Information pushing method and device, electronic equipment, readable medium and program product

Abstract

The invention relates to an information push method and device, electronic equipment, a readable medium and a program product, and the method comprises the steps: determining target push information to be pushed to a target object according to the first behavior data of the target object; determining push data which has pushed the target push information through the target push path; according to the first behavior data and the push data, push reference information for pushing target push information to the target object through a target push path is determined; and pushing the target push information to the target object according to the push reference information and the target push path. Therefore, matching between the target push information and the interest of the target object can be realized according to the behavior data of the target object, so that personalized information push is realized, and the information push effect is improved. Besides, the push reference information is determined according to the behavior data and the push data, so that the information push effect can be further improved when the target push information is pushed to the target object based on the push reference information.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular, to an information push method, device, electronic device, computer-readable medium, and computer program product. Background Art

[0002] With the rapid development of information technology, information push has become an important way of information dissemination.

[0003] When information is pushed, it is generally pushed based on the browsing data of the information (such as the number of views, click-through rate and / or browsing time). For example, for information with a large number of views, the information is pushed to more objects; for information with a small number of views, the information is not pushed. Summary of the invention

[0004] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0005] In a first aspect, the present disclosure provides an information push method, the information push method comprising: Determining target push information to be pushed to the target object according to the first behavior data of the target object; Determining that push data of the target push information has been pushed through the target push path; Determine, according to the first behavior data and the push data, push reference information for pushing the target push information to the target object through the target push path; The target push information is pushed to the target object according to the push reference information and the target push path.

[0006] In a second aspect, the present disclosure provides an information push device, the information push device comprising: A first determining module, configured to determine target push information to be pushed to the target object according to first behavior data of the target object; A second determining module, used to determine the push data of the target push information that has been pushed through the target push path; A third determining module, configured to determine, based on the first behavior data and the push data, push reference information for pushing the target push information to the target object through the target push path; A push module is used to push the target push information to the target object according to the push reference information and the target push path.

[0007] In a third aspect, the present disclosure provides an electronic device, including: a storage device having a computer program stored thereon; A processing device is used to execute the computer program in the storage device to implement the steps of the method in the first aspect.

[0008] In a fourth aspect, the present disclosure provides a computer-readable medium having a computer program stored thereon, which implements the steps of the method described in the first aspect when executed by a processing device.

[0009] In a fifth aspect, the present disclosure provides a computer program product, comprising a computer program, which implements the steps of the method described in the first aspect when executed by a processor.

[0010] Through the above technical solution, since the target push information is determined based on the behavior data of the target object, the target push information can be matched with the interests of the target object according to the behavior data of the target object, thereby realizing personalized information push and improving the information push effect. In addition, since the push data refers to the data that has pushed the target push information through the target push path, that is, the push data can represent the historical push data of the target push information under the target push path, when the push reference information is determined based on the push data and the behavior data, and the target push information is pushed to the target object based on the push reference information, the behavior data of the target object and the historical push data of the target push information can be taken into account at the same time, thereby further improving the information push effect.

[0011] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings: Figure 1 is a schematic diagram of an application environment of an information push method according to an exemplary embodiment of the present disclosure; Figure 2 is a flow chart of an information push method according to an exemplary embodiment of the present disclosure; Figure 3 is a flow chart of another information push method according to an exemplary embodiment of the present disclosure; Figure 4 is a flow chart of another information push method according to an exemplary embodiment of the present disclosure; Figure 5 The figure is a structural block diagram of an information push device according to an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION

[0013] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0014] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0015] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0016] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0017] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0018] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0019] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0020] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0021] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0022] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0023] The embodiments of the present application provide an information push method, device, electronic device, readable medium and program product. The application environment applicable to the embodiments of the present application is first described below.

[0024] See also Figure 1 , Figure 1 FIG. 1 is a schematic diagram showing an application environment of the information push method provided by an embodiment of the present application. As an implementation mode, the information push method provided by the embodiment of the present application can be applied to an electronic device. The electronic device can be as follows: Figure 1 The server 110 shown in the figure can be connected to the terminal device 120 through a network. The network is used to provide a medium for a communication link between the server 110 and the terminal device 120. The network can include various connection types, such as a wired communication link, a wireless communication link, etc., which are not limited in the embodiments of the present application. Optionally, in other embodiments, the electronic device can also be a smart phone, a laptop computer, etc.

[0025] It should be understood that Figure 1 The server 110, network and terminal device 120 are merely illustrative. Depending on the implementation requirements, there may be any number of servers, networks and terminal devices. For example, the server 110 may be a physical server or a server cluster composed of multiple servers, and the terminal device 120 may be a mobile phone, a tablet, a desktop computer, a laptop computer, and the like. It is understood that the embodiments of the present application may also allow multiple terminal devices 120 to access the server 110 at the same time.

[0026] In some embodiments, the terminal device 120 can collect data of the target object to obtain the behavior data of the object. Further, the terminal device 120 sends the behavior data of the target object to the server 110 through the network. After the server 110 receives the behavior data of the target object, the behavior data can be processed by the information push method described in the embodiment of the present application.

[0027] The embodiments of the present disclosure are further explained below with reference to the accompanying drawings.

[0028] Figure 2 is a flowchart of an information push method according to an exemplary embodiment of the present disclosure, referring to Figure 2 , the information push method may include the following steps: S201: Determine target push information to be pushed to the target object according to the first behavior data of the target object.

