Information push method, device, electronic device, and readable storage medium

Abstract

The embodiments of the present application provide an information push method, device, electronic device and readable storage medium, which relate to the fields of artificial intelligence, cloud technology, and multimedia information technology. The method includes: obtaining the first object data corresponding to the information to be pushed; for each influencing parameter, determining the historical total delivery data corresponding to the influencing parameter, and determining the object screening threshold corresponding to the parameter value of the influencing parameter and the delivery probability corresponding to the candidate delivery object; further, based on the object screening threshold and the delivery probability corresponding to each candidate delivery object, determining the target delivery object corresponding to the parameter value of the influencing parameter, and then determining the target delivery object corresponding to the information to be pushed based on the target delivery object corresponding to each influencing parameter at each parameter value. In the embodiment of the present application, the quality of the target delivery object finally determined is guaranteed, and the delivery effect is improved.

Description

Technical Field

[0001] The present application relates to the fields of artificial intelligence, cloud technology, and multimedia information technology. Specifically, the present application relates to an information push method, device, electronic device, and readable storage medium. Background Art

[0002] As information technology promotes the development of various industries, people often need to identify target groups for information recommendation in order to achieve the purpose of information promotion. In the existing technology, the target population is usually determined by Lookalike (similar population expansion) technology. Lookalike technology uses the seed customer group as the positive sample of machine learning and the non-seed customer group as the negative sample for model training. The trained model is then used to score the candidate group and take the TopN (top N) group as the target population. However, research has found that since lookalike technology generally takes a unified threshold to output the TopN as the target population, in scenarios that require continuous delivery, if the TopN group is used as the target population for each delivery, the delivery effect will continue to decline, making it difficult to meet the actual needs of users. Summary of the Invention

[0003] The embodiments of the present application provide an information push method, device, electronic device, and readable storage medium, which can better determine the target recipients of the information push, thereby improving the delivery effect of the information push and better meeting actual needs. The technical solution is as follows:

[0004] According to one aspect of an embodiment of the present application, a method for pushing information is provided, the method comprising:

[0005] Obtaining first object data corresponding to the information to be pushed, the first object data including second object data corresponding to at least one delivery effect influencing parameter, the second object data corresponding to each influencing parameter including historical delivery data corresponding to at least two parameter values ​​of the influencing parameter and object information of at least one candidate delivery object;

[0006] For each influencing parameter, determine the historical total delivery data corresponding to the influencing parameter based on the historical delivery data corresponding to each parameter value of the influencing parameter;

[0007] For each parameter value of each influencing parameter, determine the object screening threshold corresponding to the parameter value of the influencing parameter based on the historical delivery data corresponding to the parameter value of the influencing parameter and the historical total delivery data corresponding to the influencing parameter;

[0008] For each parameter value of each influencing parameter, based on the object information of each candidate delivery object corresponding to the parameter value, the delivery probability corresponding to each candidate delivery object of the influencing parameter corresponding to the parameter value is obtained through the trained object screening model, and based on the object screening threshold corresponding to the parameter value of the influencing parameter and the delivery probability corresponding to each candidate delivery object, the target delivery object corresponding to the parameter value of the influencing parameter is determined;

[0009] Based on the target delivery object corresponding to each parameter value of each influencing parameter, the target delivery object corresponding to the information to be pushed is determined.

[0010] According to another aspect of an embodiment of the present application, there is provided an information push device, the device comprising:

[0011] a data acquisition module, configured to acquire first object data corresponding to the information to be pushed, the first object data including second object data corresponding to at least one influencing parameter, the second object data corresponding to each influencing parameter including historical delivery data corresponding to at least two parameter values ​​of the influencing parameter and object information of at least one candidate delivery object;

[0012] A total delivery data determination module is used to determine, for each influencing parameter, the historical total delivery data corresponding to the influencing parameter based on the historical delivery data corresponding to each parameter value of the influencing parameter;

[0013] A threshold determination module is configured to determine, for each parameter value of each influencing parameter, an object screening threshold corresponding to the parameter value of the influencing parameter based on historical delivery data corresponding to the parameter value of the influencing parameter and historical total delivery data corresponding to the influencing parameter;

[0014] The delivery object determination module is used to obtain, for each parameter value of each influencing parameter, the delivery probability corresponding to each candidate delivery object corresponding to the parameter value of the influencing parameter based on the object information of each candidate delivery object corresponding to the parameter value through a trained object screening model, and determine the target delivery object corresponding to the parameter value of the influencing parameter based on the object screening threshold corresponding to the parameter value of the influencing parameter and the delivery probability corresponding to each candidate delivery object; and determine the target delivery object corresponding to the push information based on the target delivery object corresponding to each parameter value of each influencing parameter.

[0015] Optionally, the historical delivery data corresponding to each parameter value of each influencing parameter includes the number of historical delivery objects and the number of historical conversion objects corresponding to the parameter value of the influencing parameter;

[0016] For each parameter value of each influencing parameter, the threshold determination module is specifically configured to:

[0017] Determine the total number of historical delivery objects corresponding to the influencing parameter based on the number of historical delivery objects corresponding to each parameter value of the influencing parameter;

[0018] Determine the total number of historical conversion objects corresponding to the influencing parameter based on the number of historical conversion objects corresponding to each parameter value of the influencing parameter;

[0019] Determine the conversion share of the influencing parameter corresponding to the parameter value based on the number of historical conversion objects corresponding to the influencing parameter and the number of historical total delivery objects corresponding to the influencing parameter;

[0020] Determine the placement occupancy rate of the influencing parameter corresponding to the parameter value based on the number of historical placement objects corresponding to the influencing parameter and the total number of historical placement objects corresponding to the influencing parameter;

[0021] An object screening threshold corresponding to the parameter value of the influencing parameter is determined according to the conversion share and the delivery share of the influencing parameter corresponding to the parameter value.

[0022] Optionally, when determining the object screening threshold corresponding to the parameter value of the influencing parameter based on the conversion share and the delivery share of the influencing parameter corresponding to the parameter value, the threshold determination module is specifically configured to:

[0023] Obtaining an initial object screening threshold corresponding to the parameter value of the influencing parameter;

[0024] Determine the difference between the conversion share and the delivery share corresponding to the value of the influencing parameter;

[0025] The initial object screening threshold is adjusted according to the difference to obtain the object screening threshold corresponding to the parameter value of the influencing parameter. If the difference is greater than the set value, the initial object screening threshold is reduced; if the difference is less than or equal to the set value, the initial object screening threshold is increased.

[0026] Optionally, when the threshold determination module adjusts the initial object screening threshold according to the difference to obtain the object screening threshold corresponding to the parameter value of the influencing parameter, it is specifically configured to:

[0027] Determine the ratio between the difference and the placement share of the influencing parameter corresponding to the parameter value;

[0028] The initial object screening threshold is adjusted based on the ratio to obtain the object screening threshold corresponding to the parameter value of the influencing parameter.

[0029] Optionally, if the difference is greater than a set value, the adjustment range of the initial object screening threshold is positively correlated with the ratio; if the ratio is less than or equal to the set value, the adjustment range of the initial object screening threshold is positively correlated with the absolute value of the ratio.

[0030] Optionally, the delivery object determination module is specifically configured to:

[0031] Determine, among the delivery probabilities corresponding to the candidate delivery objects corresponding to the parameter value of the influencing parameter, the candidate delivery object corresponding to the delivery probability greater than or equal to the object screening threshold corresponding to the parameter value of the influencing parameter as the target delivery object corresponding to the parameter value of the influencing parameter;

[0032] The delivery target determination module is specifically used to determine the target delivery target corresponding to the information to be pushed based on the target delivery target corresponding to each parameter value of each influencing parameter:

[0033] The target delivery object corresponding to each parameter value of each influencing parameter is used as the candidate delivery object corresponding to the information to be pushed;

[0034] Determine the target delivery object corresponding to the information to be pushed from the candidate delivery objects corresponding to the information to be pushed.

[0035] Optionally, the device further includes a conversion object determination module, specifically configured to:

[0036] Push the information to be pushed to each target delivery object;

[0037] Obtain feedback data from each target delivery object corresponding to the information to be pushed;

[0038] Determine the conversion objects among the target delivery objects based on the feedback data of each delivery object regarding the information to be pushed.

