an estimated click rate of the alternative information, the off-line calculation referring to the calculation performed when a user client is in an off-line state; screening alternative information with the weight satisfying a preset first rule, and recording in a preselected reference list comprising at least two alternative information; acquiring characteristic information of the user operating the client when the client is in an on-line state; selecting alternative information satisfying a preset second rule from the preselected reference list according to the characteristic information; and pushing the alternative information satisfying the preset second rule in the preselected reference list to the client.

Description

一种信息处理方法、装置和***Information processing method, device and system

本申请要求于2013年12月24日提交中国专利局、申请号为201310723793.5、发明名称为“一种信息处理方法、装置和***”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。The present application claims priority to Chinese Patent Application No. 201310723793.5, entitled "Information Processing Method, Apparatus and System" on December 24, 2013, the entire contents of which are incorporated herein by reference. In the application.

技术领域Technical field

本发明涉及互联网技术，特别涉及一种信息处理方法、装置和***。The present invention relates to Internet technologies, and in particular, to an information processing method, apparatus, and system.

背景技术Background technique

目前，设置在网络侧的服务器会根据客户端用户的特征信息，如性别、爱好、地域等，在特定的时间从备选信息中选取相关的备选信息并向客户端推送，客户端显示所述选取出的备选信息，以供用户参考。Currently, the server set on the network side selects relevant candidate information from the candidate information and pushes it to the client at a specific time according to the feature information of the client user, such as gender, hobbies, geography, etc., and the client displays the location. The selected alternative information is provided for the user's reference.

但是，由于备选信息的数据量巨大，服务器从备选信息中选取相关的备选信息需要很大的运算量，导致选取时间较长，无法满足实际需求。而如果缩短选取时间，则需要对服务器进行升级，导致成本增加。However, due to the huge amount of data of the alternative information, the server needs to calculate a large amount of computational information from the candidate information, which results in a long selection time and cannot meet the actual demand. If the selection time is shortened, the server needs to be upgraded, resulting in an increase in cost.

发明内容Summary of the invention

为解决现有存在的技术问题，本发明实施例提供一种信息处理方法、装置和***。In order to solve the existing technical problems, an embodiment of the present invention provides an information processing method, apparatus, and system.

本发明提供一种信息处理方法，所述方法包括：接收多个备选信息，该多个备选信息由一个或多个投放终端发送；服务器按照所述备选信息的预估点击率离线计算出各备选信息的权重，所述离线计算是指用户所在的客户端为离线状态下执行的计算；筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息；获取所述客户端为在线状态下，操作该客户端的用户的特征信息；根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息；及将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。The present invention provides an information processing method, the method comprising: receiving a plurality of candidate information, the plurality of candidate information being sent by one or more delivery terminals; and the server calculating offline according to the estimated click rate of the candidate information The weight of each candidate information is calculated. The offline calculation refers to the calculation performed by the client where the user is located in an offline state; and the candidate letter whose weight meets the preset first rule is filtered out. And recorded in a pre-selected reference list, the pre-selected reference list includes at least two candidate information; acquiring feature information of the user who operates the client in an online state; and according to the feature information, Selecting candidate information that satisfies the preset second rule is selected in the pre-selected reference list; and the candidate information that satisfies the preset second rule in the pre-selected reference list is pushed to the client.

在一种实现方式中，所述离线计算出各备选信息的权重为：根据备选信息的预估点击率和备选信息的价格信息，离线计算出各备选信息的权重。In an implementation manner, the weight of each candidate information is calculated offline: the weight of each candidate information is calculated offline according to the estimated click rate of the candidate information and the price information of the candidate information.

在一种实现方式中，所述离线计算出各备选信息的权重之前，所述方法还包括：计算各备选信息的预估点击率。In an implementation manner, before the calculating the weight of each candidate information offline, the method further includes: calculating an estimated click rate of each candidate information.

在一种实现方式中，所述计算各备选信息的预估点击率包括：当所述备选信息在预设时间内未被推送时，将所述备选信息所属类别的平均点击率作为该备选信息的预估点击率。In an implementation manner, the calculating an estimated click rate of each candidate information includes: when the candidate information is not pushed within a preset time, using an average click rate of the category to which the candidate information belongs The estimated clickthrough rate for this alternate information.

在一种实现方式中，所述计算各备选信息的预估点击率包括：当所述备选信息在预设时间内被推送时，判断所述备选信息的被推送次数是否大于等于第一阈值；当所述备选信息的被推送次数大于等于第一阈值时，将所述备选信息在预设时间内的点击率作为该备选信息的预估点击率。In an implementation manner, the calculating an estimated click rate of each candidate information includes: when the candidate information is pushed within a preset time, determining whether the number of times the candidate information is pushed is greater than or equal to a threshold value; when the pushed number of the candidate information is greater than or equal to the first threshold, the click rate of the candidate information in the preset time is used as the estimated click rate of the candidate information.

在一种实现方式中，所述计算各备选信息的预估点击率还包括：当所述备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第一预设值的第一关联备选信息的被推送次数是否大于等于第一阈值，当所述第一关联备选信息的被推送次数大于等于第一阈值时，将所述第一关联备选信息在预设时间内的点击率作为所述备选信息的预估点击率。In an implementation manner, the calculating the estimated click rate of each candidate information further includes: when the pushed number of the candidate information is less than the first threshold, determining that the degree of association with the candidate information reaches the first Whether the number of times of pushing the first associated candidate information of a preset value is greater than or equal to the first threshold, and when the number of times of pushing the first associated candidate information is greater than or equal to the first threshold, the first associated candidate The click rate of the information within the preset time is used as the estimated click rate of the candidate information.

在一种实现方式中，所述计算各备选信息的预估点击率还包括：当所述第一关联备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第二预设值的第二关联备选信息的被推送次数是否大于等于第一阈值，当所述第二关联备选信息的被推送次数大于等于第一阈值时，将所述第二关联备选信息在预设时间内的点击率作为所述备选信息的预估点击率；当所述第二关联备选信息的被推送次数小于第一阈值时，将所述备选信息所属类别的平均点击率作为预估的点击率。In an implementation manner, the calculating the estimated click rate of each candidate information further includes: when the pushed number of the first associated candidate information is less than the first threshold, determining the candidate letter Whether the number of times of pushing the second association candidate information of the second preset value is greater than or equal to the first threshold, and when the number of times the second association candidate information is pushed is greater than or equal to the first threshold, The click rate of the second association candidate information in the preset time is used as the estimated click rate of the candidate information; when the pushed number of the second association candidate information is less than the first threshold, the preparation The average click rate of the category to which the information belongs is used as the estimated click rate.

在一种实现方式中，所述特征信息包括以下信息中的至少一种：操作所述客户端的用户的性别信息、爱好信息、地域信息。In an implementation manner, the feature information includes at least one of the following information: gender information, hobby information, and region information of a user who operates the client.

本发明还提供了一种服务器，所述服务器包括：一个或多个处理器；及存储设备，用于存储指令，该指令由所述一个或多个处理器执行，以实现信息处理方法：接收多个备选信息，该多个备选信息由一个或多个投放终端发送；服务器按照所述备选信息的预估点击率离线计算出各备选信息的权重，所述离线计算是指用户所在的客户端为离线状态下执行的计算；筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息；获取所述客户端为在线状态下，操作该客户端的用户的特征信息；根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息；将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。The present invention also provides a server, the server comprising: one or more processors; and a storage device for storing instructions executed by the one or more processors to implement an information processing method: receiving a plurality of candidate information, which is sent by one or more delivery terminals; the server calculates the weight of each candidate information offline according to the estimated click rate of the candidate information, where the offline calculation refers to the user The client is located in the offline state; the candidate information is filtered out to meet the preset first rule, and is recorded in a pre-selected reference list, where the pre-selected reference list includes at least two candidate information; The client is in the online state, and operates the feature information of the user of the client; the candidate information that meets the preset second rule is selected from the pre-selected reference list according to the feature information; and the pre-selected reference list meets the preset The alternative information of the second rule is pushed to the client.

