Information pushing method and device, equipment and storage medium

Abstract

The invention discloses an information pushing method and device, equipment and a storage medium, and relates to the technical field of computers. The method comprises the steps that first data information in an application program of at least one object is obtained, the first data information in the application program refers to data of the object executing a first operation through a channel in the application program, and the first data information comprises data associated with the first operation of the object; adopting a prediction model to predict the operation intention of the at least one object outside the application program according to the first data information to obtain a predicted object, the predicted object being an object having the operation intention of executing the first operation by adopting a channel outside the application program; and pushing first information to the prediction object, wherein the first information is used for prompting the prediction object to execute the first operation in the channel outside the application program. Through the prediction model, the accuracy of the prediction object is improved, and the benefit generated by information pushing is improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and particularly to an information push method, device, equipment and storage medium. Background Art

[0002] There are many channels for game players to obtain virtual resources in the game for virtual characters, including in-game channels and out-of-game channels. Since the acquisition operations of players are limited, to increase the number of operations of players in out-of-game channels, it is necessary to give an early reminder when players have the intention to perform out-of-game operations.

[0003] In the related art, the pushed information is usually formulated manually, which easily leads to inaccurate push population of information and low information push efficiency. Summary of the Invention

[0004] Embodiments of this application provide an information push method, device, equipment and storage medium. The technical solutions are as follows:

[0005] According to one aspect of the embodiments of this application, an information push method is provided. The method includes:

[0006] Obtain first data information in the application program of at least one object, where the first data information in the application program refers to the data of the object performing a first operation through the in-application channel, and the first data information includes data associated with the first operation of the object;

[0007] Use a prediction model to predict the operation intention of the at least one object outside the application program according to the first data information, and obtain a predicted object, where the predicted object refers to an object having the operation intention to perform the first operation through the out-of-application channel;

[0008] Push first information to the predicted object, where the first information is used to prompt the predicted object to perform the first operation through the out-of-application channel.

[0009] According to one aspect of the embodiments of this application, an information push device is provided. The device includes:

[0010] A data acquisition module, configured to obtain first data information in the application program of at least one object, where the first data information in the application program refers to the data of the object performing a first operation through the in-application channel, and the first data information includes data associated with the first operation of the object;

[0011] A willingness prediction module, configured to use a prediction model to predict, based on the first data information, the operation willingness of the at least one object outside the application program, so as to obtain a predicted object, where the predicted object refers to an object having the operation willingness to execute the first operation through a channel outside the application program;

[0012] An information push module, configured to push first information to the predicted object, where the first information is used to prompt the predicted object to execute the first operation through a channel outside the application program.

[0013] According to one aspect of the embodiments of the present application, a computer device is provided. The computer device includes a processor and a memory. A computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement the above information push method.

[0014] According to one aspect of the embodiments of the present application, a computer-readable storage medium is provided. A computer program is stored in the computer-readable storage medium, and the computer program is loaded and executed by a processor to implement the above information push method.

[0015] According to one aspect of the embodiments of the present application, a computer program product is provided. The computer program product includes a computer program, and the computer program is loaded and executed by a processor to implement the above information push method.

[0016] The technical solution provided by the embodiments of the present application can bring the following beneficial effects:

[0017] By using a prediction model, based on the data of at least one object executing a first operation through a channel inside the application program, the operation willingness of the at least one object outside the application program is predicted to obtain a predicted object, and first information is pushed to the predicted object to prompt the predicted object to execute the first operation through a channel outside the application program. Thus, relatively accurate objects with the operation willingness outside the application program are obtained, the accuracy of the predicted objects is improved, and the accuracy of the population to which the first information is pushed is also improved. And due to the improvement of the accuracy of the population to which the first information is pushed, the number of executions of the first operation outside the application program increases after the first information is pushed, the benefit generated by the information push is improved, and the efficiency of the information push is also enhanced. Description of the Drawings

[0018] Figure 1 is a schematic diagram of the implementation environment of the solution provided by an embodiment of the present application;

[0019] Figure 2 is a flowchart of the information push method provided by an embodiment of the present application;

[0020] Figure 3It is a schematic diagram of the life cycle of an object with a willingness to participate in an application provided by an embodiment of the present application;

[0021] Figure 4 It is a schematic diagram of the design process of a data logging solution provided by an embodiment of the present application;

[0022] Figure 5 It is a schematic diagram of the construction principle of a hyperplane provided by an embodiment of the present application;

[0023] Figure 6 It is a block diagram of an information push device provided by an embodiment of the present application;

[0024] Figure 7 It is a block diagram of the structure of a computer device provided by an embodiment of the present application. Detailed implementation manners

[0025] To make the objectives, technical solutions and advantages of the present application clearer, the following will further describe the embodiments of the present application in detail with reference to the accompanying drawings.

[0026] Please refer to Figure 1 , which shows a schematic diagram of the solution implementation environment provided by an embodiment of the present application. This solution implementation environment can be implemented as an information push system. This solution implementation environment may include: a terminal device 10 and a server 20.

[0027] The number of terminal devices 10 can be one or more. The terminal device 10 can be an electronic device such as a mobile phone, a tablet computer, a game console, an e-book reader, a multimedia playback device, a wearable device, a PC (Personal Computer), a vehicle-mounted terminal, etc. A client of a certain application can be installed in the terminal device 10. This application is used to implement the function of performing a first operation through a channel within the application involved in the present application. This application can be any one of the following types of applications, such as a game application, a video application, a life application, a social application, etc. Optionally, this application can be an application that needs to be downloaded and installed, or an application that can be used immediately upon clicking. The embodiments of the present application do not make any limitations in this regard.

[0028] In some embodiments, the application is a game application, including shooting applications, racing applications, multiplayer online battle arena games, etc. The game application can be an application developed based on a three-dimensional virtual environment engine or an application developed based on a two-dimensional virtual environment engine. This application does not limit this. The first operation can be an operation for an object to obtain virtual resources within the game application, and various types of virtual resources can be obtained. For example, virtual props for quickly passing through the game process, virtual clothing for decorating virtual characters, virtual equipment for enhancing the attack ability of virtual characters, etc. can be obtained, or tickets for additional game levels, additional game tasks, additional virtual scenes, etc. can also be obtained.

[0029] The execution of the first operation involved in this application can adopt different channels. For example, the first operation can be executed within the application or outside the application. Generally, executing the first operation outside the application can participate in relevant activities. In some embodiments, if executing the first operation outside the application and within the application can both participate in activities, the activity intensity of the channel outside the application is stronger than that of the channel within the application.

