CN115935054A

CN115935054A - Information pushing method and device, electronic equipment and storage medium

Info

Publication number: CN115935054A
Application number: CN202211113943.6A
Authority: CN
Inventors: 陈嘉
Original assignee: Beijing Wisdom Rongsheng Technology Co ltd
Current assignee: Beijing Wisdom Rongsheng Technology Co ltd
Priority date: 2022-09-14
Filing date: 2022-09-14
Publication date: 2023-04-07

Abstract

The invention discloses an information pushing method and device, electronic equipment and a storage medium, and relates to the field of artificial intelligence. The implementation scheme comprises the following steps: determining candidate users and user characteristics of each candidate user based on the atomic data of the operator platform; identifying the user characteristics of each candidate user by using a pre-trained transformation prediction model, and determining the probability of each candidate user responding to the push information; and selecting at least one target user from the candidate users according to the probability of each candidate user responding to the pushed information, and indicating the operator platform to push the information to the target user. In the scheme of the invention, the related information is pushed only to the target user which is likely to respond to the pushed information, and compared with the method of randomly pushing the related information to the user, the probability of converting the pushed information by the user is improved, thereby ensuring the efficiency of acquiring effective clients by enterprises.

Description

Information pushing method and device, electronic equipment and storage medium

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to an information pushing method and device, electronic equipment and a storage medium.

Background

Currently, for enterprises, related marketing information needs to be pushed to users in order to obtain customers, and customers are mainly obtained by issuing a flyer, holding offline activities and randomly sending related marketing information. But this approach is inefficient.

Disclosure of Invention

The invention provides an information pushing method, an information pushing device, electronic equipment and a storage medium, and aims to achieve the effect of improving the passenger obtaining efficiency.

According to an aspect of the present invention, there is provided an information pushing method, including:

determining candidate users and user characteristics of each candidate user based on the atomic data of the operator platform;

identifying the user characteristics of each candidate user by using a pre-trained transformation prediction model, and determining the probability of each candidate user responding to the push information;

and selecting at least one target user from the candidate users according to the probability of each candidate user responding to the push information, and indicating the operator platform to push the information of the target user.

According to another aspect of the present invention, there is provided an information pushing apparatus including:

the user and feature mining module is used for determining candidate users and the user features of each candidate user based on the atomic data of the operator platform;

the prediction module is used for identifying the user characteristics of each candidate user by utilizing a pre-trained conversion prediction model and determining the probability of each candidate user responding to the push information;

and the information pushing module is used for selecting at least one target user from the candidate users according to the probability of response of each candidate user to the pushed information and instructing the operator platform to push the information to the target user.

According to another aspect of the present invention, there is provided an electronic apparatus including:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the information pushing method according to the embodiment of the present invention.

According to another aspect of the present invention, a computer-readable storage medium is provided, and computer instructions are stored in the computer-readable storage medium, and when the computer instructions are executed, a processor implements the information pushing method according to the embodiment of the present invention.

According to the technical scheme of the embodiment of the invention, the related information is pushed only to the target user which is likely to respond to the pushed information, and compared with the method of pushing the information to the user randomly, the probability of converting the pushed information by the user is improved, so that the efficiency of acquiring effective clients by enterprises is ensured.

It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present invention, nor do they necessarily limit the scope of the invention. Other features of the present invention will become apparent from the following description.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a schematic flowchart of an information push method according to an embodiment of the present invention;

fig. 2 is a flowchart illustrating an information pushing method according to a second embodiment of the present invention;

fig. 3 is a schematic flowchart of an information pushing method according to a third embodiment of the present invention;

fig. 4 is a schematic structural diagram of an information pushing apparatus according to a fourth embodiment of the present invention;

fig. 5 is a schematic structural diagram of an electronic device of an information push method according to an embodiment of the present invention.

Detailed Description

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Example one

Fig. 1 is a flowchart of an information pushing method according to an embodiment of the present invention, where the embodiment is applicable to a scenario in which an enterprise obtains customers by pushing marketing information to a user, and the method may be executed by an information pushing apparatus, where the information pushing apparatus may be implemented in hardware and/or software, and the information pushing apparatus may be configured in an electronic device, for example, in a server device.