[0029] The first behavior data may be used to reflect the behavior and interaction mode of the target object on the digital platform. For example, the first behavior data may be data generated by the target object in activities such as browsing web pages, using applications, participating in social media and search engines, and these data may include click data, browsing data, search query data, purchase record data, video viewing data, social interaction data and / or location information data, etc. The first behavior data may be the real-time behavior data of the target object, or the behavior data of the target object within a preset time period, and the embodiments of the present disclosure do not impose any restrictions on this.

[0030] Among them, before obtaining the first behavior data of the target object, the target object should be informed and authorized by the target object in an appropriate manner in accordance with relevant laws and regulations. For example, when the target object uses the application for the first time, a prompt message can be sent to the target object to clearly prompt the target object that the use of the application will require the acquisition and use of the target object's personal information. Thereby, the target object can independently choose whether to provide personal information to the electronic device, application, server or storage medium or other software or hardware that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0031] Among them, according to the first behavior data of the target object, determining the target push information to be pushed to the target object can be to identify the subject of the target object's browsing information from the first behavior data through keyword extraction technology, thereby determining the information related to the subject as the target push information to be pushed to the target object; or it can be to construct an interest tag of the object according to the first behavior data of the target object, thereby determining the information related to the interest tag as the target push information to be pushed to the target object. Of course, it can also be determined in other ways, and the embodiments of the present disclosure do not impose any restrictions on this.

[0032] In a possible manner, the first behavior data may be used as an input to a machine learning model to obtain target push information to be pushed to a target object.

[0033] The machine learning model in this embodiment may be an existing model in the related art, or a model that is improved from an existing model in the related art, and the disclosed embodiment does not impose any restrictions on this. For example, the machine learning model in this embodiment may be a large language model in the related art, such as a Chat-based Generative Language Model (ChatGLM).

[0034] After determining the model structure of the machine learning model, in order to use the machine learning model to obtain target push information according to the first behavior data of the target object, it is also necessary to construct training data and train the target push information based on the training data.

[0035] For example, in an advertisement push scenario, historical behavior data (e.g., browsing history, purchase history, comment information, and / or click behavior data) and / or historical interest data (e.g., historically interested advertisements, etc.) of sample objects may be collected first. The collected data may then be preprocessed, including data cleaning, data deduplication, missing value filling, and / or data standardization, to obtain training data.

[0036] Since the interests and behavior patterns of different sample objects may be similar or identical, after obtaining the training data, the historical behavior data of the sample objects can be first input into the dialogue generation language model, and the historical behavior data of the sample objects can be encoded and processed through the dialogue generation language model to identify the similarities and differences between the sample objects, and based on the feature vector output by the dialogue generation language model, the sample objects are then grouped using a clustering algorithm to cluster sample objects with similar or identical feature vectors into a crowd package, thereby obtaining the grouping information of the sample objects, that is, obtaining the object category and the sample object information corresponding to each object category. That is to say, before training the dialogue generation language model, the grouping information of the sample objects can be obtained in the following way: Acquire historical behavior data of multiple sample objects; input the historical behavior data into a dialogue generation language model to obtain a feature vector for characterizing the content of interest of the sample objects; perform clustering processing on the feature vector to obtain grouping information of the sample objects.

[0037] Among them, the clustering processing of the feature vectors can be based on the K-means clustering algorithm, or based on the Gaussian mixture model clustering, or based on the density clustering algorithm. The embodiments of the present disclosure do not impose any restrictions on this.

[0038] After obtaining the grouping information of the sample objects, since each object category can represent a group of sample objects with similar interests and behavior patterns, when training the dialogue generation language model, the dialogue generation language model can be trained according to the object category of each population package, the historical behavior data of the sample objects, and the historical advertisements of interest to the sample objects.

[0039] Among the possible approaches, the dialogue generation language model can be trained in the following ways: Acquire clustering information of sample objects, where the clustering information includes multiple object categories and sample object information corresponding to each object category; For each object category, perform the following steps: Obtain historical behavior data and historical advertisements of interest for sample objects in the object category, take each sample object in the object category as a traversal object, and for the first traversal object, generate a first sample prompt text according to the object category, the historical behavior data of the traversal object, and the historical advertisements of interest for all traversal objects in the object category; input the first sample prompt text into the dialogue generation language model to obtain the score of the traversal object for each historical advertisement of interest; eliminate the historical advertisements of interest with a score less than a preset score, generate a new first sample prompt text according to the historical behavior data of the next traversal object, the object category, and the remaining historical advertisements of interest, and repeat the above steps until all traversal objects are traversed.

[0040] For example, the sample objects may be divided based on their interests, thereby obtaining object category A and object category B. For object category A, three sample objects may be included, denoted as object A1, object A2, and object A3. Among them, the historical advertisements of interest to object A1 may include: advertisement A11 and advertisement A12, and the historical behavior data may include: behavior data A11, behavior data A12, and behavior data A13. The historical advertisements of interest to object A2 may include: advertisement A21, and the historical behavior data may include: behavior data A21, behavior data A22, and behavior data A23. The historical advertisements of interest to object A3 may include: advertisement A31, advertisement A32, and advertisement A33, and the historical behavior data may include: behavior data A31 and behavior data A32.