[0030] The server 20 is used to provide background services for the client of the application installed and running in the terminal device 10. For example, the server 20 can be the background server of the above game application. The server 20 can be a single server, a server cluster composed of multiple servers, or a cloud computing service center. Optionally, the server 20 provides background services for the applications in multiple terminal devices 10 at the same time. The terminal device 10 and the server 20 can communicate with each other through a network.

[0031] In the embodiment of this application, an object executes the first operation within the application of the terminal device. The server obtains at least one object's first data information within the application through the network, and uses a prediction model to predict the operation intention of the at least one object outside the application based on the first data information of the at least one object within the application, obtaining a predicted object. The server pushes the first information to the predicted object to prompt the predicted object to execute the first operation through the channel outside the application, increasing the execution quantity of the first operation of the channel outside the application.

[0032] Please refer to Figure 2 , which shows a flowchart of the information push method provided by an embodiment of this application. The execution subject of each step of this method can be a computer device, specifically referring to the server 20 in the above information push system. This method can include at least one of the following steps 210 to 230:

[0033] Step 210, obtain first data information within the application of at least one object. The first data information within the application refers to the data of the object performing a first operation through a channel within the application. The first data information includes data associated with the first operation of the object.

[0034] The object may refer to a virtual account registered in the application through a terminal device. The first operation can be performed through a channel within the application using this virtual account. The first operation of the object can be performed through two different channels, including the first operation performed through a channel within the application and the first operation performed through a channel outside the application. Among them, the channel within the application refers to the operation channel provided within the application. The first operation performed through the channel within the application means the first operation that can be directly implemented within the application. The channel outside the application refers to the operation channel provided by other applications. The first operation performed through the channel outside the application means the first operation that needs to be implemented in other applications outside the application.

[0035] The first operation refers to an operation that can be performed either through a channel within the application or through a channel outside the application. For example, the first operation can be an operation to obtain virtual resources, a live broadcast operation, etc.

[0036] It should be noted that the "first operation" mentioned in the embodiments of the present application refers to the first operation performed by the object through a channel within the application and does not involve the first operation performed by the object through a channel outside the application. Therefore, what is obtained in step 210 is the data associated with the first operation of at least one object, that is, the data of at least one object performing the first operation through a channel within the application.

[0037] In some embodiments, the first operation is an operation to obtain virtual resources, and the first data information is the data for obtaining virtual resources. Then, obtain the data for obtaining virtual resources within the application of at least one object. The data for obtaining virtual resources within the application refers to the relevant data of the object obtaining virtual resources through a channel within the application. The data for obtaining virtual resources includes data associated with the obtaining operation of the object.

[0038] The operation of an object to obtain virtual resources can be carried out through two different channels, including the operation of obtaining virtual resources through the channel within the application and the operation of obtaining virtual resources through the channel outside the application. The channel outside the application refers to the channel for obtaining virtual resources provided by other applications. The operation of obtaining virtual resources through the channel outside the application refers to the operation of obtaining virtual resources that needs to be implemented in other applications outside the application. For example, if the virtual resource is the phone bill of a communication device, then if the obtaining operation is implemented within the communication application, it is to obtain virtual resources through the channel within the application; if the obtaining operation is implemented in other applications such as a shopping application or a payment application, it is to obtain virtual resources through the channel outside the application.

[0039] In some embodiments, the first data information obtained in step 210 is the first data information within the application of at least one object in the first time period. The end time of the first time period can be the current time, that is, step 210 needs to obtain the real-time data information of at least one object within the application, and combine the first data information of at least one object within the application in the historical time period to obtain the first data information of at least one object within the application in the first time period. The end time of the first time period can also be any historical time before the current time, then the first time period refers to the historical time period. The first time period can be any time period such as 1 year, half a year, 3 months, etc., and the present application does not make a limitation thereto. If the end time of the first time period is the current time, the first data information obtained in step 210 is the latest data information of at least one object, and the prediction of the operation intention of at least one object outside the application will be more accurate.

[0040] In some embodiments, one piece of first data information refers to the relevant data generated when an object performs a first operation. One piece of first data information includes at least one of the following: object data, operation time data, operation environment data, and operation content data.

[0041] Among them, the object data is used to indicate the attribute data of the object corresponding to the first operation. The object data may include data such as the level of the object and the registration time of the object.

[0042] The operation time data is used to indicate the occurrence time of the first operation and the data collection time of the first data information. Since there is a data transmission time from the occurrence of the first operation to the collection of the first data information, the operation time data includes the occurrence time of the first operation and the data collection time of the first data information. The occurrence time of the first operation refers to the occurrence time of the first operation obtained at the same time when the first operation occurs on the terminal device. The data collection time of the first data information refers to the time when the relevant interface of the server receives the occurrence time of the first operation after the terminal device collects the occurrence time of the first operation and transmits the occurrence time of the first operation to the relevant interface of the server.

[0043] The operating environment data is used to indicate the device environment and network environment where the first operation occurs. The device environment refers to the device environment of the terminal device where the first operation occurs, specifically including the network address of the terminal device where the first operation occurs and the device conditions of the terminal device when performing the first operation. The network address of the terminal device may be the IP address (Internet Protocol Address) of the terminal device, and the device conditions of the terminal device may include relevant parameters of the device such as the hardware data of the terminal device and the device temperature of the terminal device. Optionally, the device environment may further include the device data of the server to which the first data information is to be reported. The network environment includes the network type used when the first operation occurs and the network operating conditions when the first operation occurs. The network type may be one of a wide area network, a metropolitan area network, and a local area network, and the network operating conditions include conditions such as network speed and network interference.

[0044] The operation content data is used to indicate the operation content of the first operation. Specifically, it includes the event identifier of the first operation and the event parameters of the first operation. The event identifier of the first operation is used to indicate the first operation and can also be used to indicate this first data information. For example, the event identifier of the first operation may be data in binary form. The event parameters of the first operation are used to represent the operation parameters of the first operation and may include operation parameters such as the operation object, operation steps, and operation time of the first operation.

[0045] Step 220: Use a prediction model to predict the operation willingness of at least one object outside the application based on the first data information, and obtain a predicted object. The predicted object refers to an object that has the operation willingness to perform the first operation through a channel outside the application.