[0039] Optionally, the candidate delivery object corresponding to a parameter value of an influencing parameter is determined in the following manner:

[0040] Obtain the initial object set and historical delivery object set corresponding to the parameter value of the influencing parameter;

[0041] The initial object set is deduplicated based on the historical delivery object set, and each delivery object in the deduplicated initial object set is used as a candidate delivery object corresponding to the parameter value of the influencing parameter.

[0042] According to another aspect of an embodiment of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement any one of the method steps in the information push method.

[0043] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, any one of the method steps in the information push method is implemented.

[0044] According to one aspect of an embodiment of the present application, a computer program product is provided, including a computer program, which implements any one of the method steps in the information push method when executed by a processor.

[0045] The beneficial effects of the technical solution provided by the embodiments of the present application are:

[0046] In an embodiment of the present application, when determining the target delivery object, for each parameter value of each delivery effect influencing parameter, an object screening threshold can be determined based on the historical delivery data corresponding to the parameter value, and then the corresponding target delivery object is determined from the candidate delivery objects through the object screening model and the corresponding object screening threshold. At this time, the target delivery object finally obtained is based on the target delivery object corresponding to each parameter value of each influencing parameter. In this process, since the object screening threshold corresponding to each parameter value of each influencing parameter is obtained based on the historical delivery data, that is, the object screening threshold determined for each delivery is adjusted and changed based on the effect of the historical delivery, the quality of the target delivery object finally determined is guaranteed, and the delivery effect can be improved to better meet the actual needs of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments of the present application.

[0048] Figure 1 A schematic diagram of the composition of a target delivery object provided in an embodiment of the present application;

[0049] Figure 2 A schematic diagram of the architecture of an information push system provided in an embodiment of the present application;

[0050] Figure 3 A schematic diagram of a process for determining a target delivery object provided in an embodiment of the present application;

[0051] Figure 4 A flowchart of an information push method provided in an embodiment of the present application;

[0052] Figure 5 A schematic diagram of a curve provided in an embodiment of the present application;

[0053] Figure 6 Another curve diagram provided in an embodiment of the present application;

[0054] Figure 7 A schematic diagram of an application scenario of an information push method provided in an embodiment of the present application;

[0055] Figure 8 A schematic diagram of an application scenario of another information push method provided in an embodiment of the present application;

[0056] Figure 9 A schematic diagram of the structure of an information push device provided in an embodiment of the present application;

[0057] Figure 10 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0058] The following describes the embodiments of the present application in conjunction with the accompanying drawings. It should be understood that the embodiments described below in conjunction with the accompanying drawings are exemplary descriptions for explaining the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions of the embodiments of the present application.

[0059] Those skilled in the art will understand that, unless otherwise stated, the singular forms "a", "an", "said", and "the" used herein may also include plural forms. It should be further understood that the terms "including" and "comprising" used in the embodiments of the present application mean that the corresponding features can be implemented as the presented features, information, data, steps, operations, elements, and / or components, but do not exclude implementation as other features, information, data, steps, operations, elements, components, and / or combinations thereof supported by the present technical field. It should be understood that when we say that an element is "connected" or "coupled" to another element, the element can be directly connected or coupled to the other element, or it can refer to the element and the other element establishing a connection relationship through an intermediate element. In addition, the "connection" or "coupling" used here can include wireless connection or wireless coupling. The term "and / or" used here indicates at least one of the items defined by the term, for example, "A and / or B" can be implemented as "A", or as "B", or as "A and B".

[0060] In order to make the objectives, technical solutions and advantages of this application clearer, the implementation methods of this application will be further described in detail below with reference to the accompanying drawings.

[0061] The embodiment of the present application is an information push method proposed to address the problem that the target delivery objects determined in the prior art are inaccurate, resulting in poor delivery effects. Based on this method, the target delivery objects can be determined more accurately, thereby improving the delivery effect and better meeting the actual needs of users.

[0062] Optionally, the method provided in the embodiments of the present application can be implemented based on artificial intelligence technology. For example, the object screening model involved in the embodiments of the present application can be obtained based on machine learning training in artificial intelligence technology. Among them, machine learning (ML) is the study of how computers simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their own performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications are spread across all areas of artificial intelligence. Machine learning and deep learning generally include technologies such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and formulaic learning.

[0063] Optionally, the data processing involved in the embodiments of the present application can be implemented based on cloud technology. For example, cloud computing can be used for data calculations involved in the model training process and for data calculations on historical delivery data. Cloud computing is a computing model that distributes computing tasks across a resource pool consisting of a large number of computers, enabling various application systems to obtain computing power, storage space, and information services as needed. The network that provides resources is called a "cloud." From the user's perspective, the resources in the "cloud" are infinitely scalable and can be accessed at any time, used on demand, expanded at any time, and paid for on a per-use basis.

[0064] Optionally, the object data involved in the embodiments of the present application (such as historical delivery data and candidate delivery data) can be stored using cloud storage. Cloud storage is a new concept that extends and develops from the concept of cloud computing. A distributed cloud storage system (hereinafter referred to as a storage system) refers to a storage system that uses cluster applications, grid technology, and distributed storage file systems to bring together a large number of different types of storage devices (also known as storage nodes) in a network through application software or application interfaces to work together and provide external data storage and service access functions. Currently, the storage method of a storage system is to create a logical volume. When creating a logical volume, physical storage space is allocated for each logical volume. This physical storage space may be composed of disks on a storage device or several storage devices. When a client stores data on a logical volume, it stores the data on a file system. The file system divides the data into many parts, each of which is an object. The object contains not only the data but also additional information such as the data identifier (ID, ID entity). The file system writes each object to the physical storage space of the logical volume and records the storage location information of each object. Therefore, when a client requests access to data, the file system can allow the client to access the data based on the storage location information of each object.

[0065] In the optional embodiments of this application, when the above embodiments of this application are applied to specific products or technologies, the user-related data involved must obtain the user's permission or consent, and the collection, use, and processing of the relevant data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. In other words, if the embodiments of this application involve user-related data, this data must be obtained with the user's authorization and consent and in compliance with the relevant laws, regulations, and standards of the relevant countries and regions.

[0066] The information recommendation method provided by the present application can be executed by any electronic device, including but not limited to a terminal device or a server. Optionally, the information recommendation method provided by the embodiment of the present application can be implemented as an independent application or as a functional module / plug-in of an existing application. By applying the information recommendation method provided by the embodiment of the present application to the application, when it is necessary to determine the target delivery object, the server of the application can determine the target delivery object by executing the computer program corresponding to the functional module / plug-in, so that the information to be recommended can be delivered to these target delivery objects, that is, the information to be recommended can be sent to the terminal devices of these target delivery objects, so that the information to be recommended can be displayed to the target delivery objects through the terminal devices.

[0067] In the embodiments of the present application, the specific terminal device is not limited in the embodiments of the present application, and may include but is not limited to mobile phones, computers, intelligent voice interaction devices, smart home appliances, vehicle terminals, wearable electronic devices, AR / VR devices, etc. The server may be a cloud server or a physical server, and may be a single server or a server cluster.

[0068] In order to improve the delivery effect of the information to be recommended, the information push method provided in the embodiment of the present application can first determine the target delivery object corresponding to each delivery effect influencing parameter (hereinafter referred to as the influencing parameter) at each parameter value when determining the target delivery object corresponding to the information to be recommended, and then determine the target delivery object corresponding to the information to be pushed based on the target delivery object corresponding to each delivery effect influencing parameter at each parameter value. In other words, the target delivery object determined in the embodiment of the present application is not a simple TopN population as the target delivery object, but is jointly determined by the delivery objects corresponding to different parameter values ​​of different delivery effect influencing parameters, and since the delivery object corresponding to each parameter value of each delivery effect influencing parameter is obtained through the adjusted object screening threshold and object screening model, the target delivery object determined at this time is more accurate, and thus a better delivery effect can be achieved after the information is pushed. For example, Figure 1 As shown, it is assumed that for a certain delivery effect influencing parameter (expressed by LableX), the influencing parameter has multiple parameter values ​​(expressed by LableX_value1...LableX_value n Indicated), at this time, for each parameter value of each influencing parameter, the corresponding delivery object can be determined according to its corresponding historical delivery data and object screening model (through S1...S n Indicates the number of corresponding delivery targets), at this time, for the impact parameter LableX, its corresponding target delivery target N may include S1...S n Furthermore, the target delivery object can be obtained by the target delivery object corresponding to each influencing parameter. At this time, the target delivery object determined is more accurate, and better delivery effect can be achieved after information push.