As shown in fig. 1, the information pushing method flows as follows:

s101, determining candidate users and user characteristics of each candidate user based on the atomic data of the operator platform.

In this embodiment, the atomic data includes desensitized user call data, short message data, and user internet behavior data. On the basis, candidate users and the user characteristics of each candidate user can be determined by means of aggregation and cleaning of the atomic data. In specific implementation, aiming at each user in the operator platform full-volume users, according to different preset dimensions, user characteristics of each user are mined from user call data, short message data and user internet behavior data. Optionally, for any user, feature mining is performed from the call data according to the dimensions of the platform user identifier of the user on the operator platform (i.e., the ID of the user on the platform), the platform user identifiers of the calling and called users, the platform user identifiers of the opposite user, the number of calls, the duration, and the like, so as to obtain a plurality of user features which are multi-dimensional and related to the call, for example, the duration of each call can be used as the user feature; similarly, the short message data can be mined according to the dimensions of the platform user identification, the sending mechanism, the times and the like of the user to obtain multi-dimensional user characteristics related to the short message, for example, different sending mechanisms are used as the user characteristics related to the user; from the user internet behavior data, mining can be performed according to dimensionalities such as platform user identification, website or webpage domain names, and the like, so as to obtain multi-dimensionality user characteristics related to the internet behavior. Thus, the user characteristics of each user in the operator platform can be determined through the operation.

It should be noted that the acquisition, storage, application and the like of the personal information of the user related to the invention are authorized by the user and conform to the regulations of related laws and regulations without violating the public order and good customs.

Further, because the atomic data of the operator platform is very cluttered, when the user characteristics are mined according to the above manner, there is a condition of characteristic loss, for example, some users have no call information, short message list loss, or no internet behavior, and the users with the characteristic data loss are not suitable for being used as the objects of information pushing and all need to be removed. Meanwhile, users with few behaviors are not suitable for being used as objects for information pushing, for example, users who surf the internet less than 100 times per month, users who have less than 10 calls, or users who receive less than 100 short messages need to be filtered. In specific implementation, filtering and screening can be performed on the total number of users according to a preset feature filtering rule to obtain at least one candidate user and the user feature of each candidate user. Wherein the feature filtering rules include: filtering users with missing characteristic data; users with less behavior are filtered.

S102, identifying the user characteristics of each candidate user by using a pre-trained transformation prediction model, and determining the probability of each candidate user responding to the push information.

In this embodiment, the candidate user responding to the push information refers to a corresponding operation performed by the candidate user to convert the push information, for example, completing account registration and subscribing to a service according to the push information or purchasing a target object indicated by the push information; typically, the subject matter referred to by the push information may be an educational course, as the inventive arrangements are applicable to a scenario where an educational institution obtains a guest. It should be noted that, as long as the user converts the push information, the purpose of obtaining the guest by the enterprise is achieved.

And the probability of the candidate user responding to the push information is used for representing the intention of the candidate user to complete account registration and subscribe a certain service or purchase an object indicated by the push information according to the push information. Wherein the greater the probability, the more likely the candidate user is to respond to the push information.

When probability prediction is carried out, the user characteristics of each candidate user only need to be used as input of the conversion prediction model, and then the probability of each candidate user responding to the push information is determined according to the output of the conversion prediction model. The transformation estimation model can be obtained by training a pre-constructed neural network structure based on part of candidate users and corresponding user characteristics.

S103, selecting at least one target user from the candidate users according to the probability of each candidate user responding to the push information, and indicating the operator platform to push the information of the target user.

In this embodiment, a probability threshold may be preset, and if the probability that the candidate user responds to the push information is greater than or equal to the preset probability threshold, it indicates that the candidate user is very likely to respond to the push information, so that the candidate user is screened as a target user, and the operator platform is instructed to push information to the target user, for example, the operator platform is instructed to push information such as marketing advertisements to the target user in a manner of short message, telephone, mail, and the like, so as to achieve enterprise customer acquisition. It should be noted that, if the probability that the candidate user responds to the push information is smaller than the probability threshold, it is determined that the probability that the user converts the push information is extremely low, and at this time, the relevant information is not pushed to the candidate user, so that invalid push of the information can be avoided.