[0046] The operation willingness of an object includes the operation willingness to perform the first operation through a channel within the application and the operation willingness to perform the first operation through a channel outside the application. The prediction model is used to predict the operation willingness of an object to perform the first operation through a channel outside the application, and obtain an object that has the operation willingness to perform the first operation through a channel outside the application, which is called a predicted object.

[0047] Step 220 includes at least one sub-step among steps 221 to 222 (not shown in the figure).

[0048] Step 221: Organize the first data information to obtain the first operation data corresponding to each object respectively.

[0049] In step 210, at least one piece of first data information of channels within an application for at least one object performing a first operation over a period of time is obtained, and the at least one piece of first data information for each object is sorted out to obtain first operation data respectively corresponding to each object. The first operation data is data obtained by integrating, classifying, and summarizing at least one piece of first data information, and this first operation data can be used to indicate the operation habits and operation characteristics of the object over a period of time.

[0050] In some embodiments, the first operation data includes at least one of the following: basic feature data, operation feature data, activity feature data.

[0051] The basic feature data is used to indicate the attribute data of the object; the basic feature data includes at least one of the following: the level of the object, the registration time of the object. Optionally, if the application is a game application, the basic feature data may further include the combat power value of the object, the virtual character used by the object, etc.

[0052] The operation feature data is used to indicate the operation data during the object's execution of the first operation; the operation feature data includes at least one of the following: click data of the operation interface of the first operation, exposure data of the operation interface of the first operation, click data of the confirmation execution button, execution success data of the first operation. Each time the object's first operation accesses a web page, it can be recorded as a PV (Page Views), then according to the access records of each first operation to each web page, feature data such as the click count of the operation interface of the first operation, the exposure time of the operation interface of the first operation, the click count of the confirmation execution button, the exposure time of the execution success interface of the first operation, and the jump time of the payment page can be obtained.

[0053] Exemplarily, if the first operation is an operation for obtaining virtual resources, the operation feature data is used to indicate the operation data during the object's execution of the operation for obtaining virtual resources. The operation feature data includes at least one of the following: click data of each virtual resource, exposure data of the acquisition interface of each virtual resource, click data of the confirmation acquisition button, acquisition success data of each virtual resource, acquisition quantity data of the object's virtual resources. According to the PV volume of each acquisition operation for the selection interface of each virtual resource, the acquisition interface of each virtual resource, and the acquisition success interface of each virtual resource, feature data such as the click count of each virtual resource, the exposure time of the acquisition interface of each virtual resource, the click count of the confirmation acquisition button, the exposure time of the acquisition success interface of each virtual resource, and the jump time of the acquisition page can be obtained. The acquisition quantity data of the object's virtual resources may refer to the acquisition quantity of regular virtual resources, such as the acquisition quantity of virtual resources per week or per month, or may also be the acquisition quantity of each acquisition operation of virtual resources, and this application does not make a limitation on this.

[0054] Activity feature data is data used to indicate the first operation when an object participates in an activity; the activity feature data includes at least one of the following: the number of times of participating in each activity, the number of times of collecting each activity. An activity refers to an activity related to the first operation, for example, it can be a preferential activity, a welfare activity, etc. for performing the first operation. The number of times of participating in each activity refers to the number of times of the first operation performed when the object participates in the activity within the application. The number of times of collecting each activity refers to the number of times of collecting each activity by the object. The collected activities can be activities that the object has collected and participated in, or activities that the object has only collected but not participated in.

[0055] Exemplarily, if the first operation is an operation for obtaining virtual resources, then the activity feature data is used to indicate the data for obtaining virtual resources when the object participates in a preferential activity; the activity feature data includes at least one of the following: the number of received preferential packages at each level, the number of participations in preferential activities at each level, the total number of participations in preferential activities, the number of collections of preferential activities. The activity feature data refers to the data for obtaining virtual resources when the object participates in the preferential activity within the application. The preferential activity can be preferential packages at each level, such as coupons at each level, or a lottery activity, etc. The number of participations in preferential activities at each level refers to the number of participations in obtaining virtual resources by using preferential packages at each level to participate in preferential activities. The total number of participations in preferential activities can be the total number of participations in each preferential activity, or the total number of participations in obtaining virtual resources by using preferential packages at each level.

[0056] In some embodiments, if the first operation is an operation for obtaining virtual resources, then the first operation data may further include acquisition feature data, and the acquisition feature data is used to indicate the quantity data of the virtual resources of the object for the application. The acquisition feature data includes at least one of the following: the total amount of virtual resources acquired, the number of days for the total acquisition of virtual resources, the number of days between adjacent acquisition operations, the maximum amount of virtual resources acquired in a single acquisition operation, the amount of virtual resources acquired regularly. Optionally, it may further include the acquisition date of the most recent acquisition operation of virtual resources, the amount of virtual resources acquired in the most recent acquisition operation, the minimum amount of virtual resources acquired in a single acquisition operation, etc. Among them, the amount of virtual resources acquired regularly can be the amount of virtual resources acquired every day in the most recent week, or the amount of virtual resources acquired every week in the most recent month, or the amount of virtual resources acquired every month in the most recent half year. The present application does not make a limitation on this.

[0057] Step 222, use a prediction model to predict the operation intention of at least one object outside the application according to the first operation data, and obtain a predicted object.

[0058] The input data of the prediction model is the first operation data corresponding to at least one object respectively, and the output data of the prediction model is the prediction result for the first operation data corresponding to each object respectively. The prediction result is used to indicate whether each object has the operation intention to execute the first operation through a channel outside the application. The prediction result can be in numerical form. For example, if the prediction result is the first numerical value, it is used to indicate that the object has the operation intention to execute the first operation through a channel outside the application; if the prediction result is the second numerical value, it is used to indicate that the object does not have the operation intention to execute the first operation through a channel outside the application. Exemplarily, the first numerical value can be 1, and the second numerical value can be 0.

[0059] It should be noted that the prediction result being the second numerical value indicates that the object does not have the operation intention to execute the first operation through a channel outside the application, which does not necessarily mean that the object has the operation intention to execute the first operation through a channel inside the application. The prediction result being the second numerical value can be used to indicate that the object has the operation intention to execute the first operation through a channel inside the application, or it can also be used to indicate that the object does not have the operation intention.

[0060] According to the prediction result output by the prediction model, at least one object is classified to obtain the predicted objects having the operation intention outside the application.