在一种实现方式中，所述服务器的存储设备中还包括由所述一个或多个处理器执行以实现以下步骤的指令：在所述离线计算出各备选信息的权重之前，计算各备选信息的预估点击率。In an implementation manner, the storage device of the server further includes an instruction executed by the one or more processors to implement the following steps: calculating each device before calculating the weight of each candidate information offline The estimated clickthrough rate of the selected information.

在一种实现方式中，所述计算各备选信息的预估点击率包括：当所述备选信息在预设时间内未被推送时，将所述备选信息所属类别的平均点击率作为该备选信息的预估点击率。In an implementation manner, the calculating an estimated click rate of each candidate information includes: when the candidate information is not pushed within a preset time, an average point of a category to which the candidate information belongs The hit rate is used as the estimated click rate for this alternate information.

在一种实现方式中，所述计算各备选信息的预估点击率还包括：当所述第一关联备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第二预设值的第二关联备选信息的被推送次数是否大于等于第一阈值，当所述第二关联备选信息的被推送次数大于等于第一阈值时，将所述第二关联备选信息在预设时间内的点击率作为所述备选信息的预估点击率；当所述第二关联备选信息的被推送次数小于第一阈值时，将所述备选信息所属类别的平均点击率作为预估的点击率。In an implementation manner, the calculating the estimated click rate of each candidate information further includes: determining, when the pushed number of the first associated candidate information is less than the first threshold, determining the association with the candidate information Whether the number of times of pushing the second association candidate information of the second preset value is greater than or equal to the first threshold, and when the number of times of pushing the second association candidate information is greater than or equal to the first threshold, the second Correlating the click rate of the candidate information in the preset time as the estimated click rate of the candidate information; when the pushed number of the second associated candidate information is less than the first threshold, the candidate information belongs to The average clickthrough rate for the category is used as the estimated clickthrough rate.

本发明还提供了一种用于信息处理的计算机存储介质，其上存储有指令集，所述指令集由一个或多个处理器执行而执行以下步骤：接收多个备选信息，该多个备选信息由一个或多个投放终端发送；服务器按照所述备选信息的预估点击率离线计算出各备选信息的权重，所述离线计算是指用户所在的客户端为离线状态下执行的计算；筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息；获取所述客户端为在线状态下，操作该客户端的用户的特征信息；根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息；以及将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。The present invention also provides a computer storage medium for information processing having stored thereon a set of instructions executed by one or more processors to perform the steps of: receiving a plurality of candidate information, the plurality of The candidate information is sent by one or more delivery terminals; the server calculates the weight of each candidate information offline according to the estimated click rate of the candidate information, where the offline calculation refers to the execution of the client where the client is offline. Calculation; filter out the weight to meet the preset first rule The candidate information is recorded in a pre-selected reference list, where the pre-selected reference list includes at least two candidate information; acquiring feature information of the user who operates the client in the online state; according to the feature And selecting, from the pre-selected reference list, candidate information that meets the preset second rule; and pushing the candidate information that meets the preset second rule in the pre-selected reference list to the client.

在一种实现方式中，所述离线计算出各备选信息的权重之前，还包括：计算各备选信息的预估点击率。In an implementation manner, before the calculating the weight of each candidate information offline, the method further includes: calculating an estimated click rate of each candidate information.

由上可知，通过预先选取出权重满足预设第一规则的备选信息记录在一个预选参考列表中，当用户在线时再从所述预选参考列表中选取满足预设第二规则的备选信息并推送给客户端的用户，由于预选参考列表中的备选信息的数据量较小，因而服务器从预选参考列表中选取相关的备选信息需要的运算量大大减小，从而能够在不升级服务器的情况下，减少备选信息选取时间。It can be seen that, by selecting the candidate information that satisfies the preset first rule in advance, the candidate information is recorded in a pre-selected reference list, and when the user is online, the candidate information that meets the preset second rule is selected from the pre-selected reference list. And the user pushed to the client, because the amount of data of the candidate information in the pre-selected reference list is small, the amount of computation required by the server to select the relevant candidate information from the pre-selected reference list is greatly reduced, so that the server can be upgraded without upgrading. In case, reduce the time for selecting alternative information.

附图说明DRAWINGS

图1为本发明提供的一种信息处理方法的第一实施例的流程示意图；1 is a schematic flow chart of a first embodiment of an information processing method according to the present invention;

图2为本发明提供的一种信息处理方法的第二实施例的流程示意图；2 is a schematic flowchart of a second embodiment of an information processing method according to the present invention;

图3为本发明提供的一种信息处理方法的第三实施例的流程示意图； FIG. 3 is a schematic flowchart diagram of a third embodiment of an information processing method according to the present invention; FIG.

图4为本发明提供的一种信息处理方法中备选信息预选的实施例的流程示意图；4 is a schematic flowchart of an embodiment of pre-selection of candidate information in an information processing method according to the present invention;

图5为本发明提供的服务器的实施例的结构示意图；FIG. 5 is a schematic structural diagram of an embodiment of a server provided by the present invention; FIG.

图6为本发明提供的客户端的实施例的结构示意图；6 is a schematic structural diagram of an embodiment of a client provided by the present invention;

图7为本发明提供的一种信息处理***的实施例的结构示意图。FIG. 7 is a schematic structural diagram of an embodiment of an information processing system according to the present invention.

具体实施方式detailed description

目前，当小数据(如广告)曝光发生时，设置在网络侧的电子设备(如网络服务器)会根据在线用户的特征信息(如性别、爱好、地域等)选取最合适的、最有点击倾向的小数据对该用户进行播放。At present, when small data (such as advertising) exposure occurs, electronic devices (such as web servers) set on the network side select the most appropriate and most click-oriented tendency according to the online user's characteristic information (such as gender, hobbies, regions, etc.). The small data is played to the user.

通常备选的小数据具备预算小、数据多的特点，尤其是每次点击付费(Cost Per Click，CPC)的小数据，会导致同一小数据展现位置上存在上万甚至上十万个小数据竞价播放。Usually the alternative small data has the characteristics of small budget and many data, especially the small data of Cost Per Click (CPC), which will result in tens of thousands or even hundreds of thousands of small data in the same small data display position. Auction play.

现有技术方案根据曝光发生时在线用户的特征信息从数量巨大的备选的小数据中筛选最合适的、最有点击倾向的小数据，运算量巨大。为了保证在极短的时间(通常为毫秒级)内完成运算，往往需要使用上百台服务器构成的服务器集群来完成运算，成本过高。并且随着小数据的持续增长，还需要进一步扩充服务器集群或升级服务器，导致运营成本和服务器成本增加。The prior art solution selects the most suitable and most click-oriented small data from the huge amount of alternative small data according to the characteristic information of the online user when the exposure occurs, and the calculation amount is huge. In order to ensure that the operation is completed in a very short time (usually milliseconds), it is often necessary to use a server cluster composed of hundreds of servers to complete the calculation, which is costly. And as the small data continues to grow, it is necessary to further expand the server cluster or upgrade the server, resulting in increased operating costs and server costs.