In the embodiment, only the relevant information is pushed to the target user which is likely to respond to the pushed information, so that the problem of low customer acquisition amount or low efficiency caused by invalid pushing of the information can be avoided, and compared with the random pushing of the information to the user, the probability that the target user converts the pushed information is higher, so that the efficiency of acquiring effective customers by an enterprise can be ensured when the relevant information is pushed to the target user.

Example two

Fig. 2 is a flowchart of an information pushing method according to a second embodiment of the present invention. Referring to fig. 2, the process flow of the method includes the following steps:

s201, determining candidate users and user characteristics of each candidate user based on the atomic data of the operator platform.

For a specific implementation process, reference may be made to the foregoing embodiments, which are not described herein again. After the candidate users are obtained, the transformation prediction model may be trained by using user characteristics of some candidate users, and the specific training process is as shown in steps S202 to S205.

S202, determining at least one first-class user and at least one second-class user according to historical push data.

The history push data records information of a push object (i.e. a user) and whether the push object converts push information push. Thus, the first-class users are users who receive the push information and convert the push information, namely users who have successful marketing; the second category of users refers to users who receive the push information but do not convert the push information, namely users who fail marketing.

S203, determining the user characteristics of each first class user and the user characteristics of each second class user according to the candidate users and the user characteristics of each candidate user.

Optionally, the original user identifier of the first class of user and the original user identifier of the second class of user are converted into a platform user identifier suitable for the operator platform. Specifically, the operation Shang Ping performs desensitization processing on the original user identifier of the first type of user and the original user identifier of the second type of user to obtain platform user identifiers of the first type of user and the second type of user on the operator platform. The original user identifier refers to a device code that can uniquely characterize the user, and may be a communication number, for example. Further matching the platform user identification of the first class user with the platform user identification of the candidate user, and taking the matched user characteristic of the candidate user as the user characteristic of the first class user; and matching the platform user identification of the second class of users with the platform user identification of the candidate users, and taking the matched user characteristics of the candidate users as the user characteristics of the first class of users.

S204, screening the user characteristics of each first class of users and screening the user characteristics of each second class of users according to the importance of each user characteristic and/or the correlation between each user characteristic and the user class.

In this embodiment, since the number of the user characteristics of the first type of user serving as the positive sample is large, the user characteristics of the first type of user are screened in order to improve the training efficiency of the conversion estimation model. During specific implementation, the importance of each user characteristic of the first class of users can be calculated firstly, the users are sorted according to the importance, and then N user characteristics with the top sorting are selected; in addition, the filtering can be performed according to the correlation between each user characteristic and the user category, wherein the user category can be represented by data 1 and 0,1 represents the first class of users, and 0 represents the second class of users. Illustratively, the characteristics of which the correlation between the first type of users and each user characteristic meets the preset condition are selected. It should be noted that the two screening methods may be performed alternatively or simultaneously, and when the two screening methods are used simultaneously, N user characteristics may be selected according to the importance, so as to determine the correlations between the user categories of the first type of user and the N user characteristics, and select a part of the user characteristics that satisfy the correlation conditions.

Similarly, when the user characteristics of each second type user are screened, the importance of each user characteristic of the second type users can be calculated firstly, the users are sorted according to the importance, and then N user characteristics which are sorted in the front are selected; in addition, the filtering can be performed according to the correlation between each user characteristic and the user category, wherein the user category can be represented by data 1 and 0,1 represents the first class of users, and 0 represents the second class of users. Illustratively, the feature with the correlation between the second type of user and each user feature meeting the preset condition is selected. Here, the two screening methods may be performed alternatively or simultaneously. When the method is used, N user characteristics can be selected according to the importance degree, then the relevance between the user categories of the second type of users and the N user characteristics is judged, and part of the user characteristics meeting the relevance condition are selected.

S205, training the pre-constructed transformation estimation model according to the user characteristics of each first class of users after screening and the user characteristics of each second class of users after screening.