[0061] By sorting out the first data information of each object to obtain the first operation data of each object, the information volume of the input data of the prediction model is enriched, and the operation characteristics of each object can be measured. Thus, the prediction model can predict whether each object has the operation intention outside the application according to the operation habits of each object, improving the accuracy of the operation intention prediction of the prediction model.

[0062] Step 230: Push the first information to the predicted object. The first information is used to prompt the predicted object to execute the first operation through a channel outside the application.

[0063] The first information refers to the activity information about the first operation in other applications outside the application. The first information can include the prompt information of the preferential activity for executing the first operation, the prompt information of the welfare activity for executing the first operation, the information for receiving the coupon for executing the first operation, etc. Usually, the activity intensity corresponding to the first information is better than that of the channel inside the application.

[0064] Since the participation ability of the object is limited, such as Figure 3As shown, it shows the life cycle of an object with the willingness to participate in an application, including the pull stage, the active stage, the stable stage, the silent stage, and the churn stage. When an application is just launched, it will attract objects with the willingness to participate. Therefore, the number of objects with the willingness to participate in the pull stage will increase. When these objects with the willingness to participate actively participate within the application, they will attract other objects to participate in the application. Therefore, the number of objects with the willingness to participate in the active stage will continue to increase. Then, the participation status of the objects tends to be stable. The objects with the willingness to participate will continue to participate in the application, but it is difficult to attract other objects to participate in the application. Therefore, the growth of the objects with the willingness to participate in the stable stage is slow. After a period of time, the enthusiasm of the objects to participate in the application decreases. Therefore, the number of objects with the willingness to participate in the silent stage begins to decrease. Finally, the objects with the willingness to participate gradually churn.

[0065] Therefore, it is necessary to push the first information to the predicted objects to increase the number of executions of the first operation outside the application. The first information is used to prompt the predicted objects to execute the first operation within the application through channels outside the application. After obtaining the predicted objects by using the prediction model in step 222 above, the server will automatically push the first information to the predicted objects.

[0066] The technical solution provided by the embodiments of the present application, by using a prediction model, predicts the operation willingness of at least one object outside the application according to the data of at least one object executing the first operation through channels within the application, obtains the predicted objects, and pushes the first information to the predicted objects to prompt the predicted objects to execute the first operation through channels outside the application. Thus, the objects with the operation willingness outside the application are obtained more accurately, the accuracy of the predicted objects is improved, and the accuracy of the population to which the first information is pushed is improved. And due to the improvement of the accuracy of the population to which the first information is pushed, the number of executions of the first operation outside the application increases after the first information is pushed, the benefit generated by the information push is improved, and the efficiency of the information push is also improved.

[0067] In some embodiments, step 210 includes at least one sub-step among steps 211 to 213 (not shown in the figure).

[0068] Step 211, delivering the event model to each data collection node that needs to collect data during the process of the object executing the first operation. The event model is a data collection model designed based on the buried point scheme, and the buried point scheme is used to indicate each data collection node.

[0069] According to the first operation performed within the application, determine the data collection requirements for the first operation of the application, and based on the data collection requirements, determine the data to be collected for the first operation, and design a data tracking plan according to the data to be collected for the first operation. The generation process of the above data tracking plan is completed manually. The data tracking mentioned in this application refers to the related technologies and implementation processes of capturing, processing, and sending for each first operation of the object. The data tracking plan is used to indicate the data collection nodes corresponding to each piece of data to be collected during the execution of the first operation by the object. The data collection nodes include the front-end data collection nodes in the terminal device, the data collection nodes of the application client, and the back-end data collection nodes. Among them, the front-end data collection nodes refer to the nodes of each piece of data corresponding to the display interface of the application during the generation process of the first data information. The data collection nodes of the application client refer to the nodes that collect each piece of data during the generation process of the first data information. The back-end data collection nodes refer to the nodes that collect each piece of data during the transmission process of the first data information.

[0070] The design process of the data tracking plan can refer to Figure 4 As shown, first determine the data collection requirements, and give a complete definition of the data requirements document. Then, based on this data requirements document, check whether the existing data tracking meets the data collection requirements. If it meets the data collection requirements, obtain the data on each data collection node according to the data tracking plan, and analyze the node data through self-service query on each analysis and query platform. If it does not meet the data collection requirements, re-determine the development resources, and according to the first operation within the application, determine whether the data collection requirements in the data requirements document meet the simple requirements corresponding to the first operation within the application. If the existing data collection requirements meet the simple requirements, adjust and design the data tracking plan in the data tracking platform, and conduct self-testing in the data tracking platform based on the adjusted data tracking plan to obtain the data on each data collection node. Finally, analyze the node data through self-service query on each analysis and query platform. If the existing data collection requirements do not meet the simple requirements, it is necessary to re-evaluate and judge the data collection requirements, re-design the data tracking plan for the data group, conduct self-testing in the data tracking platform based on the re-designed data tracking plan, and jointly adjust the data tracking plan to obtain the data on each data collection node. Finally, analyze the node data through self-service query on each analysis and query platform.

[0071] In the embodiment of this application, after obtaining the data tracking plan, design an event model based on the data tracking plan. The event model is a data collection model that carries the data tracking plan and carries the data to be collected for the first operation. The event model can be expressed as:

[0072] who object data (attribute data of the object corresponding to the first operation)

[0073] when operation time data (the occurrence time of the first operation and the data collection time of the first data information)

[0074] where operation environment data (the device environment and network environment where the first operation occurs)

[0075] what operation content data (the operation content of the first operation)

[0076] The event model is an extensible model. By adjusting the event model, the data to be collected during the execution of the first operation by the object can be adjusted. For example, the data included in the event model can be added, deleted, or repositioned, etc.

[0077] Transmit the event model to each data collection node. Then, each time the object executes the first operation, the event model will be triggered to collect the first data information once, obtaining a data log corresponding to the first data information. What is obtained in step 210 is the data logs corresponding to each piece of first data information. Each piece of first data information includes the node data obtained by each data collection node.

[0078] Step 212, receive the node data sent by at least one terminal device corresponding to at least one object. The node data is the data on each data collection node obtained by the terminal device using the event model when the first operation occurs.

[0079] In response to the object executing the first operation, trigger the event model to collect the node data on each data collection node. The terminal device sends the first data information corresponding to each first operation obtained by using the event model to the server, and then the server receives the node data sent by each terminal device.

[0080] Each object can log in to different terminal devices to obtain their respective node data. Multiple objects can also log in to the same terminal device to obtain their respective node data. One object can also log in to multiple terminal devices to obtain the node data of that object.