本发明提供的一种信息处理方法的第一实施例，应用于服务器，如图1所示，所述方法包括：A first embodiment of an information processing method provided by the present invention is applied to a server. As shown in FIG. 1, the method includes:

步骤101、接收多个备选信息，该多个备选信息由一个或多个投放终端发送。Step 101: Receive a plurality of candidate information, where the multiple candidate information is sent by one or more delivery terminals.

这里，在具体应用场景中，所述备选信息可以是广告主投放的广告订单。CPC的广告订单具备预算小、数据多的特点，同一广告展现位置上会存在上万甚至上十万个订单竞价播放。Here, in a specific application scenario, the candidate information may be an advertisement order placed by an advertiser. CPC's advertising order has the characteristics of small budget and many data, and the same advertisement is displayed. There will be tens of thousands or even hundreds of thousands of orders bidding.

步骤102、服务器按照所述备选信息的预估点击率离线计算出各备选信息的权重，所述离线计算是指用户所在的客户端为离线状态下服务器执行的计算。Step 102: The server calculates the weight of each candidate information offline according to the estimated click rate of the candidate information, where the offline calculation refers to the calculation performed by the server that the user is located in the offline state.

这里，备选信息的权重指的是该备选信息在所有备选信息中被筛选出推送给客户的可能性程度，可与该备选信息的历史点击率、价格等有关。在一种实现方式中，备选信息的预估点击率越高，该备选信息的权重也就越高。在另一种实现方式中，备选信息的预估点击率越高并且备选信息的价格信息越高，该备选信息的权重也就越高。Here, the weight of the alternative information refers to the degree to which the candidate information is filtered out to be pushed to the client in all the candidate information, and may be related to the historical click rate, price, and the like of the candidate information. In one implementation, the higher the estimated click rate of the candidate information, the higher the weight of the candidate information. In another implementation, the higher the estimated click rate of the candidate information and the higher the price information of the candidate information, the higher the weight of the candidate information.

在用户所在的客户端为离线状态下执行该计算，可以提前对备选信息进行第一次筛选。之后，当客户端在线时，在第一次筛选出的备选信息中进行再次筛选所需要的计算量就大大减小了。Performing this calculation while the client where the user is located is offline, and the candidate information can be filtered for the first time in advance. Later, when the client is online, the amount of computation required to re-screen in the first filtered candidate information is greatly reduced.

步骤103、筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息。Step 103: Filter out the candidate information that the weight meets the preset first rule, and record the information in a pre-selected reference list, where the pre-selected reference list includes at least two candidate information.

这里，在实际应用中，可以依据权重的大小对权重进行排序，筛选出权重最高的一组订单作为预选订单，如从10000个订单中选取排序前1000位的订单。Here, in practical applications, the weights may be sorted according to the weight of the weights, and a group of orders with the highest weight is selected as a pre-selected order, for example, an order of 1000 orders is selected from 10,000 orders.

步骤104、获取所述客户端为在线状态下，操作该客户端的用户的特征信息。Step 104: Acquire feature information of a user who operates the client in an online state.

用户的特征信息可以包括用户的历史记录等信息，也可以包括根据用户的历史记录等信息对用户的性别、爱好、地域等进行的分类。The feature information of the user may include information such as a history of the user, and may also include classification of the gender, hobbies, regions, and the like of the user according to information such as the history of the user.

步骤105、根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息。Step 105: Select, according to the feature information, candidate information that meets a preset second rule from the pre-selected reference list.

这里，在具体应用场景中，可以根据性别、爱好、地域等特征信息从预选订单中选取一个或多个点击倾向较大的订单并推送给客户端，以使客户端显示收到的订单，所述订单可以点击播放。例如，所述第二规则可以为用户爱好与预选参考列表中的备选信息的拟合程度大于预定值，因此，可以根据用户特征从所述预选参考列表中选取与用户爱好的拟合程度大于预定值的备选信息。Here, in a specific application scenario, one or more orders with a large click preference may be selected from the pre-selected orders according to the feature information such as gender, hobbies, and regions, and pushed to the client, so that The client displays the received order, which can be clicked to play. For example, the second rule may be that the degree of fitting of the user preference with the candidate information in the pre-selected reference list is greater than a predetermined value, and therefore, the degree of fitting from the pre-selected reference list to the user's preference may be greater than the user feature. Alternative information for the predetermined value.

步骤106、将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。Step 106: Push the candidate information in the pre-selected reference list that meets the preset second rule to the client.

由此，本发明实施例预先选取出权重满足预设第一规则的备选信息记录在一个预选参考列表中，当用户在线时再从所述预选参考列表中选取满足预设第二规则的备选信息并推送，由于预选参考列表中的备选信息的数据量较小，因而服务器从预选参考列表中选取相关的备选信息需要的运算量大大减小，从而能够在不升级服务器的情况下，减少备选信息选取时间。Therefore, in the embodiment of the present invention, the candidate information that the weights satisfy the preset first rule is recorded in a pre-selected reference list, and when the user is online, the device that meets the preset second rule is selected from the pre-selected reference list. The information is selected and pushed. Since the amount of data of the candidate information in the pre-selected reference list is small, the amount of computation required by the server to select the relevant candidate information from the pre-selected reference list is greatly reduced, so that the server can be upgraded without upgrading the server. , reduce the time for selecting alternative information.

本发明提供的一种信息处理方法的第二实施例，应用于服务器，如图2所示，所述方法包括：A second embodiment of an information processing method provided by the present invention is applied to a server. As shown in FIG. 2, the method includes:

步骤201、接收多个备选信息，该多个备选信息由一个或多个投放终端发送。Step 201: Receive multiple candidate information, where the multiple candidate information is sent by one or more delivery terminals.

步骤202、计算各备选信息的预估点击率。Step 202: Calculate an estimated click rate of each candidate information.

这里，在具体应用场景中，可以根据订单所属的行业和订单投放的素材尺寸和广告位，预估一个点击率。Here, in a specific application scenario, a click rate can be estimated based on the size and the ad slot of the industry and the order to which the order belongs.

步骤203、按照所述备选信息的预估点击率离线计算出各备选信息的权重，所述离线计算是指用户所在的客户端为离线状态下执行的计算。Step 203: Calculate the weight of each candidate information offline according to the estimated click rate of the candidate information, where the offline calculation refers to the calculation performed by the client where the user is located in an offline state.

这里，在实际应用中，还可以根据各备选信息的预估点击率和价格信息，离线计算出各备选信息的权重。所述价格信息可以是CPC的订单出价。Here, in practical applications, the weight of each candidate information may also be calculated offline according to the estimated click rate and price information of each candidate information. The price information may be an order bid of the CPC.

步骤204、筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息。Step 204: Filter out the candidate information that the weight meets the preset first rule, and record in a pre- In the selected reference list, the pre-selected reference list includes at least two alternative information.

这里，在实际应用中，可以依据权重的大小对权重进行排序，筛选出权重最高的一组订单，如从10000个订单中选取排序前1000位的订单。Here, in practical applications, the weights can be sorted according to the weight of the weights, and a group of orders with the highest weight is selected, for example, the orders of the top 1000 are selected from 10,000 orders.

步骤205、获取所述客户端为在线状态下，操作该客户端的用户的特征信息。Step 205: Acquire feature information of a user who operates the client in an online state.

步骤206、根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息。Step 206: Select, according to the feature information, candidate information that meets a preset second rule from the pre-selected reference list.

步骤207、将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。Step 207: Push the candidate information that meets the preset second rule in the pre-selected reference list to the client.