Through the steps of S202-S204, the user characteristics of the first type of users and the screened first type of users can be used as positive samples; and taking the second type of users and the user characteristics of the screened second type of users as negative samples. And then training the pre-constructed transformation estimation model based on the positive and negative samples until the model converges. Wherein, the output of the conversion estimation model is the probability of the user responding to the push information.

S206, identifying the user characteristics of each candidate user by using the pre-trained transformation prediction model, and determining the probability of each candidate user responding to the push information.

S207, selecting at least one target user from the candidate users according to the probability of each candidate user responding to the push information, and indicating the operator platform to push the information of the target user.

In the embodiment, when the pre-estimation model is trained, the user characteristics of the first class of users and the user characteristics of the second class of users are screened based on the importance of the user characteristics and/or the correlation between each user characteristic and the user category, so that the model training can be carried out by using the optimal user characteristics, and the pre-estimation accuracy of the trained transformation pre-estimation model is further ensured; in addition, only the relevant information is pushed to the target user who possibly responds to the pushed information, so that the problem of low customer acquisition quantity or low efficiency caused by invalid pushing of the information can be avoided, and compared with the random pushing of the information to the user, the probability that the target user converts the pushed information is higher, so that the efficiency that an enterprise acquires valid customers can be ensured when the relevant information is pushed to the target user.

EXAMPLE III

Fig. 3 is a flowchart of an information pushing method according to a third embodiment of the present invention. Referring to fig. 3, the process flow of the method includes the following steps:

s301, determining candidate users and user characteristics of each candidate user based on the atomic data of the operator platform.

S302, determining at least one first class user and at least one second class user according to historical push data.

S303, determining the user characteristics of each first class user and the user characteristics of each second class user according to the candidate users and the user characteristics of each candidate user.

In this embodiment, a calculation scheme of the importance and the relevance of the user features is provided, specifically referring to steps S304-S305, and after calculation according to the two steps, the user features can be quickly screened by using the calculation results subsequently.

S304, calculating the correlation between the user category and each user characteristic by using a chi-square test method.

Wherein, chi-square test is a hypothesis test method of counting data. The method belongs to nonparametric test and mainly comprises the steps of comparing two or more sample rates (composition ratios) and performing relevance analysis on two classification variables. The basic idea is to compare the coincidence degree or goodness of fit between the theoretical frequency and the actual frequency. In this embodiment, the user categories are divided into 1 and 0,1 to represent the first category of users; 0 represents a second class of users. In this way, the relevance between the user category and each user feature is calculated by using a chi-square test method, that is, the relevance between the first class of users and each user feature corresponding to the first class of users is calculated by using a chi-square test method, and the relevance between the second class of users and each user feature corresponding to the second class of users is calculated by using a chi-square test method.

S305, calculating the importance of each user characteristic by using a pre-trained importance calculation model; wherein, the importance degree calculation model is a random forest model.

In this embodiment, a part of candidate users and their corresponding user features may be utilized in advance to train an importance degree calculation model, where the importance degree calculation model is a random forest model. Further, the user characteristics of the first class of users are used as the input of an importance calculation model, and the importance of each user characteristic of the first class of users is determined according to the output of the importance model; similarly, the user characteristics of the second type of users are used as the input of the importance calculation model, and the importance of each user characteristic of the second type of users is determined according to the output of the importance model.

S306, screening the user characteristics of each first class of users and screening the user characteristics of each second class of users according to the importance of each user characteristic and/or the correlation between each user characteristic and the user category.

In this embodiment, the operator platform itself may mark the characteristics of each platform user, that is, the operator platform includes the tag characteristics of the user. Therefore, the first class of users and the second class of users also include the label features corresponding to each user handled by the operator platform itself. In this way, when the conversion estimation model is trained, besides the user characteristics of the first class of users and the user characteristics of the second class of users, the label characteristics of the first class of users and the label characteristics of the second class of users are used. See S307 for a specific training process.

And S307, training the pre-constructed transformation estimation model according to the label characteristics and the screened user characteristics of each first class of users and the label characteristics and the screened user characteristics of each second class of users.