[0081] Step 213, obtain the first data information in the application of at least one object according to the node data corresponding to at least one object.

[0082] According to at least one piece of node data collected by each object on at least one terminal device, obtain at least one piece of first data information in the application of each object.

[0083] By adopting an event model to obtain the first data information within an application of at least one object, based on the scalability of the event model, the node data to be collected can be freely adjusted, thereby improving the flexibility and richness of the first data information obtained within the application, which helps to improve the accuracy of the prediction results of the prediction model.

[0084] In some embodiments, step 222 includes at least one sub-step among steps 2221 to 2222 (not shown in the figure).

[0085] Step 2221, obtain the hyperplane corresponding to the binary classification algorithm on which the prediction model is based. Among them, the prediction result of the first operation data located on the first side of the hyperplane after passing through the prediction model is the first value, and the prediction result of the first operation data located on the second side of the hyperplane after passing through the prediction model is the second value.

[0086] The prediction model is a binary classification model. Exemplarily, the prediction model can be an SVM model (Support Vector Mac, support vector machine). The SVM model can define a linear classifier with the largest margin in a feature space, and classify the input data through this linear classifier. Generally speaking, the SVM model needs to find a hyperplane in the feature space that divides the data of two categories, and make the distance from all data in the input data to this hyperplane the shortest. The data in the two categories that are closest to the hyperplane are the support vectors. Determining the support vectors can determine the hyperplane.

[0087] A hyperplane is a linear subspace in an n-dimensional Euclidean space with a co-dimension equal to 1. That is to say, a hyperplane is a linear space of n - 1 dimensions, where n is an integer greater than 3. The feature space contains the first operation data corresponding to at least one object respectively. The hyperplane divides the feature space into two subspaces, and thus also divides the first operation data corresponding to at least one object respectively into two parts of data. Among them, the prediction result of the first operation data in the subspace on the first side of the hyperplane after passing through the prediction model is the first value, and the object corresponding to the first operation data in this side of the subspace has the willingness to perform the first operation through a channel outside the application. The prediction result of the first operation data in the subspace on the second side of the hyperplane after passing through the prediction model is the second value, and the object corresponding to the first operation data in this side of the subspace does not have the willingness to perform the first operation through a channel outside the application.

[0088] Step 2222: Using the prediction model and based on the hyperplane, determine the object corresponding to the first operation data with a prediction result of the first value as the prediction object. The first value is used to indicate that the object corresponding to the first operation data on the first side has the willingness to operate outside the application, and the second value is used to indicate that the object corresponding to the first operation data on the second side does not have the willingness to operate outside the application.

[0089] Using the prediction model and based on the hyperplane, output the prediction results corresponding to each piece of first operation data, and determine the object corresponding to the first operation data with a prediction result of the first value as the prediction object.

[0090] By using a prediction model based on a binary classification algorithm and according to the first operation data corresponding to each object, predict the willingness of each object to operate outside the application, which improves the accuracy of the predicted object, and thus can improve the accuracy of the population to whom the first information is pushed.

[0091] In some embodiments, the process of obtaining the hyperplane corresponding to the binary classification algorithm on which the prediction model is based can refer to the following steps, and the schematic diagram of the construction principle of the hyperplane can refer to Figure 5 as shown.

[0092] 1. Construct the functional equation of the hyperplane.

[0093] The functional equation of the hyperplane can be expressed as:

[0094] w T x + b = 0

[0095] where x represents the first operation data corresponding to each object in the n-dimensional feature space, w = (w1, w2, …, w n ), w represents the normal vector, w T represents the transposed vector of w, b represents the displacement term, and the value of b represents the distance between the hyperplane and the origin.

[0096] Therefore, the hyperplane can be determined by determining the normal vector w and the displacement term b.

[0097] 2. According to the functional equation of the hyperplane, obtain the distance function corresponding to the hyperplane. The distance function is used to indicate the sum of the distances from the first plane and the second plane to the hyperplane. The first plane refers to the critical plane of the first operation data corresponding to the object with the willingness to operate outside the application in the first operation data, and the second plane refers to the critical plane of the first operation data corresponding to the object without the willingness to operate outside the application in the first operation data.

[0098] According to the functional equation of the hyperplane, the distance from any point x in the feature space to the hyperplane can be obtained, which can be expressed as:

[0099]

[0100] Assume that the hyperplane can correctly classify the first operation data corresponding to each object. That is, for each point x in the feature space, we can obtain:

[0101]

[0102] As Figure 7 shown, the several points closest to the hyperplane can be those where the equal sign holds in the above lions. These points are called support vectors. The sum of the distances from these two different types of support vectors to the hyperplane can be expressed as:

[0103]

[0104] Among them, the first plane refers to the plane corresponding to w T x + b = 1, that is, the critical plane where the support vectors with the first numerical prediction result are located. The second plane refers to the plane corresponding to w T x + b = -1, that is, the critical plane where the support vectors with the second numerical prediction result are located.

[0105] Taking the maximization of the distance function corresponding to the hyperplane as the goal, the following objective function can be obtained:

[0106]

[0107] To maximize the value of the distance function corresponding to the hyperplane, it is only necessary to maximize ‖w‖ -1 , which is also equivalent to minimizing ‖w‖ 2 . Therefore, the above objective function can be rewritten to obtain the rewritten objective function (i.e., the distance function corresponding to the hyperplane), which can be expressed as:

[0108]

[0109] 3. Use the Lagrange multiplier method to calculate the Lagrangian function corresponding to the distance function.

[0110] For the rewritten objective function using the Lagrange multiplier method to obtain the Lagrangian function, which can be expressed as:

[0111]

[0112] Among them, α = (α1, α2, …, α n ).

[0113] Take the partial derivatives of L(w, b, α) with respect to w and b, and set the partial derivatives of w and b to 0 to obtain the following constraint conditions:

[0114]

[0115] Substitute the above constraints into \(L(w, b, \alpha)\) and simplify \(L(w, b, \alpha)\) to obtain the simplified Lagrangian function, which can be expressed as:

[0116]

[0117] 4. According to the first operation data, with the goal of minimizing the Lagrangian function, the Lagrangian function is constrained to obtain the objective function equation when the value of the Lagrangian function is minimized.