由此，本发明实施例通过预估各备选信息的点击率，可以更准确地离线计算出各备选信息的权重，从而有利于更准确的选取出相关的备选信息。Therefore, in the embodiment of the present invention, by estimating the click rate of each candidate information, the weight of each candidate information can be calculated more accurately offline, thereby facilitating more accurate selection of relevant candidate information.

本发明提供的一种信息处理方法的第三实施例，应用于服务器，如图3所示，所述方法包括：A third embodiment of an information processing method provided by the present invention is applied to a server. As shown in FIG. 3, the method includes:

步骤301、接收多个备选信息，该多个备选信息由一个或多个投放终端发送。Step 301: Receive multiple candidate information, where the multiple candidate information is sent by one or more delivery terminals.

步骤302、判断备选信息在预设时间内是否被推送，当所述备选信息被推送时，进入步骤303；当所述备选信息未被推送时，进入步骤309。Step 302: Determine whether the candidate information is pushed within a preset time. When the candidate information is pushed, the process proceeds to step 303; when the candidate information is not pushed, the process proceeds to step 309.

步骤303、判断所述备选信息的被推送次数是否大于等于第一阈值，当所述备选信息的被推送次数大于等于第一阈值时，进入步骤304；当所述备选信息的被推送次数小于第一阈值时，进入步骤305。Step 303: Determine whether the number of times the candidate information is pushed is greater than or equal to the first threshold. When the number of times the candidate information is pushed is greater than or equal to the first threshold, proceed to step 304; when the candidate information is pushed When the number of times is less than the first threshold, the process proceeds to step 305.

这里，在具体应用场景中，所述第一阈值可以是采样阈值，所述采样阈值＝1/广告平均点击率*采样系数，所述广告平均点击率可以由服务器通过统计得出，所述采样系数为预设的值，例如为5或10等值。广告点击率是指显示广告中被点击的广告占显示广告的比例。 Here, in a specific application scenario, the first threshold may be a sampling threshold, the sampling threshold=1/advertising average click rate* sampling coefficient, and the average advertising click rate may be obtained by a server through statistics, the sampling The coefficient is a preset value, for example, a value of 5 or 10. The ad clickthrough rate is the percentage of ads that are clicked in the display ad.

步骤304、将所述备选信息在预设时间内的点击率作为预估的点击率，进入步骤310。Step 304: The click rate of the candidate information in the preset time is used as the estimated click rate, and the process proceeds to step 310.

这里，在实际应用中，所述预设时间可以根据实际情况设定，例如可以是7天或10天等。Here, in an actual application, the preset time may be set according to actual conditions, for example, may be 7 days or 10 days, and the like.

步骤305、判断与所述备选信息的关联度达到第一预设值的第一关联备选信息的被推送次数是否大于等于第一阈值，当所述第一关联备选信息的被推送次数大于等于第一阈值时，进入步骤306；当所述第一关联备选信息的被推送次数小于第一阈值时，进入步骤307。Step 305: Determine whether the number of times of pushing the first association candidate information that is related to the candidate information reaches the first preset value is greater than or equal to the first threshold, when the number of times of the first association candidate information is pushed. When the first threshold is greater than or equal to the first threshold, the process proceeds to step 306; when the number of times the first association candidate information is pushed is less than the first threshold, the process proceeds to step 307.

这里，在实际应用中，与所述备选信息的关联度达到第一预设值的第一关联备选信息可以是同一广告主的相似订单。Here, in an actual application, the first association candidate information whose degree of association with the candidate information reaches the first preset value may be a similar order of the same advertiser.

步骤306、将所述第一关联备选信息在预设时间内的点击率作为预估的点击率，进入步骤310。Step 306: The click rate of the first association candidate information in the preset time is used as the estimated click rate, and the process proceeds to step 310.

步骤307、判断与所述备选信息的关联度达到第二预设值的第二关联备选信息的被推送次数是否大于等于第一阈值，当所述第二关联备选信息的被推送次数大于等于第一阈值时，进入步骤308；当所述第二关联备选信息的被推送次数小于第一阈值时，进入步骤309。Step 307: Determine whether the pushed number of times of the second association candidate information that is related to the candidate information reaches the second preset value is greater than or equal to the first threshold, and when the second association candidate information is pushed. When the threshold is greater than or equal to the first threshold, the process proceeds to step 308; when the number of times the second association candidate information is pushed is less than the first threshold, the process proceeds to step 309.

这里，在实际应用中，与所述备选信息的关联度达到第二预设值的第二关联备选信息可以是广告主所在行业的相似订单。Here, in an actual application, the second association candidate information whose degree of association with the candidate information reaches the second preset value may be a similar order of the industry in which the advertiser is located.

步骤308、将所述第二关联备选信息在预设时间内的点击率作为预估的点击率，进入步骤310。Step 308: The click rate of the second associated candidate information in the preset time is used as the estimated click rate, and the process proceeds to step 310.

步骤309、将所述备选信息所属类别的平均点击率作为预估的点击率。Step 309: The average click rate of the category to which the candidate information belongs is used as the estimated click rate.

这里，在实际应用中，所述备选信息所属类别的平均点击率可以是订单所在播放形式或广告位的平均点击率。Here, in an actual application, the average click rate of the category to which the candidate information belongs may be the playback form of the order or the average click rate of the advertisement slot.

步骤310、根据各备选信息的点击率和价格信息，离线计算出各备选信息的权重。Step 310: Calculate each candidate offline according to the click rate and price information of each candidate information. The weight of the information.

步骤311、筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息。Step 311: Filter out the candidate information that the weight meets the preset first rule, and record it in a pre-selected reference list, where the pre-selected reference list includes at least two candidate information.

步骤312、获取所述客户端为在线状态下，操作该客户端的用户的特征信息。Step 312: Acquire feature information of a user who operates the client in an online state.

步骤313、根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息。Step 313: Select, according to the feature information, candidate information that meets a preset second rule from the pre-selected reference list.

步骤314、将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。Step 314: Push the candidate information that meets the preset second rule in the pre-selected reference list to the client.

由此，本发明实施例通过备选信息在预设时间内是否被推送、所述备选信息的被推送次数是否大于等于第一阈值、与所述备选信息的关联度达到第一预设值的第一关联备选信息的被推送次数是否大于等于第一阈值、与所述备选信息的关联度达到第二预设值的第二关联备选信息的被推送次数是否大于等于第一阈值等信息预估点击率，可以得到更准确的预估点击率，从而有利于更准确的选取出预选相关的备选信息。Therefore, in the embodiment of the present invention, whether the candidate information is pushed in the preset time, whether the number of times the candidate information is pushed is greater than or equal to the first threshold, and the degree of association with the candidate information reaches the first preset. Whether the number of times of pushing the first associated candidate information of the value is greater than or equal to the first threshold, and whether the number of times of pushing the second associated candidate information that has reached the second preset value with the candidate information is greater than or equal to the first Threshold and other information to estimate the click-through rate, you can get a more accurate estimated click-through rate, which is more conducive to more accurate selection of pre-selection related alternative information.

下面结合图4以一具体应用场景为例对本发明的一种信息处理方法中备选信息预选的流程进行介绍。本实施例中备选信息是广告主(客户)投放的广告订单。The flow of the candidate information pre-selection in an information processing method of the present invention will be described below by taking a specific application scenario as an example. The alternative information in this embodiment is an advertisement order placed by an advertiser (customer).

步骤401、判断在播的订单在7天内是否有曝光，当有时，进入步骤402；当没有时，进入步骤408。Step 401: Determine whether the broadcasted order has exposure within 7 days, and when it is, go to step 402; if not, go to step 408.