[0069] The information push method provided in the embodiments of the present application can be applied to any application scenario with information push requirements, for example, including but not limited to advertising push scenarios.

[0070] To better understand and illustrate the information push method provided by the embodiments of this application, the method provided by this application is first described below in conjunction with a specific application scenario. The application scenario can be an advertising push scenario, and the advertisement is the information to be pushed in this scenario. The specific form of the advertisement is not limited in this embodiment of the application, and the content and form of the advertisement can be different in different application requirements. For the sake of convenience, the advertisement is described using a short message as an example.

[0071] Optional, Figure 2 A schematic diagram of the architecture of an information push system applicable to the information recommendation method provided in the embodiment of the present application is shown in FIG. Figure 2 As shown in , the system may include a terminal device 10 (i.e., the terminal device corresponding to the target delivery object) and an advertising server 20. The terminal device 10 and the advertising server 20 exchange data via a network. A database 30 may be configured on one side of the advertising server 20. The data in the database 30 may be stored in the storage space of the advertising server 20 itself, or in an independent storage device of the advertising server 20 (such as a cloud server). If it is an independent storage device, the advertising server 20 may communicate with the storage device to exchange data. The database 30 may store first object data corresponding to the information to be pushed, and a trained object screening model may be deployed in the advertising server 20.

[0072] Optionally, when it is necessary to push an advertisement, it is necessary to determine the target delivery objects corresponding to the advertisement, and then push the advertisement to the target delivery objects via short messages. Figure 3 .

[0073] Step S301, the advertising server 20 obtains first object data;

[0074] Specifically, the advertising server 20 can obtain the first object data corresponding to the advertisement from the database 30, and the first object data includes second object data corresponding to at least one delivery effect influencing parameter, and the second object data corresponding to each delivery effect influencing parameter includes historical delivery data corresponding to at least two parameter values ​​of the delivery effect influencing parameter and object information of at least one candidate delivery object.

[0075] Optionally, at least one candidate delivery object corresponding to each parameter value of each delivery effect influencing parameter can be obtained by deduplicating the initial object set based on the historical delivery object set. In this case, the same object can be effectively prevented from being pushed the advertisement repeatedly, thereby reducing the waste of delivery resources.

[0076] Step S302: the advertising server 20 determines the total historical delivery data corresponding to each delivery effect influencing parameter;

[0077] Optionally, for each delivery effect influencing parameter, the advertisement server 20 may determine the total historical delivery data corresponding to the delivery effect influencing parameter based on the historical delivery data corresponding to each parameter value of the delivery effect influencing parameter.

[0078] Step S303: the advertising server 20 determines an object screening threshold corresponding to each parameter value of each delivery effect influencing parameter;

[0079] Furthermore, for each parameter value of each delivery effect influencing parameter, after obtaining the historical total delivery data corresponding to the delivery effect influencing parameter, the advertising server 20 can obtain the object screening threshold corresponding to the parameter value of the delivery effect influencing parameter based on the historical delivery data corresponding to the parameter value of the delivery effect influencing parameter and the corresponding historical total delivery data. Based on the same principle, the object screening threshold corresponding to each parameter value of each delivery effect influencing parameter is obtained, which will not be further described here.

[0080] Step S304, the advertising server 20 determines the target delivery object corresponding to each delivery effect influencing parameter and each parameter value;

[0081] Specifically, for each parameter value of each delivery effect influencing parameter, the advertising server 20 can input the object information of each candidate delivery object corresponding to the parameter value into a trained object screening model to obtain the delivery probability corresponding to each candidate delivery object. Then, based on the object screening threshold corresponding to the parameter value of the delivery effect influencing parameter, the delivery probability corresponding to each candidate delivery object is screened to obtain the target delivery object corresponding to the parameter value of the delivery effect influencing parameter. Based on the same principle, the target delivery object corresponding to each parameter value of each delivery effect influencing parameter can be obtained, which will not be repeated here.

[0082] In step S305 , the advertisement server 20 determines the target delivery object corresponding to the information to be pushed (ie, the target delivery object determined in the figure) based on the target delivery object corresponding to each delivery effect influencing parameter when each parameter takes a value.

[0083] Optionally, the advertising server 20 may use the target delivery object corresponding to each delivery effect influencing parameter when each parameter takes a value as the candidate delivery object corresponding to the advertisement, and then determine the target delivery object corresponding to the advertisement from the candidate delivery objects corresponding to the advertisement.

[0084] Step S306: The advertising server 20 pushes the advertisement to the terminal devices 10 corresponding to the target delivery objects via SMS (the figure takes the terminal devices 101 and 102 corresponding to two target delivery objects as an example, i.e., sending SMS in the figure);

[0085] Optionally, in order to better optimize the delivery effect of the advertisement in actual application, the following steps S307 and S308 may be further included. Specifically:

[0086] Step S307: the terminal device 101 and the terminal device 102 collect feedback data corresponding to the advertisement from the target delivery object and return it to the advertisement server 20;

[0087] In step S308 , the advertising server 20 receives the feedback data and determines a conversion target according to the feedback data.

[0088] Optionally, after the advertisement is pushed to each target delivery object via SMS, the terminal device corresponding to each target delivery object can collect the feedback data of the target delivery object regarding the advertisement and feed it back to the advertising server 20. The advertising server 20 determines which of the target delivery objects are converted into conversion objects based on the received feedback data, and optimizes the next delivery of the advertisement based on the result, thereby improving the delivery effect.

[0089] The following describes several exemplary embodiments to illustrate the technical solutions of the embodiments of the present application and the technical effects produced by the technical solutions of the present application. It should be noted that the following embodiments can refer to, draw on, or combine with each other, and the same terms, similar features, and similar implementation steps in different embodiments will not be repeated.

[0090] Optional, such as Figure 4 As shown, an embodiment of the present application provides an information recommendation method. Optionally, the method can be executed by a server, and the method may include:

[0091] Step S401: Obtain first object data corresponding to the information to be pushed. The first object data includes second object data corresponding to at least one delivery effect influencing parameter. The second object data corresponding to each influencing parameter includes historical delivery data corresponding to at least two parameter values ​​of the influencing parameter and object information of at least one candidate delivery object.

[0092] Among them, the specific type of information to be pushed can be pre-configured according to actual needs, such as it can refer to products that need to be recommended, or it can be information used when recommending products (such as advertisements), etc., and the embodiments of the present application do not limit this. The delivery effect influencing parameter refers to the influencing factor when determining the delivery object, that is, the influencing factor of the delivery effect. For example, when the application scenario is a promotional advertisement in a game, the delivery effect influencing factors may include region, operator, and age, etc. At this time, region, operator, and age can be used as different delivery effect influencing parameters, and the different parameter values ​​of each delivery effect influencing parameter can refer to different information corresponding to the delivery effect influencing parameter. For example, when a certain delivery effect influencing parameter is age, the different parameter values ​​of the delivery effect influencing parameter can refer to different ages, such as age 10, age 20, etc. The historical delivery data of each parameter value of each delivery effect influencing parameter can refer to the information of the historical delivery object corresponding to the parameter value of the delivery effect influencing parameter within a certain time period, or it can only refer to the information of the historical object corresponding to the parameter value of the delivery effect influencing parameter at the time of the last delivery, and the embodiments of the present application do not limit this.

[0093] Step S402 : For each influencing parameter, determine the historical total delivery data corresponding to the influencing parameter based on the historical delivery data corresponding to each parameter value of the influencing parameter.

[0094] Step S403 : for each parameter value of each influencing parameter, determine the object screening threshold corresponding to the parameter value of the influencing parameter according to the historical delivery data corresponding to the parameter value of the influencing parameter and the historical total delivery data corresponding to the influencing parameter.

[0095] Optionally, each parameter value of each delivery effect influencing parameter may have a corresponding object screening threshold, and the delivery objects may be screened based on the object screening threshold to obtain the final target delivery object, and the object screening threshold corresponding to each parameter value of each delivery effect influencing parameter may be determined based on the historical delivery data corresponding to the parameter value of the delivery effect influencing parameter and the historical total delivery data corresponding to the delivery effect influencing parameter, and the historical total delivery data is determined based on the historical delivery data corresponding to each parameter value of the delivery effect influencing parameter.