[0118] With the goal of minimizing the Lagrangian function, the dual problem of minimizing the Lagrangian function can be obtained, which can be expressed as:

[0119]

[0120] The constraint conditions of this dual problem can be synthesized as:

[0121]

[0122] In addition, the KKT conditions (Karush-Kuhn-Tucker Conditions) of this dual problem can be expressed as

[0123]

[0124] 5. Determine the hyperplane according to the objective function equation.

[0125] According to the constraint conditions and KKT conditions of the above dual problem, the optimal solutions of \(w\) and \(b\) can be determined, and thus the hyperplane can be determined.

[0126] In some embodiments, a classification decision function is used to obtain the prediction results of each point in the feature space according to the hyperplane, which can be expressed as:

[0127] \(f(x)=\text{sign}(w * x + b * )

[0128] where \(w * and \(b * are the optimal solutions of \(w\) and \(b\), and \(w * and \(b * can be expressed as:

[0129]

[0130]

[0131] Through the above method for determining the hyperplane, the operation intentions of each object outside the application can be output, and the determination method of the above hyperplane helps to improve the accuracy of predicting objects.

[0132] In some embodiments, the process of obtaining the above hyperplane can be optimized, and the optimization process can refer to the following steps.

[0133] 1. Construct the functional equation of the hyperplane.

[0134] The functional equation of the hyperplane can be expressed as:

[0135] w T x + b = 0

[0136] 2. Introduce slack variables for the first operation data of each object respectively. According to the functional equation of the hyperplane, the updated distance function is obtained. The slack variable is used to indicate the interval by which the first operation data deviates from its corresponding critical surface.

[0137] The slack variable is used to allow some of the first operation data to be on the wrong side of the hyperplane. At this time, the constraint condition after introducing the slack variable is updated to:

[0138] y i (w T x i + b) ≥ 1 - ξ i , i = 1, 2, …, n

[0139] where ξ i represents the interval by which the i-th point in the feature space deviates from its correct critical surface. If ξ i is allowed to be arbitrarily large, then any hyperplane meets the conditions. Therefore, on the basis of the original objective, it is necessary to make the sum of ξ i as small as possible.

[0140] Then the updated objective function can be expressed as:

[0141]

[0142] where C is used to control the maximum margin between two different types of support vectors and ensure that the functional margins of most points are less than 1, the weights of these two objectives. After the above updated objective function is further optimized, the updated distance function can be obtained, which can be expressed as:

[0143]

[0144] 3. Use the Lagrange multiplier method to calculate the updated Lagrangian function.

[0145] The updated Lagrangian function can be expressed as:

[0146]

[0147] Take the partial derivatives of \(L(w, b, \alpha, \xi)\) with respect to \(w\), \(b\), and \(\xi\), and set the partial derivatives of \(w\), \(b\), and \(\xi\) to 0. After substituting into \(L(w, b, \alpha, \xi)\), simplify \(L(w, b, \alpha)\) to obtain the simplified Lagrangian function, which can be expressed as:

[0148]

[0149] It can be seen that the expression of the updated Lagrangian function is the same as the Lagrangian function without introducing slack variables above. The difference lies in the constraint conditions of the two.

[0150] 4. According to the first operation data, with the goal of minimizing the updated objective Lagrangian function and minimizing the sum of each slack variable, obtain the updated objective function equation when the value of the updated objective Lagrangian function is minimized.

[0151] Since the expression of the updated Lagrangian function has not changed, the dual problem of minimizing the updated Lagrangian function has not changed, and the KKT conditions of this dual problem have not changed.

[0152] The constraint conditions of this dual problem can be updated to:

[0153]

[0154] The constraint condition of the slack variable can be simplified to: \(\xi\) i = max(0, 1 - y i (w T x i + b)).

[0155] 5. Determine the hyperplane according to the updated objective function equation.

[0156] According to the updated constraint conditions and KKT conditions of the above dual problem, the optimal solutions of \(w\) and \(b\) can be determined, and thus the hyperplane can be determined.

[0157] By introducing slack variables, it is allowed that some of the first operation data can be on the wrong side of the hyperplane, avoiding the situation where the hyperplane cannot meet the distribution of most of the first operation data due to some deviated first operation data, and also avoiding the situation where the hyperplane cannot be determined due to some deviated first operation data, increasing the flexibility and self - adaptability of the prediction model and reducing the prediction error rate of the prediction object.

[0158] The following is an apparatus embodiment of the present application, which can be used to execute the method embodiment of the present application. For details not disclosed in the apparatus embodiment of the present application, please refer to the method embodiment of the present application.

[0159] Please refer to Figure 6 , which shows a block diagram of an information push device provided by an embodiment of the present application. This device has the function of implementing the above-mentioned information push method, and this function can be implemented by hardware or by hardware executing corresponding software. This device can be the server introduced above or can be set in the server. As Figure 6 shown, the device 600 may include: a data acquisition module 610, a willingness prediction module 620, and an information push module 630.

[0160] The data acquisition module 610 is used to acquire first data information in the application of at least one object. The first data information in the application refers to data of the object performing a first operation through a channel in the application, and the first data information includes data associated with the first operation of the object.

[0161] The willingness prediction module 620 is used to use a prediction model to predict the operation willingness of the at least one object outside the application according to the first data information, and obtain a predicted object. The predicted object refers to an object having the operation willingness to perform the first operation through a channel outside the application.

[0162] The information push module 630 is used to push first information to the predicted object, and the first information is used to prompt the predicted object to perform the first operation through a channel outside the application.

[0163] In some embodiments, the data acquisition module 610 is used for:

[0164] Transmit an event model to each data acquisition node that needs to collect data during the process of the object performing the first operation. The event model is a data acquisition model designed based on a data embedding scheme, and the data embedding scheme is used to indicate each data acquisition node;

[0165] Receive node data sent by at least one terminal device corresponding to the at least one object. The node data is data on each data acquisition node obtained by the terminal device using the event model when the first operation occurs;

[0166] Obtain the first data information in the application of the at least one object according to the node data corresponding to the at least one object.

[0167] In some embodiments, the first data information includes at least one of the following: object data, operation time data, operation environment data, and operation content data;

[0168] Among them, the object data is used to indicate the attribute data of the object corresponding to the first operation, the operation time data is used to indicate the occurrence time of the first operation and the data acquisition time of the first data information, the operation environment data is used to indicate the device environment and network environment where the first operation occurs, and the operation content data is used to indicate the operation content of the first operation.

[0169] In some embodiments, the willingness prediction module 620 includes:

[0170] A data sorting unit, configured to sort the first data information to obtain first operation data corresponding to each object respectively.