[0041] Therefore, when training the dialogue generation language model based on data related to object category A, a sample prompt text can be first generated based on object category A, advertisement A11, advertisement A12, advertisement A21, advertisement A31, advertisement A32, advertisement A33, behavior data A11, behavior data A12, and behavior data A13. For example, the sample prompt text is: Please generate the scores of object A1 for advertisement A11, advertisement A12, advertisement A21, advertisement A31, advertisement A32, and advertisement A33 based on "category: object category A; interested advertisements: advertisement A11, advertisement A12, advertisement A21, advertisement A31, advertisement A32, advertisement A33; behavior data: behavior data A11, behavior data A12, and behavior data A13". Then, the sample prompt text is input into the dialogue generation language model to obtain the scores of object A1 for advertisement A11, advertisement A12, advertisement A21, advertisement A31, advertisement A32, and advertisement A33. Then, compare the score of object A1 for each historical advertisement of interest with the preset score. If the scores of object A1 for advertisement A11, advertisement A12, advertisement A21, advertisement A31, advertisement A32 and advertisement A33 are not less than the preset score, generate sample prompt text according to object category A, advertisement A11, advertisement A12, advertisement A21, advertisement A31, advertisement A32, advertisement A33, behavior data A21, behavior data A22 and behavior data A23, and repeat the above steps; if the score of object A1 for a certain historical advertisement of interest is less than the preset score, for example, the score of object A1 for advertisement A21 is less than the preset score, then eliminate advertisement A21, and generate sample prompt text according to object category A, advertisement A11, advertisement A12, advertisement A31, advertisement A32, advertisement A33, behavior data A21, behavior data A22 and behavior data A23, and repeat the above steps.

[0042] The format of the sample prompt text generated each time may be fixed or variable, and the embodiment of the present disclosure does not impose any limitation on this.

[0043] When the format of the sample prompt text varies, it is sufficient as long as the sample prompt text includes the object category, the historical behavior data of the traversed object, the historical advertisements of interest for all traversed objects in the object category, and text for prompting the dialogue generation language model to output the score of each historical advertisement of interest based on the above information.

[0044] When the format of the sample prompt text is fixed, a prompt text template can be pre-set. After obtaining the object category, the historical behavior data of the traversed object, and the historical advertisements of interest of all traversed objects in the object category, the first sample prompt text can be quickly generated by filling the object category, the historical behavior data of the traversed object, and the historical advertisements of interest of all traversed objects in the object category into the corresponding positions of the prompt text template.

[0045] The prompt text template may be set according to actual conditions, and the embodiments of the present disclosure do not impose any restrictions on this. For example, the prompt text template may be set as follows: Please generate a score of object: XX for interested advertisement: XXX based on "object category: XX; interested advertisement: XXX; behavior data: XXXX".

[0046] Therefore, after obtaining the object category, the historical behavior data of the traversed object, and the historical advertisements of interest of all traversed objects in the object category, sample prompt text can be quickly generated based on the prompt text template, thereby improving the training speed of the dialogue generation language model.

[0047] Among the possible approaches, the dialogue generation language model can be trained in the following ways: Obtaining object grouping information, where the object grouping information includes multiple object categories and sample object information corresponding to each object category; For each object category, perform the following steps: Obtain historical behavior data and historical advertisements of interest for sample objects in the object category, take each historical advertisement of interest in the object category as a traversal object, and for the first traversal object, generate a second sample prompt text according to the object category, the traversal object, and the historical behavior data of all sample objects in the object category; input the second sample prompt text into the dialogue generation language model to obtain the score of each sample object for the traversal object, and remove the traversal object when the average score of the traversal object is less than a preset value; generate a new second sample prompt text according to the next traversal object, the object category, and the historical behavior data of all sample objects in the object category, and repeat the above steps until all traversal objects are traversed.

[0048] For example, objects may be divided based on object behavior patterns to obtain object categories C and object categories D. For object category C, three sample objects may be included, denoted as object C1, object C2, and object C3. Among them, the historical advertisements of interest to object C1 may include: advertisement C11 and advertisement C12, and the historical behavior data may include: behavior data C11 and behavior data C12. The historical advertisements of interest to object C2 may include: advertisement C21, and the historical behavior data may include: behavior data C21, behavior data C22, and behavior data C23. The historical advertisements of interest to object C3 may include: advertisement C31, advertisement C32, and advertisement C33, and the historical behavior data may include: behavior data C31 and behavior data C32.

[0049] When training the dialogue generation language model based on data related to object category C, a sample prompt text may be first generated based on object category C, advertisement C11, behavior data C11, behavior data C12, behavior data C21, behavior data C22, behavior data C23, behavior data C31, and behavior data C32. For example, the sample prompt text is: Please generate the scores of object C1, object C2, and object C3 for advertisement C11 based on "category: object category C; interested advertisement: advertisement C11; behavior data: behavior data C11, behavior data C12, behavior data C21, behavior data C22, behavior data C23, behavior data C31, and behavior data C32". Then, the sample prompt text is input into the dialogue generation language model to obtain the scores of object C1, object C2, and object C3 for advertisement C11. Then, the average score of the scores of object C1, object C2, and object C3 for advertisement C11 is calculated, and the size of the average score is compared with the preset value. If the mean score is not less than the preset value, generate sample prompt text based on object category C, advertisement C12, behavior data C11, behavior data C12, behavior data C21, behavior data C22, behavior data C23, behavior data C31 and behavior data C32, and repeat the above steps; if the mean score is less than the preset value, eliminate advertisement C11, and generate sample prompt text based on object category C, advertisement C12, behavior data C11, behavior data C12, behavior data C21, behavior data C22, behavior data C23, behavior data C31 and behavior data C32, and repeat the above steps.

[0050] After the training of the dialogue generation language model is completed, when the first behavior data of the target object and the dialogue generation language model are used to generate the target push advertisement to be pushed to the target object, a prompt text can be first generated according to the first behavior data of the target object, and then the prompt text can be input into the dialogue generation language model to obtain the target push advertisement to be pushed to the target object.