[0096] For example, assuming that a certain delivery effect influencing parameter includes three different parameter values ​​(respectively a, b, and c), and the historical delivery data corresponding to the three different parameter values ​​are the historical delivery object number A, the historical delivery object number B, and the historical delivery object number C, respectively. At this time, the historical total delivery object number corresponding to the delivery effect influencing parameter is A+B+C. Furthermore, for each parameter value of each delivery effect influencing parameter, such as parameter value a, the object screening threshold corresponding to the parameter value a can be obtained based on the historical object number A corresponding to the parameter value a of the delivery effect influencing parameter and the historical total delivery object number A+B+C corresponding to the delivery effect influencing parameter. Furthermore, the object screening threshold corresponding to the parameter value c and the parameter value b can be obtained based on the same principle, which will not be repeated here.

[0097] Step S404: For each parameter value of each influencing parameter, based on the object information of each candidate delivery object corresponding to the parameter value, the delivery probability corresponding to each candidate delivery object of the influencing parameter corresponding to the parameter value is obtained through the trained object screening model, and based on the object screening threshold corresponding to the parameter value of the influencing parameter and the delivery probability corresponding to each candidate delivery object, the target delivery object of the influencing parameter corresponding to the parameter value is determined.

[0098] The object screening model may be pre-trained and is used to determine a delivery probability corresponding to input object information of a candidate delivery object. The delivery probability represents the likelihood that the candidate delivery object can be used as a target delivery object. The object information of the candidate delivery object may include relevant data of the candidate delivery object, and the relevant data of each candidate delivery object represents the candidate delivery object.

[0099] Correspondingly, for each parameter value of each delivery effect influencing parameter, when determining the target delivery object corresponding to the parameter value of the delivery effect influencing parameter, the relevant data of the corresponding candidate delivery objects can be input into the object screening model, and the object screening model can obtain the delivery probability corresponding to each candidate delivery object. The delivery probability represents the possibility of the candidate delivery object as the target delivery object. When the delivery probability of the candidate delivery object is greater, the possibility of the candidate delivery object as the target delivery object is greater. Conversely, when the delivery probability of the candidate delivery object is smaller, the possibility of the candidate delivery object as the target delivery object is smaller.

[0100] Step S405 : determining the target delivery object corresponding to the information to be pushed based on the target delivery object corresponding to each parameter value of each influencing parameter.

[0101] Optionally, based on an object screening threshold corresponding to the parameter value of the influencing parameter and a delivery probability corresponding to each candidate delivery object, determining a target delivery object corresponding to the parameter value of the influencing parameter includes:

[0102] The candidate delivery object corresponding to the delivery probability corresponding to the parameter value of the influencing parameter, which is greater than or equal to the object screening threshold corresponding to the parameter value of the influencing parameter, is determined as the target delivery object corresponding to the parameter value of the influencing parameter.

[0103] Determining the target delivery object corresponding to the information to be pushed based on the target delivery object corresponding to each delivery effect influencing parameter when each parameter takes a value includes at least one of the following:

[0104] Determining the target delivery object corresponding to the information to be pushed based on the target delivery object corresponding to each parameter value of each influencing parameter includes:

[0105] The target delivery object corresponding to each parameter value of each influencing parameter is used as the candidate delivery object corresponding to the information to be pushed;

[0106] Determine the target delivery object corresponding to the information to be pushed from the candidate delivery objects corresponding to the information to be pushed.

[0107] Optionally, the delivery probability of each candidate delivery object can be compared with a determined object screening threshold to screen out delivery probabilities greater than the object screening threshold, and the candidate delivery objects corresponding to delivery probabilities greater than or equal to the object screening threshold can be determined as the target delivery objects whose delivery effect influencing parameter corresponds to the parameter value.

[0108] Furthermore, after obtaining the target delivery objects corresponding to each delivery effect influencing parameter when each parameter is taken as a value, the target delivery objects corresponding to each delivery effect influencing parameter when each parameter is taken as a value can be used as a candidate delivery object corresponding to the information to be pushed, and then the target delivery object is determined from the candidate delivery objects. For example, if there are sufficient delivery resources, the target delivery objects corresponding to each delivery effect influencing parameter when each parameter is taken as a value can be used as the target delivery object; or if the number of target delivery objects is fixed, a set number of target objects can be selected from the target delivery objects corresponding to each parameter of each delivery effect influencing parameter when each parameter is taken as a value, as the target delivery objects corresponding to the information to be pushed, etc. The embodiments of the present application do not limit this.

[0109] In an embodiment of the present application, when determining the target delivery object, for each parameter value of each delivery effect influencing parameter, an object screening threshold can be determined based on the historical delivery data corresponding to the parameter value, and then the corresponding target delivery object is determined from the candidate delivery objects through the object screening model and the corresponding object screening threshold, and the target delivery object finally obtained is based on the target delivery object corresponding to each delivery effect influencing parameter at each parameter value. In this process, since the object screening threshold corresponding to each parameter value of each delivery effect influencing parameter is obtained based on the historical delivery data, that is, the object screening threshold determined for each delivery is adjusted and changed based on the effect of the historical delivery, the quality of the target delivery object finally determined is guaranteed, and the delivery effect can be improved to better meet the actual needs of users.

[0110] In an optional implementation manner in the embodiment of the present application, the historical delivery data corresponding to each parameter value of each influencing parameter includes the number of historical delivery objects and the number of historical conversion objects corresponding to the parameter value of the influencing parameter;

[0111] For each parameter value of each influencing parameter, determining an object screening threshold corresponding to the parameter value of the influencing parameter based on the historical delivery data corresponding to the parameter value of the influencing parameter and the historical total delivery data corresponding to the influencing parameter includes:

[0112] Determine the total number of historical delivery objects corresponding to the influencing parameter based on the number of historical delivery objects corresponding to each parameter value of the influencing parameter;

[0113] Determine the total number of historical conversion objects corresponding to the influencing parameter based on the number of historical conversion objects corresponding to each parameter value of the influencing parameter;

[0114] Determine the conversion share of the influencing parameter corresponding to the parameter value based on the number of historical conversion objects corresponding to the influencing parameter and the number of historical total delivery objects corresponding to the influencing parameter;

[0115] Determine the placement occupancy rate of the influencing parameter corresponding to the parameter value based on the number of historical placement objects corresponding to the influencing parameter and the total number of historical placement objects corresponding to the influencing parameter;

[0116] An object screening threshold corresponding to the parameter value of the influencing parameter is determined according to the conversion share and the delivery share of the influencing parameter corresponding to the parameter value.

[0117] Optionally, the historical delivery data corresponding to each parameter value of each delivery effect influencing parameter may include the number of historical delivery objects and the number of historical conversion objects corresponding to the parameter value of the delivery effect influencing parameter. The historical conversion objects refer to objects in the historical delivery objects that are interested in the historical push information. For example, if the historical push information is about a particular game, the historical conversion objects may refer to objects that registered the game in the historical delivery objects corresponding to the game.

[0118] Furthermore, for each parameter value of each delivery effect influencing parameter, after knowing the historical number of delivery objects corresponding to each parameter value of the delivery effect influencing parameter, the historical total number of delivery objects corresponding to the delivery effect influencing parameter can be determined based on the historical number of delivery objects corresponding to each parameter value of the delivery effect influencing parameter, and then the ratio of the historical number of delivery objects corresponding to the parameter value of the delivery effect influencing parameter to the historical total number of delivery objects corresponding to the delivery effect influencing parameter can be determined to obtain the delivery occupancy rate of the delivery effect influencing parameter corresponding to the parameter value.

[0119] For example, suppose that for the tth delivery, there are n values ​​of the parameters that affect the delivery effect, s i The number of historical delivery objects corresponding to the parameter value i of the delivery effect impact parameter. At this time, the total number of historical delivery objects corresponding to the delivery effect impact parameter can be expressed as The distribution of the number of historical delivery objects corresponding to the values ​​of the parameters affecting the delivery effect can be expressed as Furthermore, for parameter value i, based on the number of historical delivery objects corresponding to the delivery effect impact parameter when the parameter value is i, and the total number of historical delivery objects corresponding to the delivery effect impact parameter, the delivery share of the delivery effect impact parameter corresponding to the parameter value can be expressed as

[0120] Correspondingly, for the parameter value of the delivery effect influencing parameter, after knowing the number of historical conversion objects corresponding to each parameter value of the delivery effect influencing parameter, the total historical number of conversion objects corresponding to the delivery effect influencing parameter can be determined based on the number of historical conversion objects corresponding to each parameter value of the delivery effect influencing parameter. Then, the ratio of the number of historical conversion objects corresponding to the parameter value of the delivery effect influencing parameter to the total historical number of conversion objects corresponding to the delivery effect influencing parameter can be determined to obtain the conversion share of the delivery effect influencing parameter corresponding to the parameter value.