In this embodiment, the pre-constructed transformation prediction model is a WDL (Wide & Deep Learning) prediction model. The WDL pre-estimation model structure comprises two parts. The Wide part is cross features of manual prior, mainly user tag features directly printed by an operator, and direct prediction is performed in the form of an LR model if the number of the user tag features is small. The Deep part on the right is a forward neural network, and the forward neural network is grouped to enter a full-connection neural network according to the characteristic type (conversation, short message or networking behavior information). WDL uses FTRL to train Wide part, uses AdaGrad to train Deep part, for the two-classification problem, the model estimation function is expressed as:

where σ is the sigmoid function, φ () is a cross-transform,

is Wide partial weight, based on>

Is the last layer weight of the Deep portion, b is a bias parameter, <' >>

Is an input feature of the Deep section.

The loss function is then the classical logarithmic loss function J (θ):

where m is the number of samples, h _θ () Is a prediction function.

After the model structure and the loss function are defined, the label features of the first class of users and the screened user features, and the label features of the second class of users and the screened user features are used for training until the model converges.

S308, identifying the user characteristics of each candidate user by using the pre-trained transformation prediction model, and determining the probability of each candidate user responding to the push information.

S309, selecting at least one target user from the candidate users according to the probability of each candidate user responding to the push information, and instructing the operator platform to push the information to the target user.

In the embodiment, the relevance between the user category and each user characteristic can be accurately calculated in a chi-square test mode, and the importance of each user characteristic can be quickly determined by using an importance calculation model, so that the follow-up characteristic screening is guaranteed; and the user characteristics and the label characteristics of the first class of users and the user characteristics and the label characteristics of the second class of users are simultaneously utilized to train, transform and predict the model, so that the accuracy of model prediction can be ensured.

Example four

Fig. 4 is a schematic structural diagram of an information pushing apparatus according to a fourth embodiment of the present invention, which is applicable to a scenario in which an enterprise obtains customers by pushing marketing information to a user. As shown in fig. 4, the apparatus includes:

a user and feature mining module 401, configured to determine candidate users and user features of each candidate user based on atomic data of an operator platform;

a prediction module 402, configured to identify a user characteristic of each candidate user by using a pre-trained transformation prediction model, and determine a probability that each candidate user responds to the push information;

an information pushing module 403, configured to select at least one target user from the candidate users according to a probability that each candidate user responds to the pushed information, and instruct the operator platform to push information to the target user.

On the basis of the embodiment, optionally, the atomic data comprises desensitized user call data, short message data and user internet behavior data;

correspondingly, the user and feature mining module is further configured to:

aiming at each user in the operator platform full-volume users, mining the user characteristics of each user from user call data, short message data and user internet behavior data according to different preset dimensions;

and filtering and screening the full amount of users according to a preset feature filtering rule to obtain at least one candidate user and the user feature of each candidate user.

On the basis of the foregoing embodiment, optionally, the system further includes a model training module, where the model training module includes:

the first data mining unit is used for determining at least one first class user and at least one second class user according to historical push data;

the second data mining unit is used for determining the user characteristics of each first class of users and the user characteristics of each second class of users according to the candidate users and the user characteristics of each candidate user;

the screening unit is used for screening the user characteristics of each first class of users and screening the user characteristics of each second class of users according to the importance of each user characteristic and/or the correlation between each user characteristic and the user class;

and the training unit is used for training the pre-constructed transformation estimation model according to the user characteristics of each first class of users after screening and the user characteristics of each second class of users after screening.

On the basis of the foregoing embodiment, optionally, the second data mining unit is further configured to:

converting the original user identification of the first class of users and the original user identification of the second class of users into platform user identification suitable for an operator platform;

matching the platform user identification of the first class user with the platform user identification of the candidate user, and taking the matched user characteristic of the candidate user as the user characteristic of the first class user;

and matching the platform user identification of the second class of users with the platform user identification of the candidate users, and taking the matched user characteristics of the candidate users as the user characteristics of the first class of users.