[0051] For example, the first behavior data of the target object is: the target object's recent stock trading records and browsing history. Then the prompt text generated based on the first behavior data can be: Based on your recent stock trading and browsing history, the following stock investment products are recommended for you. Finally, the prompt text is input into the dialogue to generate a language model to obtain personalized stock investment products for the target object.

[0052] S202: Determine that push data of the target push information has been pushed through the target push path.

[0053] The target push path may be a multimedia platform or an information display page in the multimedia platform. Push data may include push times, conversion rate, click-through rate, push cost and / or push revenue, etc., which may be determined according to actual conditions, and the embodiments of the present disclosure do not impose any restrictions on this.

[0054] S203: Determine, according to the first behavior data and the push data, push reference information for pushing target push information to the target object through the target push path.

[0055] The push reference information may be used to characterize the predicted push result of pushing the target push information to the target object through the target push path, and the predicted push result may be a predicted conversion rate, a predicted click-through rate, or a predicted push revenue, etc. It may also be used to characterize the push time of pushing the target push information to the target object through the target push path, and it may also be used to characterize the push frequency of pushing the target push information to the target object through the target push path, etc., which is determined according to actual conditions, and the embodiments of the present disclosure do not impose any restrictions on this.

[0056] In a possible manner, the push reference information may include first push reference information and second push reference information; accordingly, determining the push reference information for pushing the target push information to the target object through the target push path according to the first behavior data and the push data may include: Determine first push reference information based on the first behavior data and the push data; and determine second push reference information based on the push data; wherein the first push reference information is used to represent predicted conversion information of pushing target push information to the target object via the target push path; and the second push reference information is used to represent push resource information reserved by the target push path for pushing the target push information to the target object.

[0057] For example, the first behavior data and the push data can be used as inputs of the machine learning model to obtain the first push reference information. For example, the total push resources used to push the target push information to the target object through the target push path can be obtained in advance, and then the second push reference information is determined based on the total push resources and the push data (for example, the number of pushes and the push cost).

[0058] The total push resource in this embodiment may be the total number of pushes or the total push cost, etc.

[0059] The machine learning model in this embodiment may be an existing model in the related art, or may be a model that is improved upon an existing model in the related art, and the disclosed embodiment does not impose any restrictions on this. For example, the machine learning model in this embodiment may be a large language model in the related art, such as a ChatGLM model.

[0060] After determining the model structure of the machine learning model, in order to use the machine learning model to obtain the first push reference information according to the first behavior data and the push data, training data can be first constructed, and the machine learning model can be trained based on the training data.

[0061] For example, in an advertisement push scenario, historical advertisement push data (e.g., historical conversion data, historical click data, historical download data, etc.) of different advertisement positions in sample media and historical behavior data of sample objects may be collected first. The collected historical data is then cleaned, deduplicated, and feature extracted, so that a machine learning model is trained based on the processed historical data to obtain a machine learning model for obtaining first push reference information based on the first behavior data and the push data.

[0062] After obtaining the machine learning model for obtaining the first push reference information based on the first behavior data and the push data, you can first determine the prompt text used by the machine learning model to output the first push reference information based on the first behavior data and the push data; then input the prompt text into the machine learning model to obtain the first push reference information.

[0063] The format of the prompt text generated each time can be fixed or variable, and the disclosed embodiment does not impose any restrictions on this. When the format of the prompt text is fixed, a prompt text template can be pre-set, and after obtaining the first behavior data and the push data, the prompt text can be quickly generated by filling the first behavior data and the push data into the corresponding positions of the prompt text template.

[0064] In this embodiment, by using the second behavior data and the push data as inputs to the machine learning model, the first push reference information can be automatically output without human intervention, thereby improving the efficiency of information push.

[0065] S204: Push the target push information to the target object according to the push reference information and the target push path.

[0066] In a possible manner, there may be multiple target push paths. Accordingly, according to the push reference information and the target push path, the target push information is pushed to the target object, which may include: According to the push reference information, a first target push path is determined from multiple target push paths; and the target push information is pushed to the target object through the first target push path.

[0067] For example, when push reference information is used to characterize the predicted push results of pushing target push information to a target object through a target push path, a first target push path with the best predicted push result can be determined from multiple target push paths based on the push reference information, so that when the target push information is pushed to the target object through the first target push path, the push effect of the target push information can be improved, for example, the conversion rate of the target push information, the click-through rate of the target push information, or the push revenue of the target push information can be improved.

[0068] For example, when push reference information is used to characterize the push time of pushing target push information to a target object via a target push path, a first target push path for pushing the target push information at the current time can be determined from multiple target push paths based on the push reference information and the current time, so that when the target push information is pushed to the target object via the first target push path, the push effect of the target push information can be improved, for example, the conversion rate of the target push information, the click-through rate of the target push information, or the push revenue of the target push information can be improved.

[0069] In a possible manner, there may be multiple target push paths, and the push reference information may include first push reference information and second push reference information, wherein the first push reference information is used to represent predicted conversion information of pushing the target push information to the target object through the target push path; the second push reference information is used to represent push resource information reserved by the target push path for pushing the target push information to the target object. Accordingly, according to the push reference information and the target push path, pushing the target push information to the target object may include: According to the second push reference information, multiple second target push paths are determined from multiple target recommended paths; according to the first push reference information, a third target push path is determined from multiple second target push paths; and the target push information is pushed to the target object through the third target push path.