[0121] For example, suppose that for the tth delivery, there are n values ​​of the parameters that affect the delivery effect, r iIt indicates the number of historical conversion objects corresponding to the impact parameter of the delivery effect when the parameter value is i. The total number of historical conversion objects corresponding to the impact parameter of the delivery effect can be expressed as The distribution of the number of historical conversion objects corresponding to the values ​​of the parameters affecting the delivery effect can be expressed as Furthermore, for parameter value i, based on the number of historical conversion objects corresponding to the delivery effect impact parameter when the parameter value is i, and the total number of historical conversion objects corresponding to the delivery effect impact parameter, the conversion share of the delivery effect impact parameter corresponding to the parameter value can be expressed as

[0122] An optional implementation in the embodiment of the present application is to determine the object screening threshold corresponding to the parameter value of the influencing parameter based on the conversion share and the delivery share of the influencing parameter corresponding to the parameter value, including:

[0123] Obtaining an initial object screening threshold corresponding to the parameter value of the influencing parameter;

[0124] Determine the difference between the conversion share and the delivery share corresponding to the value of the influencing parameter;

[0125] The initial object screening threshold is adjusted according to the difference to obtain the object screening threshold corresponding to the parameter value of the influencing parameter. If the difference is greater than the set value, the initial object screening threshold is reduced; if the difference is less than or equal to the set value, the initial object screening threshold is increased.

[0126] Optionally, each parameter value of each delivery effect influencing parameter may have a corresponding initial object screening threshold value. The initial object screening threshold value may be the object screening threshold value corresponding to the parameter value of the delivery effect influencing parameter when the target delivery object was determined last time, or it may be a preset default object screening threshold value for the parameter value of the delivery effect influencing parameter. This embodiment of the present application does not limit this. During the first delivery, each parameter value of each delivery effect influencing parameter may correspond to the same initial object screening threshold value, or different initial object screening threshold values ​​may be set for each parameter. This embodiment of the present application does not limit this.

[0127] In practical applications, assuming that the initial object screening threshold is Score 0 , and when the target delivery object is determined for the first time, when the delivery probability of the candidate object is greater than the initial object screening threshold (ie Score>=Score 0 ) is used as the delivery target. At this time, based on Score>=Score 0 The number of delivery objects obtained is (D+X)*N, where X is the relaxation coefficient (specifically a hyperparameter), D is the number of delivery days, and N is the expected number of delivery objects.

[0128] Accordingly, after obtaining the conversion share and delivery share of the delivery effect influencing parameter corresponding to the parameter value, the difference between the conversion share and the delivery share can be determined. If the determined difference is greater than the set value, it means that the delivery object corresponding to the parameter value of the delivery effect influencing parameter is more likely to become a conversion object, that is, the degree of reception of the push information is higher. At this time, the initial object screening threshold can be adjusted down, thereby increasing the number of target delivery objects corresponding to the parameter value of the delivery effect influencing parameter. On the contrary, if the determined ratio is or is equal to the set value, it means that the delivery object corresponding to the parameter value of the delivery effect influencing parameter is not likely to become a conversion object, that is, the degree of reception of the push information is low. At this time, the initial object screening threshold can be increased, thereby reducing the number of target delivery objects corresponding to the parameter value of the delivery effect influencing parameter, thereby reducing the corresponding delivery resources and saving delivery costs. Optionally, the set value can be zero or a relatively small absolute value, which can be set according to experience or experimental values. Among them, when the set value is zero, if the determined difference is equal to zero, it means that the delivery object corresponding to the parameter value of the delivery effect influencing parameter has a moderate degree of reception of push information. Here, the initial object screening threshold can be left unchanged and the original number of delivery objects can be maintained.

[0129] In an optional implementation of the embodiment of the present application, the initial object screening threshold is adjusted according to the difference to obtain the object screening threshold corresponding to the parameter value of the influencing parameter, including:

[0130] Determine the ratio between the difference and the placement share of the influencing parameter corresponding to the parameter value;

[0131] The initial object screening threshold is adjusted based on the ratio to obtain the object screening threshold corresponding to the parameter value of the influencing parameter.

[0132] Optionally, when adjusting the initial screening threshold according to the difference, the ratio between the difference and the delivery occupancy rate of the delivery effect influencing parameter corresponding to the parameter value can be further determined, and then the initial object screening threshold is adjusted based on the ratio. The ratio represents the distribution difference rate between the conversion object and the delivery object when the delivery effect influencing parameter corresponds to the parameter value. Optionally, for any delivery effect influencing parameter, the distribution difference rate between the conversion object and the delivery object when the delivery effect influencing parameter corresponds to each parameter value can be expressed as Among them, w1,w2,…,w n Indicates the distribution difference rate between conversion objects and delivery objects corresponding to parameter values ​​1 to n.

[0133] Optionally, the distribution difference rate between the conversion object and the delivery object when each delivery effect influencing parameter corresponds to each parameter value can be determined based on the following formula:

[0134]

[0135] Among them, w i Indicates the distribution difference rate between the conversion object and the delivery object when the delivery effect impact parameter corresponds to the parameter value i. Indicates that the effect of the delivery affects the conversion rate of the parameter when the parameter takes the value i. Indicates the delivery effect impact parameter's delivery occupancy rate when the parameter takes value i.

[0136] In an optional implementation of the embodiment of the present application, if the difference is greater than the set value, the adjustment amplitude of the initial object screening threshold is positively correlated with the ratio; if the ratio is less than or equal to the set value, the adjustment amplitude of the initial object screening threshold is positively correlated with the absolute value of the ratio.

[0137] Optionally, when the determined difference is greater than a set value, the adjustment range of the initial screening threshold is positively correlated with the ratio, i.e., when the ratio is greater than the set value and the value is larger, the initial object screening threshold is reduced by a larger amount. Conversely, when the ratio is less than or equal to the set value, the adjustment range of the initial screening threshold is positively correlated with the absolute value of the ratio, i.e., when the ratio is less than or equal to the set value and the value is larger, the initial object screening threshold is increased by a smaller amount.

[0138] In this embodiment of the present application, the object screening threshold can be determined based on the following formula.

[0139] Score t+1 =Score t -lr*f(w i )

[0140] Among them, Score t+1 Indicates the object screening threshold corresponding to a parameter value of a certain delivery effect influencing parameter when the target delivery object is determined t+1 times, Score t The initial object screening threshold corresponding to the parameter value of the delivery effect parameter is represented. lr is the base adjustment score (a hyperparameter). The specific parameter value of the base adjustment score can be set according to the actual situation. Usually, the parameter value of the base adjustment score is 0.05 or 0.1, etc. i ) is the adjustment coefficient, and its parameter range is [0,1]. Calculate the f(w i ), there are two types of strategies, namely linear strategy and nonlinear strategy. The calculation formula of linear strategy is as follows:

[0141]

[0142] Among them, α and β are hyperparameters, and the specific α and β can be set based on the actual situation. At this time, f(w i ) and w i The relationship between Figure 5 shown.

[0143] The optional non-linear strategy is calculated as follows:

[0144]

[0145] Among them, f(w i ) and w i The relationship between Figure 6 As shown. Correspondingly, at this time based on Figure 5 and Figure 6 As can be seen from the graph shown, f(w i ) parameter value range is [0,1], and in w i When greater than zero, the amplitude of the initial object screening threshold is reduced and w i Positive correlation; in w i When it is less than zero, the magnitude of the increase in the initial object screening threshold is positively correlated with the absolute value of the ratio.

[0146] In an optional implementation manner in the embodiment of the present application, the method further includes:

[0147] Push the information to be pushed to each target delivery object;

[0148] Obtain feedback data from each target delivery object corresponding to the information to be pushed;

[0149] Determine the conversion objects among the target delivery objects based on the feedback data of each delivery object regarding the information to be pushed.

[0150] Optionally, after determining the target delivery objects, the push information can be pushed to each target delivery object, such as by sending text messages to send the information to be pushed to each target delivery object, or by using in-application information push to send the information to be pushed to each target delivery object. The embodiment of the present application does not limit the method of pushing the recommended information to each target delivery object.