步骤402、判断曝光次数是否大于等于采样阈值，当大于等于采样阈值时，进入步骤403；当小于采样阈值时，进入步骤404。Step 402: Determine whether the number of exposures is greater than or equal to the sampling threshold. When the sampling threshold is greater than or equal to the sampling threshold, proceed to step 403. If the sampling threshold is less than the sampling threshold, proceed to step 404.

这里，所述采样阈值＝1/广告平均点击率*采样系数，所述广告平均点击率可以由服务器通过统计得出，所述采样系数为预设的值，例如为5或10等值。广告点击率是指显示广告中被点击的广告占显示广告的比例。 Here, the sampling threshold = 1 / advertising average click rate * sampling coefficient, the average advertising click rate can be obtained by the server through statistics, the sampling coefficient is a preset value, for example, 5 or 10 equal value. The ad clickthrough rate is the percentage of ads that are clicked in the display ad.

步骤403，将所述订单7天历史点击率作为预估的点击率，进入步骤409。Step 403: The 7-day historical click rate of the order is used as the estimated click rate, and the process proceeds to step 409.

步骤404、判断客户是否有相似订单在7天内曝光次数大于等于采样阈值，当有时，进入步骤405；当没有时，进入步骤406。Step 404: Determine whether the customer has a similar order within 7 days, the number of exposures is greater than or equal to the sampling threshold, and when not, go to step 405; if not, go to step 406.

步骤405、将同客户相似订单7天历史点击率作为预估的点击率，进入步骤409。In step 405, the 7-day historical click rate of the customer similar order is used as the estimated click rate, and the process proceeds to step 409.

步骤406、判断客户所在行业是否有相似订单在7天曝光次数是否大于等于采样阈值，当有时，进入步骤407；当没有时，进入步骤408。Step 406: Determine whether the customer's industry has a similar order, whether the number of exposures in the 7 days is greater than or equal to the sampling threshold, and when not, proceed to step 407; if not, proceed to step 408.

步骤407、将同行业相似订单7天历史点击率作为预估的点击率，进入步骤409。Step 407: The 7-day historical click rate of the similar order in the same industry is taken as the estimated click rate, and the process proceeds to step 409.

步骤408、将订单所在播放形式或广告位的平均点击率作为预估的点击率。Step 408: The average click rate of the play form or the ad slot of the order is used as the estimated click rate.

步骤409、根据订单出价和预估的点击率计算订单的权重。Step 409: Calculate the weight of the order according to the order bid and the estimated click rate.

由此，可以根据计算出的权重将订单排序并进入或淘汰出订单池，得到预选订单。Thus, the orders can be sorted according to the calculated weights and entered or eliminated from the order pool to obtain a pre-selected order.

本发明提供的另一种信息处理方法的实施例，应用于客户端，所述方法包括：接收推送的备选信息；以及显示所述推送的备选信息。Another embodiment of the information processing method provided by the present invention is applied to a client, the method comprising: receiving push candidate information; and displaying the push candidate information.

所述推送的备选信息为客户端在线状态下根据客户端用户的特征信息从预选参考列表中选取的满足预设第二规则的备选信息，所述预选参考列表中记录了筛选出的权重满足预设第一规则的备选信息，其中备选信息的权重是按照备选信息的预估点击率离线计算出的。The candidate information of the push is candidate information that is selected from the pre-selected reference list according to the feature information of the client user in the online state, and the selected second reference rule is recorded, and the filtered weight is recorded in the pre-selected reference list. The candidate information satisfying the preset first rule, wherein the weight of the candidate information is calculated offline according to the estimated click rate of the candidate information.

所述应用于客户端的方法与上述应用于服务器的方法相对应，具体细节可以参见上述应用于服务器的方法。The method applied to the client corresponds to the above method applied to the server, and the specific details can be referred to the above method applied to the server.

本发明提供的一种服务器，如图5所示，所述服务器包括：A server provided by the present invention, as shown in FIG. 5, the server includes:

一个或多个处理器501；及 One or more processors 501; and

存储设备502，用于存储指令，该指令由所述一个或多个处理器执行，以实现：The storage device 502 is configured to store an instruction executed by the one or more processors to implement:

接收多个备选信息，该多个备选信息由一个或多个投放终端发送；Receiving a plurality of candidate information, the plurality of candidate information being sent by one or more delivery terminals;

服务器按照所述备选信息的预估点击率离线计算出各备选信息的权重，所述离线计算是指用户所在的客户端为离线状态下执行的计算；The server calculates the weight of each candidate information offline according to the estimated click rate of the candidate information, where the offline calculation refers to the calculation performed by the client where the user is located in an offline state;

筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息；Filtering out the candidate information that the weight meets the preset first rule, and records the information in a pre-selected reference list, where the pre-selected reference list includes at least two candidate information;

获取所述客户端为在线状态下，操作该客户端的用户的特征信息；Obtaining feature information of the user who operates the client in the online state;

根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息；及Selecting, from the pre-selected reference list, candidate information that meets a preset second rule according to the feature information; and

将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。The candidate information in the pre-selected reference list that satisfies the preset second rule is pushed to the client.

在一实施例中，所述离线计算出各备选信息的权重为：根据备选信息的预估点击率和备选信息的价格信息，离线计算出各备选信息的权重。In an embodiment, the weight of each candidate information is calculated offline: the weight of each candidate information is calculated offline according to the estimated click rate of the candidate information and the price information of the candidate information.

在一实施例中，所述服务器的存储设备中还包括由所述一个或多个处理器执行以实现以下步骤的指令：In an embodiment, the storage device of the server further includes instructions executed by the one or more processors to implement the following steps:

所述离线计算出各备选信息的权重之前，计算各备选信息的预估点击率。Before the offline calculation of the weight of each candidate information, the estimated click rate of each candidate information is calculated.

在一实施例中，所述计算各备选信息的预估点击率包括：判断备选信息在预设时间内是否被推送，当所述备选信息被推送时，判断所述备选信息的被推送次数是否大于等于第一阈值，当所述备选信息的被推送次数大于等于第一阈值时，将所述备选信息在预设时间内的点击率作为预估的点击率；In an embodiment, the calculating the estimated click rate of each candidate information comprises: determining an alternative Whether the information is pushed in a preset time, and when the candidate information is pushed, determining whether the number of times of pushing the candidate information is greater than or equal to a first threshold, when the number of times the candidate information is pushed is greater than or equal to a threshold value, the click rate of the candidate information in a preset time is used as an estimated click rate;

当所述备选信息未被推送时，将所述备选信息所属类别的平均点击率作为预估的点击率。When the candidate information is not pushed, the average click rate of the category to which the candidate information belongs is used as the estimated click rate.

在一实施例中，所述计算各备选信息的预估点击率包括：当所述备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第一预设值的第一关联备选信息的被推送次数是否大于等于第一阈值，当所述第一关联备选信息的被推送次数大于等于第一阈值时，将所述第一关联备选信息在预设时间内的点击率作为预估的点击率。In an embodiment, the calculating the estimated click rate of each candidate information comprises: when the pushed number of the candidate information is less than the first threshold, determining that the degree of association with the candidate information reaches the first pre- Whether the number of times the first associated candidate information is pushed is greater than or equal to the first threshold, and when the pushed number of the first associated candidate information is greater than or equal to the first threshold, the first associated candidate information is The click rate in the preset time is used as the estimated click rate.