[0070] For example, when the predicted conversion information is the predicted conversion rate, and the push resource information is the number of deliveries that can also be used to push the target push information to the target object through the target push path, you can first screen out multiple second target push paths with a number of deliveries greater than 0 based on the second push reference information corresponding to each target recommended path, and then screen out the third target push path based on the predicted conversion rate corresponding to each second target push path, where the third target push path has the highest predicted conversion rate, and finally push the target push information to the target object through the third target push path.

[0071] In this embodiment, multiple second target push paths are first determined from multiple target recommended paths based on the first push reference information, thereby avoiding pushing target push information to the target object through invalid target push paths, thereby improving the accuracy of information push. Then, a third target push path for pushing target push information to the target object is determined from multiple target recommended paths based on the second push reference information, thereby improving the conversion rate of target push information and thus improving the information push effect.

[0072] Through the above technical solution, the target push information to be pushed to the target object can be determined based on the behavior data of the target object, and the push data that has pushed the target push information through the target push path can be determined, and then the push reference information can be determined based on the behavior data and the push data, and the target push information can be pushed to the target object based on the push reference information and the target push path. Since the target push information is determined based on the behavior data of the target object, the target push information can be matched with the interests of the target object based on the behavior data of the target object, thereby realizing personalized information push and improving the information push effect. In addition, since the push data refers to the data that has pushed the target push information through the target push path, that is, the push data can represent the historical push data of the target push information under the target push path, when the push reference information is determined based on the push data and the behavior data, and the target push information is pushed to the target object based on the push reference information, both the behavior data of the target object and the historical push data of the target push information can be taken into account at the same time, thereby further improving the information push effect.

[0073] It should be understood that, due to the changeable behavior of the target object, after pushing the target push information to the target object, there is a problem that the pushed target push information does not match the current behavior data of the target object, resulting in unsatisfactory information push effect. Therefore, in order to further improve the information push effect and information push accuracy, after pushing the target push information to the target object, the second behavior data of the target object and / or the real conversion information of the target push information can be obtained, so as to dynamically adjust the push information pushed to the target object based on the second behavior data and / or the real conversion information, and / or adjust the push path of the push information to the target object. That is to say, in a possible way, after pushing the target push information to the target object, the information push method can also include: Determine the second behavior data of the target object and / or the real conversion information of the target push information; adjust the push information pushed to the target object and / or adjust the push path of the push information to the target object according to the second behavior data and / or the real conversion information.

[0074] The second behavior data may be the real-time behavior data of the target object after the target push information is pushed to the target object, or the behavior data of the target object within a preset duration or a preset time period after the target push information is pushed to the target object. For example, if the moment when the target push information is pushed to the target object is moment T, then the second behavior data may be the behavior data of the target object between moment T and moment T+t, where T and t are positive integers.

[0075] Among them, the real conversion data can be used to characterize the proportion of the target object that performs the expected behavior of the information publisher after the target push information reaches the target object. Among them, the expected behavior of the information publisher can be determined according to the actual situation, and the embodiments of the present disclosure do not impose any restrictions on this. For example, in an advertising push scenario, the information publisher may be an advertiser, the information may be an advertisement published by the advertiser, and the expected behavior of the information publisher may be that the advertisement is clicked by the object after being displayed. Accordingly, the real conversion data may be the click-through conversion rate. The expected behavior of the information publisher may also be that the advertisement is purchased by the object after being displayed. Accordingly, the real conversion data may be the purchase conversion rate. The expected behavior of the information publisher may also be that the advertisement is downloaded or registered by the object after being displayed. Accordingly, the real conversion data may be the registration conversion rate or the download conversion rate, etc.

[0076] Through the above technical solution, after the target push information is pushed to the target object, the second behavior data of the target object and / or the real conversion information of the target push information can be obtained, and based on the second behavior data and / or the real conversion information, the push information pushed to the target object and / or the push path of the push information to the target object can be adjusted. In this way, the target information and the push path can be dynamically adjusted, so that the target information and the push path match the behavior data of the target object, further improving the information push effect and the accuracy of the information push.

[0077] In a possible manner, the target push information may include first target push information and second target push information; accordingly, pushing the target push information to the target object may include: Pushing the first target push information to the target object; Accordingly, adjusting the push information pushed to the target object according to the second behavior data and the real conversion information may include: Based on the second behavior data and the actual conversion information, the third push reference information is determined, and the third push reference information is used to represent the predicted conversion information of continuing to push the first target recommended information to the target object through the target push path; if the third push reference information does not meet the preset conditions, the second target push information is pushed to the target object through the target push path.

[0078] The preset condition may be set according to actual conditions, and the present disclosure does not impose any restrictions on this. For example, the preset condition may be that the predicted conversion rate in the predicted conversion information is less than a threshold conversion rate.

[0079] The second target push information may include one or more. When the second target push information includes more than one, any one of the second target push information may be pushed to the target object through the target push path, or the predicted conversion information of the target push information pushed to the target object through the target push path may be determined for each second target push information based on the second behavior data and the push data. Then, the conversion information is predicted based on the multiple second target push information, and the second target push information pushed to the target object through the target push path is screened out from the multiple second target push information.