[0151] Furthermore, after receiving the information to be pushed, each target recipient may generate feedback regarding the information to be pushed. This feedback represents the target recipient's level of interest in the information to be pushed. For example, if the information to be pushed is game information, the target recipient may download and register for the game if they are interested. Accordingly, feedback data generated by each target recipient regarding the information to be pushed can be obtained. Since this feedback data reflects the target recipient's level of interest in the information to be pushed, the conversion targets among the target recipients can be determined based on the feedback data obtained from each recipient regarding the information to be pushed.

[0152] Optionally, based on the feedback data obtained from each delivery target regarding the pushed information, conversion targets within the target delivery target are determined. One optional implementation method is that if the feedback data generated by the target delivery target regarding the information to be pushed indicates that the target delivery target has a high interest in the information to be pushed, then the target delivery target can be considered a conversion target. For example, if information about a game to be pushed is sent to the target delivery target via SMS, and the target delivery target generates behavioral data such as downloading and registering the game after receiving the SMS, this indicates that the target delivery target has a high interest in the information to be pushed, and the target delivery target is considered a conversion target.

[0153] In an optional implementation manner of the embodiment of the present application, the candidate delivery object corresponding to a parameter value of an influencing parameter is determined in the following manner:

[0154] Obtain the initial object set and historical delivery object set corresponding to the parameter value of the influencing parameter;

[0155] The initial object set is deduplicated based on the historical delivery object set, and each delivery object in the deduplicated initial object set is used as a candidate delivery object corresponding to the parameter value of the influencing parameter.

[0156] Optionally, in an embodiment of the present application, an initial object set and a historical delivery object set can be obtained. The historical delivery object set includes various historical target delivery objects, and the initial object set includes various objects, but the various objects included in the initial object set may include some historical target delivery objects. At this time, after obtaining the initial object set and the historical delivery object set, the initial object set can be deduplicated based on the historical delivery object set, and the objects in the deduplicated initial object set can be used as at least one candidate delivery object corresponding to the parameter value of the delivery effect influencing parameter.

[0157] Optionally, in actual applications, the obtained initial object set can also be filtered based on historical delivery receipts. For example, based on historical delivery data, it can be determined which delivery effect-influencing parameters and which parameter values ​​produce poor delivery effects. At this time, the objects corresponding to the value of the delivery effect-influencing parameters in the initial object set can be filtered out, thereby reducing the amount of data processing and reducing the waste of delivery resources.

[0158] Optional, such as Figure 7 As shown, the information push method provided in the embodiment of the present application can be applied to scenarios such as continuous delivery of customer products. In this scenario, it can specifically include the client side, the algorithm side, the delivery side, and the client side. At this time, when the product is delivered, the client side can obtain relevant data of a batch of seed package users (i.e., seed package users in the figure) and various sub-package users as training samples to train the initial neural network model to obtain an object screening model; further, the client side can obtain relevant data of candidate delivery objects corresponding to each parameter value of each delivery effect influencing parameter (such as including model (with adopted object screening model), strategy (referring to specific delivery strategy), model (model of terminal device), region, operator (operator adopted by terminal device, etc.) (i.e., relevant data in the figure). At this time, the candidate delivery objects obtained include both potential audiences of the customer product and a large number of non-audience groups; further, the relevant data of the candidate delivery objects corresponding to each parameter value of each delivery effect influencing parameter are input into the object screening model on the algorithm side, and the candidate delivery objects with a delivery probability greater than the initial object screening threshold are determined as high-potential groups and a delivery package is generated.

[0159] Correspondingly, the delivery side delivers the recommended information to the target delivery object (i.e., the client side) through the SMS delivery platform based on the delivery system. After the target delivery object on the client side receives the SMS (i.e., reached), when the target delivery object clicks the SMS (i.e., click in the figure), the SMS delivery platform can collect the target delivery object's post-link data (i.e., feedback data) for the information to be recommended, such as registration, completion (the documents required for the product are fully prepared), credit (such as payment, credit), etc. Further, the obtained post-link data can be returned to the algorithm side (i.e., reach & click return data in the figure)

[0160] Furthermore, the algorithm can attribute the delivery effect of each delivery effect influencing parameter based on the post-link data, determine the conversion population, and thus analyze the delivery effect influencing parameters with positive effects and delivery effect influencing parameters with negative effects, and optimize the next delivery strategy based on the obtained analysis results, so as to achieve the goal of optimizing the delivery effect.

[0161] Optionally, in order to better understand the method provided in the embodiments of the present application, Figure 8 As shown, the following describes this method in detail in conjunction with scenarios such as the continuous launch of customer products.

[0162] In actual applications, if this product launch is the first, historical conversion populations (i.e., historical conversion objects) can be obtained and the corresponding identifiers (i.e., ID mapping) of each conversion population can be recorded as positive samples, and populations randomly sampled from the general population (such as the population in the database) can be used as negative samples. Furthermore, the initial neural network model can be trained based on the obtained positive and negative samples to form a training set, a validation set, and a test set. During the training process, the relevant data corresponding to each sample (i.e., the relevant data of the user in the figure) can be obtained and features can be extracted. In order to ensure that the extracted features more accurately represent the sample objects, the extracted features can be preprocessed to obtain features corresponding to each sample object. The initial neural network model is then trained based on the features corresponding to each sample object to obtain an initial object screening model. The obtained initial object screening model is then evaluated offline for effectiveness, and the final object screening model is obtained after meeting the requirements.

[0163] Furthermore, when launching a customer product, relevant data of users in the general population can be obtained, and the population corresponding to each parameter value of each launch effect influencing parameter can be determined, and the obtained population can be filtered (i.e., the target population can be circled). For example, the population corresponding to a certain parameter value of a certain launch effect influencing parameter can be filtered out, and the population obtained after filtering can be used as a candidate population, which includes candidate populations corresponding to different parameter values ​​of each launch effect influencing parameter.

[0164] Furthermore, the candidate population is input into the object screening model, which uses a classification algorithm to obtain the delivery probability of each candidate (i.e., the model prediction score in the figure), and filters the candidate objects according to the initial object screening threshold (i.e., population filtering in the figure), outputting the population with a delivery probability higher than the initial object screening threshold (i.e., the TopN similar population in the figure) and recording the corresponding identifiers (i.e., ID mapping). Furthermore, the TopN similar population can be filtered for delivery (e.g., filtering out the already delivered population (i.e., the excluded population in the figure)) to obtain the final delivery population. The final delivery population is then targeted for customer product delivery (i.e., unpacking and targeted delivery in the figure), and feedback data on the customer product from the delivery population is collected to evaluate the delivery effect and determine the conversion population within the delivery population.

[0165] Furthermore, statistics are taken on the target population to obtain the target population distribution corresponding to each parameter value of each target effect influencing parameter (i.e., label LableX distribution 1 in the figure), and statistics are taken on the conversion population to obtain the conversion population distribution corresponding to each parameter value of each target effect influencing parameter (i.e., label LableX distribution 2 in the figure). Then, the difference between label LableX distribution 1 and LableX distribution 2 is determined (i.e., the label distribution difference in the figure), and the initial object screening threshold is adjusted based on the difference (i.e., the threshold adjustment in the figure) to obtain the object screening threshold corresponding to each parameter value of each target effect influencing parameter. Furthermore, when the customer's product is launched next time, for each parameter value of each target effect influencing parameter, the target population corresponding to the parameter value of the target effect influencing parameter is determined based on the object screening threshold and object screening model corresponding to the parameter value of the target effect influencing parameter. At this time, the target population corresponding to each parameter value of all target effect influencing parameters can constitute the final target population (i.e., the target delivery object).

[0166] In order to better understand the process of adjusting the initial object screening threshold according to the difference between label LableX distribution 1 and label LableX distribution 2, the following is an example of the parameter X (i.e., label X) that affects the delivery effect. The delivery data at label X is shown in Table 1.

[0167] Table 1

[0168]

[0169] From Table 1, we can see that the label X has n parameter values ​​(i.e., LableX_value1…LableX_valuen), and each parameter value has a corresponding number of people (s i ), share of investment (srate i ), number of conversions (r i ), conversion share (srate i ), proportion difference (w i ) and the ratio between the difference in proportion and the share of delivery (f(w i )). Furthermore, if the determined f(w i ) is greater than 0, then reduce the initial object threshold. If the determined f(w i ) is less than 0, then the initial object screening threshold is increased, and when f(w i ) is equal to 0, the initial object threshold is not adjusted.