在一实施例中，所述计算各备选信息的预估点击率包括：当所述第一关联备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第二预设值的第二关联备选信息的被推送次数是否大于等于第一阈值，当所述第二关联备选信息的被推送次数大于等于第一阈值时，将所述第二关联备选信息在预设时间内的点击率作为预估的点击率；In an embodiment, the calculating the estimated click rate of each candidate information comprises: when the pushed number of the first associated candidate information is less than the first threshold, determining that the degree of association with the candidate information is reached Whether the number of times of the second association candidate information of the second preset value is greater than or equal to the first threshold, and when the number of times the second association candidate information is pushed is greater than or equal to the first threshold, the second association is prepared. Select the click rate of the information within the preset time as the estimated click rate;

当所述第二关联备选信息的被推送次数小于第一阈值时，将所述备选信息所属类别的平均点击率作为预估的点击率。When the number of times the second association candidate information is pushed is less than the first threshold, the average click rate of the category to which the candidate information belongs is used as the estimated click rate.

本发明提供的一种客户端的实施例，如图6所示，所述客户端包括：An embodiment of the client provided by the present invention is as shown in FIG. 6. The client includes:

一个或多个处理器601；及One or more processors 601; and

存储设备602，用于存储指令，该指令由所述一个或多个处理器执行，以实现：The storage device 602 is configured to store an instruction executed by the one or more processors to implement:

接收推送的备选信息，所述推送的备选信息为客户端在线状态下根据客户端用户的特征信息从预选参考列表中选取的满足预设第二规则的备选信息，所述预选参考列表中记录了筛选出的权重满足预设第一规则的备选信息，其中备选信息的权重是按照备选信息的预估点击率离线计算出的；以及Receiving the push candidate information, where the pushed candidate information is candidate information that is selected from the pre-selected reference list and meets the preset second rule according to the feature information of the client user in the online state, the pre-selected reference list. Recording the filtered weights to meet the preset first rule Selecting information, wherein the weight of the candidate information is calculated offline according to the estimated click rate of the candidate information;

显示所述推送的备选信息。The candidate information of the push is displayed.

本发明提供的一种信息处理***的实施例，如图7所示，所述***包括服务器701和客户端702。An embodiment of an information processing system provided by the present invention, as shown in FIG. 7, includes a server 701 and a client 702.

所述服务器701，用于接收多个备选信息，该多个备选信息由一个或多个投放终端发送；按照所述备选信息的预估点击率离线计算出各备选信息的权重，所述离线计算是指用户所在的客户端为离线状态下执行的计算；筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息；获取所述客户端为在线状态下，操作该客户端的用户的特征信息；根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息；以及将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。The server 701 is configured to receive multiple candidate information, where the multiple candidate information is sent by one or more delivery terminals; and calculate the weight of each candidate information offline according to the estimated click rate of the candidate information. The offline calculation refers to the calculation performed by the client where the user is located in an offline state; the candidate information whose weight meets the preset first rule is selected and recorded in a pre-selected reference list, and the pre-selected reference list includes at least two Optional information; acquiring feature information of the user who operates the client in the online state; selecting candidate information that meets the preset second rule from the pre-selected reference list according to the feature information; The candidate information in the pre-selected reference list that satisfies the preset second rule is pushed to the client.

所述客户端702，用于接收并显示从所述预选参考列表中选取满足预设第二规则的备选信息。The client 702 is configured to receive and display, by using the pre-selected reference list, candidate information that meets a preset second rule.

综上所述，本发明实施例在客户端离线时预先选取出权重满足预设第一规则的备选信息记录在一个预选参考列表中，例如几百个或一千个订单，当客户端在线时再从所述预选参考列表中选取满足预设第二规则的备选信息，由于预选参考列表中的备选信息的数据量较小，因而服务器从预选参考列表中选取相关的备选信息需要的运算量大大减小，从而能够在不升级服务器(升级服务器指增加服务器集群中服务器的数量或提升单个服务器的性能)的情况下，减少备选信息选取时间，例如控制选取时间在毫秒级。本发明在客户端离线时，即没有客户端用户的特征信息的情况下，预先按订单所属的行业和相似订单等信息对订单进行权重排序，在此基础上筛选出权重最高的一定数量的订单，能够有效节省服务器开销，并解决订单数量无限增长的问题。In summary, in the embodiment of the present invention, when the client is offline, the candidate information whose weight is determined to meet the preset first rule is recorded in a pre-selected reference list, for example, hundreds or one thousand orders, when the client is online. Then, the candidate information that satisfies the preset second rule is selected from the pre-selected reference list. Because the data amount of the candidate information in the pre-selected reference list is small, the server needs to select the relevant candidate information from the pre-selected reference list. The amount of computation is greatly reduced, so that the candidate information selection time can be reduced without upgrading the server (the upgrade server refers to increasing the number of servers in the server cluster or improving the performance of a single server), for example, the control selection time is in the order of milliseconds. When the client is offline, that is, without the feature information of the client user, the order is weighted according to the industry and similar orders to which the order belongs, and a certain number of orders with the highest weight are selected. Can effectively save server overhead and solve The problem of an unlimited number of orders.

本申请中的装置实施例与方法实施例相对应，因此没有对装置实施例进行详细的描述，具体的细节可以参照相应方法实施例的描述。The device embodiments in the present application correspond to the method embodiments, and thus the device embodiments are not described in detail. For specific details, refer to the description of the corresponding method embodiments.

本发明还提供了一种用于信息处理的计算机存储介质，其上存储有指令集，所述指令集由一个或多个处理器执行而执行以下步骤：The present invention also provides a computer storage medium for information processing having stored thereon a set of instructions executed by one or more processors to perform the following steps:

根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息；以及Selecting, from the pre-selected reference list, candidate information that meets a preset second rule according to the feature information;

在一实施例中，所述离线计算出各备选信息的权重之前，还包括：计算各备选信息的预估点击率。In an embodiment, before the calculating the weight of each candidate information offline, the method further includes: calculating an estimated click rate of each candidate information.

在一实施例中，所述计算各备选信息的预估点击率包括：当所述备选信息在预设时间内未被推送时，将所述备选信息所属类别的平均点击率作为该备选信息的预估点击率。In an embodiment, the calculating the estimated click rate of each candidate information comprises: when the candidate information is not pushed within a preset time, using an average click rate of the category to which the candidate information belongs Estimated clickthrough rate for alternate information.

在一实施例中，所述计算各备选信息的预估点击率包括：当所述备选信息在预设时间内被推送时，判断所述备选信息的被推送次数是否大于等于第一阈值；当所述备选信息的被推送次数大于等于第一阈值时，将所述备选信息在预设时间内的点击率作为该备选信息的预估点击率。In an embodiment, the calculating the estimated click rate of each candidate information comprises: determining, when the candidate information is pushed within a preset time, whether the number of times the candidate information is pushed is greater than or equal to the first Threshold; when the number of times the candidate information is pushed is greater than or equal to the first threshold, The click rate of the candidate information in the preset time is used as the estimated click rate of the candidate information.

在一实施例中，所述计算各备选信息的预估点击率还包括：当所述备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第一预设值的第一关联备选信息的被推送次数是否大于等于第一阈值，当所述第一关联备选信息的被推送次数大于等于第一阈值时，将所述第一关联备选信息在预设时间内的点击率作为所述备选信息的预估点击率。In an embodiment, the calculating the estimated click rate of each candidate information further includes: when the pushed number of the candidate information is less than the first threshold, determining that the degree of association with the candidate information reaches the first Whether the number of times of the first association candidate information of the preset value is greater than or equal to the first threshold, and when the number of times the first association candidate information is pushed is greater than or equal to the first threshold, the first association candidate information is used. The click rate in the preset time is used as the estimated click rate of the candidate information.