[0080] For example, the second target push information includes target push information a, target push information b and target push information c, and the target push information b is pushed to the target object through the target push path, and the real conversion information is the real conversion rate, the predicted conversion information is the predicted conversion rate, and the push data is the historical conversion rate. When the predicted conversion rate obtained according to the second behavior data and the real conversion rate is less than the threshold conversion rate, the predicted conversion rate of the target push information b can be obtained according to the second behavior data and the historical conversion rate of the target push information b, and the predicted conversion rate of the target push information c can be obtained according to the second behavior data and the historical conversion rate of the target push information c. Then, the historical conversion rate of the target push information b and the historical conversion rate of the target push information c are compared. If the historical conversion rate of the target push information b is greater than the historical conversion rate of the target push information c, the target push information b is pushed to the target object through the target push path; if the historical conversion rate of the target push information c is greater than the historical conversion rate of the target push information b, the target push information c is pushed to the target object through the target push path.

[0081] Among them, the method of determining the third push reference information according to the second behavior data and the real conversion information can refer to the method of determining the first push reference information according to the first behavior data and the push data, which will not be repeated here.

[0082] Through the above method, after the target push information is pushed to the target object, the third push reference information can be determined based on the second behavior data and the real conversion information, and when the third push reference information does not meet the preset conditions, the second target push information is pushed to the target object through the target push path. In this way, the push information pushed to the target object can be dynamically adjusted to further improve the information push effect, for example, to improve the conversion rate, click-through rate or delivery revenue of the push information.

[0083] In a possible manner, the target push path may include a fourth target push path and a fifth target push path; accordingly, adjusting the push path for pushing information to the target object according to the second behavior data may include: If the target push information is pushed through the fourth target push path, and the second behavior data represents that the push success rate of the target push information is less than or equal to the preset threshold, the target push information is pushed to the target object through the fifth target push path; the push success rate corresponding to the fifth target push path is greater than the preset threshold.

[0084] Among them, the push success rate can be used to represent the proportion of the target object performing a preset behavior after the target push information reaches the target object. Among them, the preset behavior can be the behavior of clicking on the target push information, the behavior of browsing the target push information for a preset time, or the behavior of sharing the target push information. Correspondingly, the push success rate can be the click rate, browsing rate, sharing rate, etc.

[0085] Among them, the fifth target push path may include one or multiple. When the fifth target push path includes multiple, the target push information can be pushed to the target object through any fifth target push path, or through the fifth target push path with the highest push success rate. A number can also be set for each target push path in advance, and then according to the number of the fourth target push path, the target number that is after the number of the fourth target push path and closest to the number of the fourth target push path is determined among multiple fifth target push paths, and finally the target push information is pushed to the target object through the fifth target push path corresponding to the target number.

[0086] Among them, the push success rate corresponding to the fifth target push path can be a historical push success rate, or it can be a real-time push success rate determined based on the second behavior data, and the embodiments of the present disclosure do not impose any restrictions on this. For example, if the target push information is pushed to the target user through the fourth target push path, and the fifth target push path does not push other information to the target user, then the push success rate corresponding to the fifth target push path can be the historical push success rate. If the target push information is pushed to the target user through the fourth target push path, and the fifth target push path pushes other information to the target user, then the push success rate corresponding to the fifth target push path can be the real-time push success rate determined based on the second behavior data.

[0087] The preset threshold value may be determined according to actual conditions, and the embodiments of the present disclosure do not impose any restrictions on this. For example, the preset threshold value may be set to 65% or 80%.

[0088] Through the above scheme, after pushing the target push information to the target object, the push success rate of the target push information can be determined, and when the push success rate is lower than the preset threshold, the target push information can be pushed to the fifth push path with a push success rate higher than the preset threshold, thereby improving the push success rate of the target push information.

[0089] It should be understood that the above adjustment method is only for illustrative purposes and does not constitute a limitation to the present solution. In a possible manner, the push amount of the target push information to the target object can also be adjusted according to the real conversion data. For example, when the real conversion data indicates that the conversion rate of the push information in the target time period is higher than the preset conversion rate threshold, the delivery amount of the push information in the target time period is increased; when the real conversion data indicates that the conversion rate of the target push information in the target time period is lower than or equal to the preset conversion rate threshold, the delivery amount of the target push information in the target time period is reduced.

[0090] For example, when the target push information is financial information, after the financial information is pushed to the target object, the conversion rate of the financial information in different time periods can be obtained. If the conversion rate of the financial information is higher than the preset conversion rate threshold between 10 pm and 2 am, the amount of financial information delivered in this time period can be automatically increased to obtain more conversions. If the conversion rate of financial information is lower than the preset conversion rate threshold between 10 am and 6 pm, the amount of conversion delivered in this time period can be automatically reduced to reduce the cost of information push.

[0091] The preset conversion rate threshold can be determined according to actual conditions, and the embodiments of the present disclosure do not impose any restrictions on this. For example, the preset conversion rate threshold can be set to 55% or 60%.

[0092] Through the above scheme, the delivery volume of the target push information in the target time period can be dynamically adjusted according to the conversion data of the target push information in different time periods. As a result, the delivery volume of the target push information can be increased in the time period with high conversion data to improve the conversion data of the target push information; the delivery volume of the target push information can be reduced in the time period with low conversion data to reduce the push cost of the target push information.

[0093] To facilitate understanding of the information push method disclosed herein, the information push method disclosed herein is described below by taking financial product advertisement push as an example. In this embodiment, the target object is a user, the push information is an advertisement, and the push path is an advertisement position.