[0171] A willingness prediction unit, configured to use the prediction model to predict the operation willingness of the at least one object outside the application program according to the first operation data, and obtain the predicted object.

[0172] In some embodiments, the first operation data includes at least one of the following: basic feature data, operation feature data, and activity feature data;

[0173] The basic feature data is used to indicate the attribute data of the object; the basic feature data includes at least one of the following: the level of the object, the registration time of the object;

[0174] The operation feature data is used to indicate the operation data during the process of the object executing the first operation; the click data of the operation interface of the first operation, the exposure data of the operation interface of the first operation, the click data of the confirmation execution button, the execution success data of the first operation;

[0175] The activity feature data is used to indicate the data of the first operation when the object participates in an activity; the activity feature data includes at least one of the following: the participation times of each activity, the collection times of each activity.

[0176] In some embodiments, the willingness prediction unit includes:

[0177] A hyperplane acquisition subunit, configured to acquire the hyperplane corresponding to the binary classification algorithm on which the prediction model is based, wherein the prediction result of the first operation data located on the first side of the hyperplane after passing through the prediction model is a first value, and the prediction result of the first operation data located on the second side of the hyperplane after passing through the prediction model is a second value.

[0178] A prediction subunit, configured to use the prediction model to determine, according to the hyperplane, an object corresponding to first operation data with a prediction result of the first value as the prediction object, where the first value is used to indicate that the object corresponding to the first operation data on the first side has an operation intention outside the application program, and the second value is used to indicate that the object corresponding to the first operation data on the second side does not have an operation intention outside the application program.

[0179] In some embodiments, the hyperplane acquisition subunit is configured to:

[0180] Construct a functional equation of the hyperplane;

[0181] According to the functional equation of the hyperplane, obtain a distance function corresponding to the hyperplane, where the distance function is used to indicate the sum of the distances from a first plane and a second plane to the hyperplane, the first plane refers to a critical plane of the first operation data corresponding to an object having an operation intention outside the application program in the first operation data, and the second plane refers to a critical plane of the first operation data corresponding to an object not having an operation intention outside the application program in the first operation data;

[0182] Use the Lagrange multiplier method to calculate a Lagrangian function corresponding to the distance function;

[0183] According to the first operation data, with the goal of minimizing the Lagrangian function, constrain the Lagrangian function to obtain an objective function equation when the value of the Lagrangian function is minimized;

[0184] Determine the hyperplane according to the objective function equation.

[0185] In some embodiments, the hyperplane acquisition subunit is further configured to:

[0186] Introduce slack variables for the first operation data of each object respectively, and according to the functional equation of the hyperplane, obtain an updated distance function, where the slack variable is used to indicate the interval by which the first operation data deviates from its corresponding critical plane;

[0187] Use the Lagrange multiplier method to calculate the updated Lagrangian function;

[0188] According to the first operation data, with the goal of minimizing the updated objective Lagrangian function and with the goal of minimizing the sum of all the slack variables, obtain an updated objective function equation when the value of the updated objective Lagrangian function is minimized;

[0189] Determine the hyperplane according to the updated objective function equation.

[0190] The technical solution provided by the embodiments of the present application uses a prediction model to predict the operation willingness of at least one object outside the application based on the data of at least one object performing a first operation through a channel within the application, obtaining predicted objects, and pushing first information to the predicted objects to prompt the predicted objects to perform the first operation through channels outside the application. Thus, the objects with the operation willingness outside the application are obtained more accurately, improving the accuracy of the predicted objects and the accuracy of the population to which the first information is pushed. And due to the improvement of the accuracy of the population to which the first information is pushed, the number of executions of the first operation outside the application increases after the first information is pushed, improving the benefits generated by the information push and also enhancing the efficiency of the information push.

[0191] It should be noted that for the device provided in the above embodiment, when implementing its functions, only the above division of each functional module is used for illustration. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the content structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device provided in the above embodiment and the method embodiment belong to the same concept, and the specific implementation process can be seen in the method embodiment, which will not be elaborated here.

[0192] Please refer to Figure 7 , which shows the structural block diagram of a computer device 700 provided by an embodiment of the present application. The computer device 700 can be any electronic device with data calculation, processing, and storage functions. The computer device 700 can be used to implement the information push method provided in the above embodiment.

[0193] Generally, the computer device 700 includes: a processor 701 and a memory 702.

[0194] The processor 701 may include one or more processing cores, such as a quad-core processor, an octa-core processor, etc. The processor 701 may be implemented in at least one of the following hardware forms: DSP (Digital Signal Processing), FPGA (Field Programmable Gate Array), and PLA (Programmable Logic Array). The processor 701 may also include a main processor and a coprocessor. The main processor is a processor used to process data in the wake state, also known as the CPU (Central Processing Unit); the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 701 may be integrated with a GPU (Graphics Processing Unit), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 701 may further include an AI processor, which is used to process computational operations related to machine learning.

[0195] The memory 702 may include one or more computer-readable storage media, and the computer-readable storage media may be non-transitory. The memory 702 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash storage devices. In some embodiments, the non-transitory computer-readable storage media in the memory 702 is used to store a computer program, and the computer program is configured to be executed by one or more processors to implement the above-mentioned information push method.

[0196] Those skilled in the art can understand that Figure 7 the structure shown in

[0197] does not constitute a limitation on the computer device 700, and may include more or fewer components than shown in the figure, or combine certain components, or adopt different component arrangements.

[0198] In an exemplary embodiment, a computer program product is further provided. The computer program product includes a computer program stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the above information push method.

[0199] It should be noted that, before and during the process of collecting relevant data of a user, this application can display a prompt interface, a pop-up window or output voice prompt information. The prompt interface, pop-up window or voice prompt information is used to prompt the user that their relevant data is being collected currently, so that this application only starts to execute the relevant steps of obtaining the user's relevant data after obtaining the confirmation operation sent by the user for the prompt interface or the pop-up window. Otherwise (that is, when the confirmation operation sent by the user for the prompt interface or the pop-up window is not obtained), the relevant steps of obtaining the user's relevant data are ended, that is, the relevant data of the user is not obtained. In other words, all user data collected by this application (including the first data information of the object) is processed strictly in accordance with the requirements of relevant national laws and regulations. Obtaining the informed consent or separate consent of the personal information subject is carried out under the condition that the user agrees and authorizes, and subsequent data use and processing behaviors are carried out within the scope authorized by laws and regulations and the personal information subject. The collection, use and processing of relevant user data need to comply with the relevant laws, regulations and standards of relevant countries and regions.