在一实施例中，所述计算各备选信息的预估点击率还包括：当所述第一关联备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第二预设值的第二关联备选信息的被推送次数是否大于等于第一阈值，当所述第二关联备选信息的被推送次数大于等于第一阈值时，将所述第二关联备选信息在预设时间内的点击率作为所述备选信息的预估点击率；当所述第二关联备选信息的被推送次数小于第一阈值时，将所述备选信息所属类别的平均点击率作为预估的点击率。In an embodiment, the calculating the estimated click rate of each candidate information further includes: determining the degree of association with the candidate information when the pushed number of the first associated candidate information is less than the first threshold Whether the number of times of the second association candidate information that reaches the second preset value is greater than or equal to the first threshold, and when the number of times the second association candidate information is pushed is greater than or equal to the first threshold, the second association is The click rate of the candidate information in the preset time is used as the estimated click rate of the candidate information; when the pushed number of the second associated candidate information is less than the first threshold, the category of the candidate information belongs to The average clickthrough rate is used as the estimated clickthrough rate.

本领域内的技术人员应明白，本发明的实施例可提供为方法、***、或计算机程序产品。因此，本发明可采用硬件实施例、软件实施例、或结合软件和硬件方面的实施例的形式。而且，本发明可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器和光学存储器等)上实施的计算机程序产品的形式。Those skilled in the art will appreciate that embodiments of the present invention can be provided as a method, system, or computer program product. Accordingly, the present invention can take the form of a hardware embodiment, a software embodiment, or a combination of software and hardware. Moreover, the invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) including computer usable program code.

本发明是参照根据本发明实施例的方法、设备(***)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。The present invention has been described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (system), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or FIG. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing device to produce a machine for the execution of instructions for execution by a processor of a computer or other programmable data processing device. The work specified in one or more blocks of a flow or a flow and/or a block diagram of a flowchart Able device.

这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。The computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device. The apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.

这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device. The instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.

以上所述，仅为本发明的较佳实施例而已，并非用于限定本发明的保护范围。 The above is only the preferred embodiment of the present invention and is not intended to limit the scope of the present invention.

Claims

一种信息处理方法，其特征在于，所述方法包括：An information processing method, characterized in that the method comprises:

接收多个备选信息，该多个备选信息由一个或多个投放终端发送；Receiving a plurality of candidate information, the plurality of candidate information being sent by one or more delivery terminals;

服务器按照所述备选信息的预估点击率离线计算出各备选信息的权重，所述离线计算是指用户所在的客户端为离线状态下服务器执行的计算；The server calculates the weight of each candidate information offline according to the estimated click rate of the candidate information, where the offline calculation refers to the calculation performed by the server that the user is located in the offline state;

从所述多个备选信息中筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息；Selecting, from the plurality of candidate information, candidate information that the weight meets the preset first rule is recorded in a pre-selected reference list, where the pre-selected reference list includes at least two candidate information;

获取所述客户端为在线状态下，操作该客户端的用户的特征信息；Obtaining feature information of the user who operates the client in the online state;

根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息；及Selecting, from the pre-selected reference list, candidate information that meets a preset second rule according to the feature information; and

将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。The candidate information in the pre-selected reference list that satisfies the preset second rule is pushed to the client.
根据权利要求1所述的方法，其特征在于，所述离线计算出各备选信息的权重为：The method according to claim 1, wherein the weight of each candidate information calculated offline is:

根据备选信息的预估点击率和备选信息的价格信息，离线计算出各备选信息的权重。The weight of each candidate information is calculated offline based on the estimated click rate of the candidate information and the price information of the candidate information.
根据权利要求1所述的方法，其特征在于，所述离线计算出各备选信息的权重之前，所述方法还包括：The method according to claim 1, wherein before the calculating the weight of each candidate information offline, the method further comprises:

计算各备选信息的预估点击率。Calculate the estimated clickthrough rate for each candidate.
根据权利要求3所述的方法，其特征在于，所述计算各备选信息的预估点击率包括：The method according to claim 3, wherein said calculating a predicted click rate of each candidate information comprises:

当所述备选信息在预设时间内未被推送时，将所述备选信息所属类别的平均点击率作为该备选信息的预估点击率。When the candidate information is not pushed within a preset time, the average click rate of the category to which the candidate information belongs is used as the estimated click rate of the candidate information.
根据权利要求3所述的方法，其特征在于，所述计算各备选信息的预估点击率包括：The method according to claim 3, wherein said calculating a predicted click rate of each candidate information comprises:

当所述备选信息在预设时间内被推送时，判断所述备选信息的被推送次数是否大于等于第一阈值；When the candidate information is pushed within a preset time, determining whether the pushed number of times of the candidate information is greater than or equal to a first threshold;

当所述备选信息的被推送次数大于等于第一阈值时，将所述备选信息在预设时间内的点击率作为该备选信息的预估点击率。When the number of times the candidate information is pushed is greater than or equal to the first threshold, the candidate information is used The click rate in the preset time is used as the estimated click rate of the candidate information.
根据权利要求5所述的方法，其特征在于，所述计算各备选信息的预估点击率还包括：The method according to claim 5, wherein said calculating the estimated click rate of each candidate information further comprises:

当所述备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第一预设值的第一关联备选信息的被推送次数是否大于等于第一阈值，当所述第一关联备选信息的被推送次数大于等于第一阈值时，将所述第一关联备选信息在预设时间内的点击率作为所述备选信息的预估点击率。When the number of times of pushing the candidate information is less than the first threshold, determining whether the number of times of pushing the first association candidate information that has reached the first preset value with the candidate information is greater than or equal to the first threshold, When the number of times the first association candidate information is pushed is greater than or equal to the first threshold, the click rate of the first association candidate information in the preset time is used as the estimated click rate of the candidate information.
根据权利要求6所述的方法，其特征在于，所述计算各备选信息的预估点击率还包括：The method according to claim 6, wherein the calculating the estimated click rate of each candidate information further comprises:

当所述第一关联备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第二预设值的第二关联备选信息的被推送次数是否大于等于第一阈值，当所述第二关联备选信息的被推送次数大于等于第一阈值时，将所述第二关联备选信息在预设时间内的点击率作为所述备选信息的预估点击率；When the number of times of pushing the first association candidate information is less than the first threshold, determining whether the number of times of pushing the second association candidate information that has reached the second preset value with the candidate information is greater than or equal to a threshold value, when the number of times the second association candidate information is pushed is greater than or equal to the first threshold, the click rate of the second association candidate information in the preset time is used as the estimated click of the candidate information. rate;

当所述第二关联备选信息的被推送次数小于第一阈值时，将所述备选信息所属类别的平均点击率作为预估的点击率。When the number of times the second association candidate information is pushed is less than the first threshold, the average click rate of the category to which the candidate information belongs is used as the estimated click rate.
根据权利要求1所述的方法，其特征在于，所述特征信息包括以下信息中的至少一种：操作所述客户端的用户的性别信息、爱好信息、地域信息。The method according to claim 1, wherein the feature information comprises at least one of the following: sex information, hobby information, and region information of a user operating the client.
一种服务器，其特征在于，所述服务器包括：A server, wherein the server comprises:

一个或多个处理器；及One or more processors; and

存储设备，用于存储指令，该指令由所述一个或多个处理器执行，以实现信息处理方法：a storage device for storing instructions executed by the one or more processors to implement an information processing method:

接收多个备选信息，该多个备选信息由一个或多个投放终端发送；Receiving a plurality of candidate information, the plurality of candidate information being sent by one or more delivery terminals;

服务器按照所述备选信息的预估点击率离线计算出各备选信息的权重，所述离线计算是指用户所在的客户端为离线状态下执行的计算；The server calculates the weight of each candidate information offline according to the estimated click rate of the candidate information, where the offline calculation refers to the calculation performed by the client where the user is located in an offline state;

筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息； Filtering out the candidate information that the weight meets the preset first rule, and records the information in a pre-selected reference list, where the pre-selected reference list includes at least two candidate information;

获取所述客户端为在线状态下，操作该客户端的用户的特征信息；Obtaining feature information of the user who operates the client in the online state;

根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息；Selecting candidate information that meets the preset second rule from the pre-selected reference list according to the feature information;

将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。The candidate information in the pre-selected reference list that satisfies the preset second rule is pushed to the client.
根据权利要求9所述的服务器，其特征在于，所述离线计算出各备选信息的权重为：The server according to claim 9, wherein the weight of each candidate information calculated offline is:

根据备选信息的预估点击率和备选信息的价格信息，离线计算出各备选信息的权重。The weight of each candidate information is calculated offline based on the estimated click rate of the candidate information and the price information of the candidate information.
根据权利要求8所述的服务器，其特征在于，所述服务器的存储设备中还包括由所述一个或多个处理器执行以实现以下步骤的指令：The server according to claim 8, wherein the storage device of the server further comprises instructions executed by the one or more processors to implement the following steps:

在所述离线计算出各备选信息的权重之前，计算各备选信息的预估点击率。The estimated click rate of each candidate information is calculated before the offline calculation of the weight of each candidate information.
根据权利要求11所述的服务器，其特征在于，所述计算各备选信息的预估点击率包括：The server according to claim 11, wherein said calculating a predicted click rate of each candidate information comprises:

当所述备选信息在预设时间内未被推送时，将所述备选信息所属类别的平均点击率作为该备选信息的预估点击率。When the candidate information is not pushed within a preset time, the average click rate of the category to which the candidate information belongs is used as the estimated click rate of the candidate information.
根据权利要求11所述的服务器，其特征在于，所述计算各备选信息的预估点击率包括：The server according to claim 11, wherein said calculating a predicted click rate of each candidate information comprises:

当所述备选信息在预设时间内被推送时，判断所述备选信息的被推送次数是否大于等于第一阈值；When the candidate information is pushed within a preset time, determining whether the pushed number of times of the candidate information is greater than or equal to a first threshold;

当所述备选信息的被推送次数大于等于第一阈值时，将所述备选信息在预设时间内的点击率作为该备选信息的预估点击率。When the pushed number of the candidate information is greater than or equal to the first threshold, the click rate of the candidate information in the preset time is used as the estimated click rate of the candidate information.
根据权利要求13所述的服务器，其特征在于，所述计算各备选信息的预估点击率还包括：The server according to claim 13, wherein the calculating the estimated click rate of each candidate information further comprises:

当所述备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第一预设值的第一关联备选信息的被推送次数是否大于等于第一阈值，当所述第一关联备选信息的被推送次数大于等于第一阈值时，将所述第一关联备选信息在预设时间内的点击率作为所述备选信息的预估点击率。When the number of times of pushing the candidate information is less than the first threshold, determining whether the number of times of pushing the first association candidate information that has reached the first preset value with the candidate information is greater than or equal to the first threshold, When the number of times the first association candidate information is pushed is greater than or equal to the first threshold, the click rate of the first association candidate information in the preset time is used as an estimate of the candidate information. Clickthrough rate.
根据权利要求14所述的服务器，其特征在于，所述计算各备选信息的预估点击率还包括：The server according to claim 14, wherein the calculating the estimated click rate of each candidate information further comprises:

当所述第一关联备选信息的被推送次数小于第一阈值时，判断与所述备选信息的关联度达到第二预设值的第二关联备选信息的被推送次数是否大于等于第一阈值，当所述第二关联备选信息的被推送次数大于等于第一阈值时，将所述第二关联备选信息在预设时间内的点击率作为所述备选信息的预估点击率；When the number of times of pushing the first association candidate information is less than the first threshold, determining whether the number of times of pushing the second association candidate information that has reached the second preset value with the candidate information is greater than or equal to a threshold value, when the number of times the second association candidate information is pushed is greater than or equal to the first threshold, the click rate of the second association candidate information in the preset time is used as the estimated click of the candidate information. rate;

当所述第二关联备选信息的被推送次数小于第一阈值时，将所述备选信息所属类别的平均点击率作为预估的点击率。When the number of times the second association candidate information is pushed is less than the first threshold, the average click rate of the category to which the candidate information belongs is used as the estimated click rate.
根据权利要求9所述的服务器，其特征在于，所述特征信息包括以下信息中的至少一种：操作所述客户端的用户的性别信息、爱好信息、地域信息。The server according to claim 9, wherein the feature information comprises at least one of the following: sex information, hobby information, and region information of a user who operates the client.
一种用于信息处理的计算机存储介质，其上存储有指令集，所述指令集由一个或多个处理器执行而执行以下步骤：A computer storage medium for information processing having stored thereon a set of instructions executed by one or more processors to perform the following steps:

接收多个备选信息，该多个备选信息由一个或多个投放终端发送；Receiving a plurality of candidate information, the plurality of candidate information being sent by one or more delivery terminals;

服务器按照所述备选信息的预估点击率离线计算出各备选信息的权重，所述离线计算是指用户所在的客户端为离线状态下执行的计算；The server calculates the weight of each candidate information offline according to the estimated click rate of the candidate information, where the offline calculation refers to the calculation performed by the client where the user is located in an offline state;

筛选出权重满足预设第一规则的备选信息，记录在一个预选参考列表中，该预选参考列表中包括至少两个备选信息；Filtering out the candidate information that the weight meets the preset first rule, and records the information in a pre-selected reference list, where the pre-selected reference list includes at least two candidate information;

获取所述客户端为在线状态下，操作该客户端的用户的特征信息；Obtaining feature information of the user who operates the client in the online state;

根据所述特征信息从所述预选参考列表中选取满足预设第二规则的备选信息；以及Selecting, from the pre-selected reference list, candidate information that meets a preset second rule according to the feature information;

将该预选参考列表中满足预设第二规则的备选信息推送给所述客户端。The candidate information in the pre-selected reference list that satisfies the preset second rule is pushed to the client.
根据权利要求17所述的计算机存储介质，其特征在于，所述离线计算出各备选信息的权重为：The computer storage medium according to claim 17, wherein the weight of each candidate information calculated offline is:

根据备选信息的预估点击率和备选信息的价格信息，离线计算出各备选信息的权重。 The weight of each candidate information is calculated offline based on the estimated click rate of the candidate information and the price information of the candidate information.
根据权利要求17所述的计算机存储介质，其特征在于，所述离线计算出各备选信息的权重之前，还包括：The computer storage medium according to claim 17, wherein before the calculating the weight of each candidate information offline, the method further comprises:

计算各备选信息的预估点击率。Calculate the estimated clickthrough rate for each candidate.
根据权利要求19所述的计算机存储介质，其特征在于，所述计算各备选信息的预估点击率包括：当所述备选信息在预设时间内被推送时，判断所述备选信息的被推送次数是否大于等于第一阈值；The computer storage medium according to claim 19, wherein said calculating a predicted click rate of each candidate information comprises: determining said candidate information when said candidate information is pushed within a preset time Whether the number of times pushed is greater than or equal to the first threshold;

当所述备选信息的被推送次数大于等于第一阈值时，将所述备选信息在预设时间内的点击率作为该备选信息的预估点击率。 When the pushed number of the candidate information is greater than or equal to the first threshold, the click rate of the candidate information in the preset time is used as the estimated click rate of the candidate information.