[0170] In one example, assuming that label X is a model, different models (brand A to brand D) have different parameter values ​​for label X. The corresponding delivery data after one delivery is shown in Table 2:

[0171] Table 2

[0172]

[0173] At this time, based on the f(w i ) we can see that the f(w i ) is greater than zero, indicating a positive advertising effect, but the corresponding advertising population ratio is low. Brands A and B, on the other hand, not only have negative advertising effects but also have a high advertising population ratio. In this case, the target screening threshold for Brands F, C, and D, which have positive effects, can be reduced. Meanwhile, the target screening threshold for Brands A and B, which have negative effects, can be increased, while the target screening threshold for Brand E remains unchanged, thereby reducing conversion costs.

[0174] In another example, assuming that label X is a region, different regions (region 1 to region 4) have different parameter values ​​for label X. The corresponding delivery data after one delivery is shown in Table 3:

[0175] Table 3

[0176]

[0177] At this time, based on the f(w i ) we can see that f(w i ) is greater than zero, indicating a positive delivery effect, but the corresponding delivery population ratio is low. Regions 1 and 3, on the other hand, not only have negative delivery effects but also have a high delivery population ratio. In this case, the target screening threshold for Regions 2 and 4 (with positive effects) can be lowered, while the target screening threshold for Regions 1 and 3 (with negative effects) can be increased, thereby reducing conversion costs.

[0178] In the embodiment of the present application, the difference between the delivery population and the conversion population is studied, and the influence of each delivery effect parameter on f(w i ) The parameter values ​​with positive and negative effects are taken and the corresponding object screening thresholds are adjusted, thereby increasing the proportion of the delivery population corresponding to the positive label in the total delivery population and reducing the proportion of the delivery population corresponding to the negative label in the total delivery population, thereby improving the quality of the overall delivery population.

[0179] The embodiment of the present application provides an information push device, such as Figure 9 As shown, the information push device 90 may include: a data acquisition module 901, a total delivery data determination module 902, a threshold determination module 903 and a delivery target determination module 904, wherein:

[0180] Data acquisition module 901 is configured to acquire first object data corresponding to the information to be pushed, the first object data including second object data corresponding to at least one influencing parameter, the second object data corresponding to each influencing parameter including historical delivery data corresponding to at least two parameter values ​​of the influencing parameter and object information of at least one candidate delivery object;

[0181] Total delivery data determination module 902, for determining, for each influencing parameter, the historical total delivery data corresponding to the influencing parameter based on the historical delivery data corresponding to each parameter value of the influencing parameter;

[0182] A threshold determination module 903 is configured to determine, for each parameter value of each influencing parameter, an object screening threshold corresponding to the parameter value of the influencing parameter based on the historical delivery data corresponding to the parameter value and the historical total delivery data corresponding to the influencing parameter;

[0183] The delivery object determination module 904 is used to obtain, for each parameter value of each influencing parameter, the delivery probability corresponding to each candidate delivery object corresponding to the parameter value of the influencing parameter based on the object information of each candidate delivery object corresponding to the parameter value through a trained object screening model, and determine the target delivery object corresponding to the parameter value of the influencing parameter based on the object screening threshold corresponding to the parameter value of the influencing parameter and the delivery probability corresponding to each candidate delivery object; and determine the target delivery object corresponding to the push information based on the target delivery object corresponding to each parameter value of each influencing parameter.

[0184] Optionally, the historical delivery data corresponding to each parameter value of each influencing parameter includes the number of historical delivery objects and the number of historical conversion objects corresponding to the parameter value of the influencing parameter;

[0185] For each parameter value of each influencing parameter, the threshold determination module is specifically configured to:

[0186] Determine the total number of historical delivery objects corresponding to the influencing parameter based on the number of historical delivery objects corresponding to each parameter value of the influencing parameter;

[0187] Determine the total number of historical conversion objects corresponding to the influencing parameter based on the number of historical conversion objects corresponding to each parameter value of the influencing parameter;

[0188] Determine the conversion share of the influencing parameter corresponding to the parameter value based on the number of historical conversion objects corresponding to the influencing parameter and the number of historical total delivery objects corresponding to the influencing parameter;

[0189] Determine the placement occupancy rate of the influencing parameter corresponding to the parameter value based on the number of historical placement objects corresponding to the influencing parameter and the total number of historical placement objects corresponding to the influencing parameter;

[0190] An object screening threshold corresponding to the parameter value of the influencing parameter is determined according to the conversion share and the delivery share of the influencing parameter corresponding to the parameter value.

[0191] Optionally, when determining the object screening threshold corresponding to the parameter value of the influencing parameter based on the conversion share and the delivery share of the influencing parameter corresponding to the parameter value, the threshold determination module is specifically configured to:

[0192] Obtaining an initial object screening threshold corresponding to the parameter value of the influencing parameter;

[0193] Determine the difference between the conversion share and the delivery share corresponding to the value of the influencing parameter;

[0194] The initial object screening threshold is adjusted according to the difference to obtain the object screening threshold corresponding to the parameter value of the influencing parameter. If the difference is greater than the set value, the initial object screening threshold is reduced; if the difference is less than or equal to the set value, the initial object screening threshold is increased.

[0195] Optionally, when the threshold determination module adjusts the initial object screening threshold according to the difference to obtain the object screening threshold corresponding to the parameter value of the influencing parameter, it is specifically configured to:

[0196] Determine the ratio between the difference and the placement share of the influencing parameter corresponding to the parameter value;

[0197] The initial object screening threshold is adjusted based on the ratio to obtain the object screening threshold corresponding to the parameter value of the influencing parameter.

[0198] Optionally, if the difference is greater than a set value, the adjustment range of the initial object screening threshold is positively correlated with the ratio; if the ratio is less than or equal to the set value, the adjustment range of the initial object screening threshold is positively correlated with the absolute value of the ratio.

[0199] Optionally, the delivery object determination module is specifically configured to:

[0200] Determine, among the delivery probabilities corresponding to the candidate delivery objects corresponding to the parameter value of the influencing parameter, the candidate delivery object corresponding to the delivery probability greater than or equal to the object screening threshold corresponding to the parameter value of the influencing parameter as the target delivery object corresponding to the parameter value of the influencing parameter;

[0201] The delivery target determination module is specifically used to determine the target delivery target corresponding to the information to be pushed based on the target delivery target corresponding to each parameter value of each influencing parameter:

[0202] The target delivery object corresponding to each parameter value of each influencing parameter is used as the candidate delivery object corresponding to the information to be pushed;

[0203] Determine the target delivery object corresponding to the information to be pushed from the candidate delivery objects corresponding to the information to be pushed.

[0204] Optionally, the device further includes a conversion object determination module, specifically configured to:

[0205] Push the information to be pushed to each target delivery object;

[0206] Obtain feedback data from each target delivery object corresponding to the information to be pushed;

[0207] Determine the conversion objects among the target delivery objects based on the feedback data of each delivery object regarding the information to be pushed.

[0208] Optionally, the candidate delivery object corresponding to a parameter value of an influencing parameter is determined in the following manner:

[0209] Obtain the initial object set and historical delivery object set corresponding to the parameter value of the influencing parameter;

[0210] The initial object set is deduplicated based on the historical delivery object set, and each delivery object in the deduplicated initial object set is used as a candidate delivery object corresponding to the parameter value of the influencing parameter.

[0211] The device of the embodiment of the present application can execute the method provided by the embodiment of the present application, and its implementation principle is similar. The actions performed by each module in the device of each embodiment of the present application correspond to the steps in the method of each embodiment of the present application. For the detailed functional description of each module of the device, please refer to the description in the corresponding method shown in the previous text, and will not be repeated here.

[0212] An embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory, and the processor executes the above computer program to implement the steps of the information push method.

[0213] In an alternative embodiment, an electronic device is provided, such as Figure 10 As shown, Figure 10 The electronic device 4000 shown includes: a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, for example, via a bus 4002. Optionally, the electronic device 4000 may further include a transceiver 4004, which may be used for data exchange between the electronic device and other electronic devices, such as data transmission and / or data reception. It should be noted that in actual applications, the number of transceivers 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present application.

[0214] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 4001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0215] Bus 4002 may include a path for transmitting information between the above components. Bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. Bus 4002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 10 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0216] The memory 4003 can be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium that can be used to carry or store computer programs and can be read by a computer, without limitation here.