[0094] For example, Figure 3 and Figure 4As shown, financial companies can collect user behavior data (these behavior data may include user stock trading history, fund investment status, loan application records, and website browsing behaviors) and interested advertisements through websites and mobile applications with user authorization, as well as collect advertising placement data of the financial company for advertising placement in different advertising positions. Then, the behavior data, interested advertisements and / or advertising placement data are preprocessed, such as removing duplicate data and filling missing values, and converted into a format that the model can understand, such as text-based user behavior records or advertising placement records. Then, the converted behavior data and interested advertisements are used to fine-tune the advertising push model integrated in the advertising push system to generate personalized recommendations for the financial field. The converted behavior data and advertising placement data are used to fine-tune the push reference information prediction model integrated in the advertising push system to generate push reference information for the financial field. Therefore, after obtaining the user's real-time behavior data, personalized recommendations can be made based on the real-time behavior data, the advertising push model, and the push reference information prediction model. For example, for users who are interested in stock investment, advertisements for stock trading software can be pushed; for users who have loan needs, advertisements for loan products can be pushed.

[0095] Specifically, when a user browses financial-related information through a website or mobile application. For example, when a user browses a stock page, the advertising push system can receive the user's behavioral data, such as browsing history and search keywords. The behavioral data is transmitted to the fine-tuned advertising push model, and the advertising push model generates personalized financial product recommendations based on the user's behavioral data. For example, the advertising push system obtains the user's behavioral data as "recent stock trading records and browsing history", and then generates a prompt text based on the behavioral data "Based on your recent stock trading and browsing history, the following stock investment products are recommended for you", and inputs the prompt text into the fine-tuned advertising push model to obtain the pushed investment products, such as investment product a and investment product b.

[0096] After obtaining investment product a and investment product b, the advertising push system can obtain push data a of investment product a in ad slot 1 and push data b of investment product b in ad slot 1, and generate prompt text a according to push data a and user behavior data, and input prompt text a into the push reference information prediction model, thereby obtaining push reference information for pushing investment product a to the target object through ad slot 1, such as predicted delivery income a. And generate prompt text b according to push data b and user behavior data, and input prompt text b into the push reference information prediction model, thereby obtaining push reference information for pushing investment product b to the target object through ad slot 1, such as predicted delivery income b. If the predicted delivery income b is greater than the predicted delivery income a, investment product b is delivered to the user through ad slot 1; if the predicted delivery income a is greater than the predicted delivery income b, investment product a is delivered to the user through ad slot 1.

[0097] After pushing investment product a or investment product b, the advertising push system can continue to collect user behavior data and real conversion information of investment product a or investment product b, such as conversion rate, to dynamically adjust the delivery of investment product a and / or investment product b. For example, if the advertising push system finds based on user behavior data and / or real conversion data of investment product a that the click-through rate of investment product a is low when investment product a is delivered through ad slot 1, then investment product a can be automatically switched to an ad slot with a higher click-through rate. For example, investment product a is switched to ad slot 2 for delivery, so as to improve the conversion data of investment product a.

[0098] Through the above method, the user's first behavior data and historical behavior data can be combined to push the user's interested advertisements to the user in real time, thereby realizing personalized advertisement recommendations and improving the effect of advertisement push. In addition, after the advertisement is pushed, the user's second behavior data and / or the real conversion data of the pushed advertisement can also be obtained, and based on the second behavior data and / or the real conversion data, the advertisement delivery situation can be dynamically adjusted, such as adjusting the push time, push volume, push advertisements and / or advertisement positions, etc. In this way, dynamic adjustment of advertisements can be realized, so that the pushed advertisements match the user's behavior data, further improving the effect and accuracy of advertisement push.

[0099] Based on the same concept, the embodiment of the present disclosure also provides an information push device, which can be part or all of the electronic device through software, hardware, or a combination of software and hardware. Figure 5 As shown, the information push device 500 may include: A first determining module 501 is used to determine target push information to be pushed to the target object according to the first behavior data of the target object; A second determining module 502 is used to determine the push data of the target push information that has been pushed through the target push path; A third determining module 503 is used to determine, according to the first behavior data and the push data, push reference information for pushing the target push information to the target object through the target push path; The push module 504 is configured to push the target push information to the target object according to the push reference information and the target push path.

[0100] Through the above-mentioned information push device 500, the target push information to be pushed to the target object can be determined based on the behavior data of the target object, and the push data that has pushed the target push information through the target push path can be determined, and then the push reference information can be determined based on the behavior data and the push data, and the target push information can be pushed to the target object based on the push reference information and the target push path. Since the target push information is determined based on the behavior data of the target object, the target push information can be matched with the interests of the target object based on the behavior data of the target object, thereby realizing personalized information push and improving the information push effect. In addition, since the push data refers to the data that has pushed the target push information through the target push path, that is, the push data can represent the historical push data of the target push information under the target push path, when the push reference information is determined based on the push data and the behavior data, and the target push information is pushed to the target object based on the push reference information, the behavior data of the target object and the historical push data of the target push information can be taken into account at the same time, thereby further improving the information push effect.

[0101] In a possible implementation, the push reference information may include first push reference information and second push reference information. Accordingly, the third determination module 503 may be used to: determine the first push reference information based on the first behavior data and the push data; and determine the second push reference information based on the push data; wherein the first push reference information is used to characterize the predicted conversion information of pushing the target push information to the target object through the target push path; and the second push reference information is used to characterize the push resource information reserved by the target push path for pushing the target push information to the target object.

[0102] In a possible implementation, there may be multiple target push paths, and accordingly, the push module 504 may include: A first determining unit, configured to determine a first target push path from a plurality of target push paths according to the push reference information; The first pushing unit is used to push target pushing information to the target object through a first target pushing path.