[0200] It should be understood that the "plurality" mentioned herein refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally represents an "or" relationship between the associated objects before and after. In addition, the step numbers described herein only exemplarily show a possible execution sequence between steps. In some other embodiments, the above steps may not be executed in the order of the numbers. For example, two steps with different numbers are executed simultaneously, or two steps with different numbers are executed in the reverse order of the illustration. The embodiments of this application do not make any limitations in this regard.

[0201] The above are only exemplary embodiments of this application, and are not intended to limit this application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of this application shall be included in the protection scope of this application.

Claims

1. An information push method, characterized in that, The method includes: Obtaining first data information within an application of at least one object, where the first data information within the application refers to data of the object performing a first operation through a channel within the application, and the first data information includes data associated with the first operation of the object; Using a prediction model to predict the willingness of the at least one object to perform an operation outside the application based on the first data information, obtaining predicted objects, where the predicted objects refer to objects having the willingness to perform the first operation through a channel outside the application; Pushing first information to the predicted objects, where the first information is used to prompt the predicted objects to perform the first operation through a channel outside the application.

2. The method according to claim 1, wherein The obtaining of the first data information within the application of the at least one object includes: Transmitting an event model to each data collection node that needs to collect data during the process of the object performing the first operation, where the event model is a data collection model designed based on a data embedding scheme, and the data embedding scheme is used to indicate each data collection node; Receiving node data sent by at least one terminal device corresponding to the at least one object, where the node data is data on each data collection node obtained by the terminal device using the event model when the first operation occurs; Obtaining the first data information within the application of the at least one object based on the node data corresponding to the at least one object.

3. The method according to claim 1, wherein The first data information includes at least one of the following: object data, operation time data, operation environment data, operation content data; Among them, the object data is used to indicate the attribute data of the object corresponding to the first operation, the operation time data is used to indicate the occurrence time of the first operation and the data collection time of the first data information, the operation environment data is used to indicate the device environment and network environment when the first operation occurs, and the operation content data is used to indicate the operation content of the first operation.

4. The method according to claim 1, wherein The using of the prediction model to predict the willingness of the at least one object to perform an operation outside the application based on the first data information and obtaining predicted objects includes: Sorting out the first data information to obtain first operation data corresponding to each object respectively; Using the prediction model to predict the willingness of the at least one object to perform an operation outside the application based on the first operation data, obtaining the predicted objects.

5. The method according to claim 4, wherein The first operation data includes at least one of the following: basic feature data, operation feature data, activity feature data; The basic feature data is used to indicate the attribute data of the object; the basic feature data includes at least one of the following: the level of the object, the registration time of the object; The operation feature data is used to indicate the operation data during the process of the object performing the first operation; the operation feature data includes at least one of the following: click data of the operation interface of the first operation, exposure data of the operation interface of the first operation, click data of the confirmation execution button, execution success data of the first operation; The activity feature data is data for indicating a first operation when the object participates in an activity; the activity feature data includes at least one of the following: the number of times of participation in each activity, the number of times of collection of each activity.

6. The method according to claim 4, characterized in that, The predicting, by using the prediction model according to the first operation data, the operation willingness of the at least one object outside the application program to obtain the predicted object includes: Obtaining a hyperplane corresponding to a binary classification algorithm on which the prediction model is based, wherein a prediction result of first operation data located on a first side of the hyperplane after passing through the prediction model is a first value, and a prediction result of first operation data located on a second side of the hyperplane after passing through the prediction model is a second value; Using the prediction model according to the hyperplane to determine, as the predicted object, an object corresponding to first operation data with a prediction result of the first value, where the first value is used to indicate that the object corresponding to the first operation data on the first side has an operation willingness outside the application program, and the second value is used to indicate that the object corresponding to the first operation data on the second side does not have an operation willingness outside the application program.

7. The method according to claim 6, wherein The obtaining a hyperplane corresponding to a binary classification algorithm on which the prediction model is based includes: Constructing a function equation of the hyperplane; According to the function equation of the hyperplane, obtaining a distance function corresponding to the hyperplane, where the distance function is used to indicate the sum of the distances from a first plane and a second plane to the hyperplane, the first plane being a critical plane of first operation data corresponding to an object having an operation willingness outside the application program among the first operation data, and the second plane being a critical plane of first operation data corresponding to an object not having an operation willingness outside the application program among the first operation data; Calculating a Lagrangian function corresponding to the distance function by using the Lagrange multiplier method; According to the first operation data, with the goal of minimizing the Lagrangian function, constraining the Lagrangian function to obtain an objective function equation when the value of the Lagrangian function is minimized; Determining the hyperplane according to the objective function equation.

8. The method according to claim 7, wherein The method further includes: Introducing a slack variable for each first operation data of the objects respectively, and according to the function equation of the hyperplane, obtaining an updated distance function, where the slack variable is used to indicate the interval by which the first operation data deviates from its corresponding critical plane; Calculating an updated Lagrangian function by using the Lagrange multiplier method; According to the first operation data, with the goal of minimizing the updated objective Lagrangian function and with the goal of minimizing the sum of each slack variable, obtaining an updated objective function equation when the value of the updated objective Lagrangian function is minimized; Determining the hyperplane according to the updated objective function equation.

9. An information push device, characterized in that, The device includes: A data acquisition module, configured to acquire first data information within an application of at least one object, where the first data information within the application refers to data of the object performing a first operation through a channel within the application, and the first data information includes data associated with the first operation of the object; A willingness prediction module, configured to use a prediction model to predict the operation willingness of the at least one object outside the application according to the first data information, and obtain a predicted object, where the predicted object refers to an object having the operation willingness to perform the first operation through a channel outside the application; An information push module, configured to push first information to the predicted object, where the first information is used to prompt the predicted object to perform the first operation through a channel outside the application.

10. A computer device, characterized in that, The computer device includes a processor and a memory, and a computer program is stored in the memory. The computer program is loaded and executed by the processor to implement the information push method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, A computer-readable storage medium stores a computer program, and the computer program is loaded and executed by a processor to implement the information push method according to any one of claims 1 to 8.

12. A computer program product, characterized in that, The computer program product includes a computer program, and the computer program is loaded and executed by a processor to implement the information push method according to any one of claims 1 to 8.

Images