[0217] The memory 4003 is used to store the computer program for executing the embodiment of the present application, and the execution is controlled by the processor 4001. The processor 4001 is used to execute the computer program stored in the memory 4003 to implement the steps shown in the above method embodiment.

[0218] An embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps and corresponding contents of the aforementioned method embodiment can be implemented.

[0219] An embodiment of the present application also provides a computer program product, including a computer program, which can implement the steps and corresponding contents of the aforementioned method embodiment when executed by a processor.

[0220] The terms "first," "second," "third," "fourth," "1," "2," and the like (if any) in the specification and claims of this application and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or sequential sequence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of the application described herein can be implemented in an order other than that shown or described in the drawings.

[0221] It should be understood that, although each operation step is indicated by arrows in the flowchart of the embodiment of the present application, the order of implementation of these steps is not limited to the order indicated by the arrows. Unless otherwise clearly stated herein, in some implementation scenarios of the embodiment of the present application, the implementation steps in each flowchart can be performed in other orders according to demand. In addition, some or all of the steps in each flowchart can include multiple sub-steps or multiple stages based on actual implementation scenarios. Some or all of these sub-steps or stages can be executed at the same time, and each sub-step or stage in these sub-steps or stages can also be executed at different times respectively. Under different scenarios at the execution time, the execution order of these sub-steps or stages can be flexibly configured according to demand, and the embodiment of the present application does not limit this.

[0222] The above description is only an optional implementation method for some implementation scenarios of this application. It should be pointed out that for ordinary technicians in this technical field, without departing from the technical concept of the solution of this application, the use of other similar implementation methods based on the technical ideas of this application also falls within the protection scope of the embodiments of this application.

Claims

1. An information push method, characterized in that: include: Obtaining first object data corresponding to the information to be pushed, the first object data including second object data corresponding to at least one delivery effect influencing parameter, the second object data corresponding to each of the influencing parameters including historical delivery data corresponding to at least two parameter values ​​of the influencing parameter and object information of at least one candidate delivery object; For each of the influencing parameters, determine the historical total delivery data corresponding to the influencing parameter based on the historical delivery data corresponding to each parameter value of the influencing parameter; For each parameter value of each of the influencing parameters, determining an object screening threshold corresponding to the parameter value of the influencing parameter based on the historical delivery data corresponding to the parameter value of the influencing parameter and the historical total delivery data corresponding to the influencing parameter; For each parameter value of each of the influencing parameters, based on the object information of each candidate delivery object corresponding to the parameter value, the delivery probability corresponding to each candidate delivery object of the influencing parameter corresponding to the parameter value is obtained through the trained object screening model, and based on the object screening threshold corresponding to the parameter value of the influencing parameter and the delivery probability corresponding to each candidate delivery object, the target delivery object corresponding to the parameter value of the influencing parameter is determined; The target delivery object corresponding to the information to be pushed is determined based on the target delivery object corresponding to each parameter value of each of the influencing parameters.

2. The method according to claim 1, characterized in that The historical delivery data corresponding to each parameter value of each of the influencing parameters includes the number of historical delivery objects and the number of historical conversion objects corresponding to the parameter value of the influencing parameter; For each parameter value of each of the influencing parameters, determining the object screening threshold corresponding to the parameter value of the influencing parameter based on the historical delivery data corresponding to the parameter value of the influencing parameter and the historical total delivery data corresponding to the influencing parameter includes: Determine the total number of historical delivery objects corresponding to the influencing parameter based on the number of historical delivery objects corresponding to each parameter value of the influencing parameter; Determine the total number of historical conversion objects corresponding to the influencing parameter based on the number of historical conversion objects corresponding to each parameter value of the influencing parameter; Determine the conversion share of the influencing parameter corresponding to the parameter value based on the number of historical conversion objects corresponding to the influencing parameter and the number of historical total delivery objects corresponding to the influencing parameter; Determine the placement occupancy rate of the influencing parameter corresponding to the parameter value based on the number of historical placement objects corresponding to the influencing parameter and the total number of historical placement objects corresponding to the influencing parameter; An object screening threshold corresponding to the parameter value of the influencing parameter is determined according to the conversion share and the delivery share of the influencing parameter corresponding to the parameter value.

3. The method according to claim 2, characterized in that Determining the object screening threshold corresponding to the parameter value of the influencing parameter based on the conversion share and the delivery share of the influencing parameter corresponding to the parameter value includes: Obtaining an initial object screening threshold corresponding to the parameter value of the influencing parameter; Determine the difference between the conversion share and the delivery share corresponding to the value of the influencing parameter; The initial object screening threshold is adjusted according to the difference to obtain the object screening threshold corresponding to the parameter value of the influencing parameter, wherein if the difference is greater than the set value, the initial object screening threshold is reduced; if the difference is less than or equal to the set value, the initial object screening threshold is increased.

4. The method according to claim 3, characterized in that The adjusting the initial object screening threshold according to the difference to obtain the object screening threshold corresponding to the parameter value of the influencing parameter includes: Determine the ratio between the difference and the placement occupancy rate of the influencing parameter corresponding to the parameter value; The initial object screening threshold is adjusted based on the ratio to obtain the object screening threshold corresponding to the parameter value of the influencing parameter.

5. The method according to claim 3 or 4, characterized in that If the difference is greater than the set value, the adjustment amplitude of the initial object screening threshold is positively correlated with the ratio; if the ratio is less than or equal to the set value, the adjustment amplitude of the initial object screening threshold is positively correlated with the absolute value of the ratio.

6. The method according to claim 1, characterized in that The determining of the target delivery object corresponding to the parameter value of the influencing parameter based on the object screening threshold corresponding to the parameter value of the influencing parameter and the delivery probability corresponding to each candidate delivery object includes: Determine, among the delivery probabilities corresponding to the candidate delivery objects corresponding to the parameter value of the influencing parameter, the candidate delivery object corresponding to the delivery probability greater than or equal to the object screening threshold corresponding to the parameter value of the influencing parameter as the target delivery object corresponding to the parameter value of the influencing parameter; The determining the target delivery object corresponding to the information to be pushed based on the target delivery object corresponding to each parameter value of each of the influencing parameters includes: The target delivery object corresponding to each parameter value of each of the influencing parameters is used as the candidate delivery object corresponding to the information to be pushed; A target delivery object corresponding to the information to be pushed is determined from candidate delivery objects corresponding to the information to be pushed.

7. The method according to claim 1, characterized in that The method further comprises: Pushing the information to be pushed to each target delivery object corresponding to each recommendation; Obtain feedback data corresponding to the information to be pushed from each target delivery object corresponding to the recommendation; According to the obtained feedback data, the conversion objects among the target delivery objects to be recommended are determined.

8. The method according to claim 1, characterized in that The candidate delivery object corresponding to a parameter value of an influencing parameter is determined in the following way: Obtain the initial object set and historical delivery object set corresponding to the parameter value of the influencing parameter; The initial object set is deduplicated based on the historical delivery object set, and each delivery object in the deduplicated initial object set is used as a candidate delivery object corresponding to the parameter value of the influencing parameter.

9. An information push device, characterized in that: include: a data acquisition module, configured to acquire first object data corresponding to information to be pushed, wherein the first object data includes second object data corresponding to at least one influencing parameter, wherein the second object data corresponding to each influencing parameter includes historical delivery data corresponding to at least two parameter values ​​of the influencing parameter and object information of at least one candidate delivery object; A total delivery data determination module is configured to determine, for each of the influencing parameters, the historical total delivery data corresponding to the influencing parameter based on the historical delivery data corresponding to each parameter value of the influencing parameter; a threshold determination module configured to determine, for each parameter value of each of the influencing parameters, an object screening threshold corresponding to the parameter value of the influencing parameter based on historical delivery data corresponding to the parameter value of the influencing parameter and historical total delivery data corresponding to the influencing parameter; a delivery object determination module, configured to obtain, for each parameter value of each influencing parameter, based on the object information of each candidate delivery object corresponding to the parameter value, a delivery probability corresponding to each candidate delivery object corresponding to the parameter value of the influencing parameter through a trained object screening model, and determine, based on the object screening threshold corresponding to the parameter value of the influencing parameter and the delivery probability corresponding to each candidate delivery object, a target delivery object corresponding to the parameter value of the influencing parameter; And based on the target delivery object corresponding to each parameter value of each of the influencing parameters, the target delivery object corresponding to the push information is determined.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory, wherein: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images