[0103] In a possible implementation, there may be multiple target push paths, and accordingly, the push module 504 may include: A second determining unit, configured to determine a plurality of second target push paths from the plurality of target push paths according to the second push reference information; A third determining unit, configured to determine a third target push path from a plurality of second target push paths according to the first push reference information; The second pushing unit is used to push the target pushing information to the target object through the third target pushing path.

[0104] In a possible implementation manner, the information push device 500 may further include: A fourth determination module, used to determine the second behavior data of the target object and / or the real conversion information of the target push information; The adjustment module is used to adjust the push information pushed to the target object and / or adjust the push path of the push information to the target object according to the second behavior data and / or the real conversion information.

[0105] In a possible implementation, the target push information may include first target push information and second target push information. Accordingly, the push module 504 may be used to push the first target push information to the target object. Accordingly, the adjustment module may include: A fourth determining unit, configured to determine third push reference information according to the second behavior data and the actual conversion information, wherein the third push reference information is used to represent predicted conversion information for continuing to push the first target push information to the target object through the target push path; The third pushing unit is configured to push the second target pushing information to the target object through the target pushing path if the third push reference information does not meet the preset condition.

[0106] In a possible implementation, the target push path may include a fourth target push path and a fifth target push path. Accordingly, the adjustment module may be used for: if the target push information is pushed through the fourth target push path, and the second behavior data represents that the push success rate of the target push information is less than or equal to a preset threshold, then the target push information is pushed to the target object through the fifth target push path; the push success rate corresponding to the fifth target push path is greater than the preset threshold.

[0107] Regarding the information push device 500 in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0108] Based on the same concept, the embodiment of the present disclosure further provides an electronic device, including: a storage device having a computer program stored thereon; A processing device is used to execute the computer program in the storage device to perform the steps of any of the above-mentioned information push methods.

[0109] Based on the same concept, an embodiment of the present disclosure further provides a computer-readable medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned information push methods are implemented.

[0110] Based on the same concept, an embodiment of the present disclosure further provides a computer program product, including a computer program, which implements the steps of any of the above-mentioned information push methods when executed by a processor.

[0111] It should be understood that the above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) to form a technical solution.

[0112] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0113] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims. Regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Claims

1. An information push method, characterized in that: The information pushing method comprises: Determining target push information to be pushed to the target object according to the first behavior data of the target object; Determining that push data of the target push information has been pushed through the target push path; Determine, according to the first behavior data and the push data, push reference information for pushing the target push information to the target object through the target push path; The target push information is pushed to the target object according to the push reference information and the target push path.

2. The information push method according to claim 1, characterized in that: The push reference information includes first push reference information and second push reference information; the push reference information for pushing the target push information to the target object through the target push path is determined according to the first behavior data and the push data, including: determining the first push reference information according to the first behavior data and the push data; and determining the second push reference information according to the push data; The first push reference information is used to represent predicted conversion information of pushing the target push information to the target object through the target push path; The second push reference information is used to represent push resource information reserved by the target push path for pushing the target push information to the target object.

3. The information push method according to claim 1 or 2, characterized in that: There are multiple target push paths, and pushing the target push information to the target object according to the push reference information and the target push path includes: Determining a first target push path from a plurality of target push paths according to the push reference information; The target push information is pushed to the target object through the first target push path.

4. The information push method according to claim 2, characterized in that: There are multiple target push paths, and pushing the target push information to the target object according to the push reference information and the target push path includes: determining a plurality of second target push paths from a plurality of target push paths according to the second push reference information; determining a third target push path from the plurality of second target push paths according to the first push reference information; The target push information is pushed to the target object through the third target push path.

5. The information push method according to claim 1 or 2, characterized in that: After pushing the target push information to the target object, the method further includes: Determine the second behavior data of the target object and / or the real conversion information of the target push information; According to the second behavior data and / or the real conversion information, the push information pushed to the target object is adjusted, and / or the push path of the push information to the target object is adjusted.

6. The information push method according to claim 5, characterized in that: The target push information includes first target push information and second target push information; The pushing the target push information to the target object includes: pushing the first target push information to the target object; The adjusting the push information pushed to the target object according to the second behavior data and the real conversion information includes: determining third push reference information according to the second behavior data and the actual conversion information, wherein the third push reference information is used to represent predicted conversion information of continuing to push the first target push information to the target object through the target push path; If the third push reference information does not meet the preset condition, the second target push information is pushed to the target object through the target push path.

7. The information push method according to claim 5, characterized in that: The target push path includes a fourth target push path and a fifth target push path; and adjusting the push path for pushing information to the target object according to the second behavior data includes: If the target push information is pushed through the fourth target push path, and the second behavior data indicates that the push success rate of the target push information is less than or equal to a preset threshold, then the target push information is pushed to the target object through the fifth target push path; the push success rate corresponding to the fifth target push path is greater than the preset threshold.

8. An information push device, characterized in that: The information pushing device comprises: A first determining module, configured to determine target push information to be pushed to the target object according to first behavior data of the target object; A second determining module, used to determine the push data of the target push information that has been pushed through the target push path; A third determining module, configured to determine, based on the first behavior data and the push data, push reference information for pushing the target push information to the target object through the target push path; A push module is used to push the target push information to the target object according to the push reference information and the target push path.

9. An electronic device, characterized in that: include: a storage device having a computer program stored thereon; A processing device, configured to execute the computer program in the storage device to implement the steps of the method according to any one of claims 1 to 7.

10. A computer readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processing device, the steps of the method according to any one of claims 1 to 7 are implemented.

11. 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 7 are implemented.

Images