On the basis of the above embodiment, optionally, the method further includes:

the correlation calculation module is used for calculating the correlation between the user category and each user characteristic by using a chi-square test method;

the importance calculation module is used for calculating the importance of each user characteristic by utilizing a pre-trained importance calculation model; wherein, the importance degree calculation model is a random forest model.

On the basis of the above embodiment, optionally, the user characteristics further include a tag characteristic corresponding to each user, which is processed by the operator platform itself;

correspondingly, the training unit is further configured to:

training the pre-constructed transformation estimation model according to the label characteristics and the screened user characteristics of each first class user and the label characteristics and the screened user characteristics of each second class user; the pre-constructed transformation prediction model is a WDL prediction model.

The information pushing device provided by the embodiment of the invention can execute the information pushing method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

EXAMPLE five

FIG. 5 illustrates a block diagram of an electronic device 10 that may be used to implement an embodiment of the invention. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.

As shown in fig. 5, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, and the like, wherein the memory stores a computer program executable by the at least one processor, and the processor 11 can perform various suitable actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from a storage unit 18 into the Random Access Memory (RAM) 13. In the RAM13, various programs and data necessary for the operation of the electronic apparatus 10 can also be stored. The processor 11, the ROM12, and the RAM13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to the bus 14.

A number of components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, or the like; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.

The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, or the like. The processor 11 performs the various methods and processes described above, such as performing an information push method.

In some embodiments, the information pushing method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM12 and/or the communication unit 19. When the computer program is loaded into the RAM13 and executed by the processor 11, one or more steps of the information push method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the information push method by any other suitable means (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), system on a chip (SOCs), complex Programmable Logic Devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.

Computer programs for implementing the methods of the present invention can be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be performed. A computer program can execute entirely on a machine, partly on a machine, as a stand-alone software package partly on a machine and partly on a remote machine or entirely on a remote machine or server.

In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.

The computing system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical host and VPS service are overcome.

It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present invention may be executed in parallel, sequentially, or in different orders, and are not limited herein as long as the desired results of the technical solution of the present invention can be achieved.

The above-described embodiments should not be construed as limiting the scope of the invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An information pushing method, comprising:

determining candidate users and user characteristics of each candidate user based on atomic data of an operator platform;

identifying the user characteristics of each candidate user by using a pre-trained transformation estimation model, and determining the probability of each candidate user responding to the push information;

and selecting at least one target user from the candidate users according to the probability of each candidate user responding to the pushed information, and indicating the operator platform to push the information to the target user.

2. The method of claim 1, wherein the atomic data comprises desensitized user call data, short message data, and user surfing behavior data;

correspondingly, the candidate users and the user characteristics of each candidate user are determined based on the atomic data of the operator platform, and the method comprises the following steps:

aiming at each user in the operator platform full-volume users, mining the user characteristics of each user from the user call data, the short message data and the user internet behavior data according to different preset dimensions;

3. The method of claim 1, wherein the training process of the conversion estimation model comprises:

determining at least one first class user and at least one second class user according to historical push data;

determining the user characteristics of each first class of users and the user characteristics of each second class of users according to the candidate users and the user characteristics of each candidate user;

screening the user characteristics of each first class of users and screening the user characteristics of each second class of users according to the importance of each user characteristic and/or the correlation between each user characteristic and the user category;

and training the pre-constructed transformation estimation model according to the user characteristics of each first class of users after being screened and the user characteristics of each second class of users after being screened.

4. The method of claim 3, wherein determining the user characteristics of each first class of users and the user characteristics of each second class of users according to the candidate users and the user characteristics of each candidate user comprises:

converting the original user identification of the first class of users and the original user identification of the second class of users into platform user identification suitable for the operator platform;

and matching the platform user identification of the second class of users with the platform user identification of the candidate user, and taking the matched user characteristics of the candidate user as the user characteristics of the first class of users.

5. The method of claim 3, further comprising:

calculating the correlation between the user category and each user characteristic by using a chi-square test method;

calculating the importance of each user characteristic by using a pre-trained importance calculation model; and the importance calculation model is a random forest model.

6. The method of claim 3, wherein the user characteristics further include a label characteristic for each user handled by the operator platform itself;

correspondingly, training the pre-constructed transformation estimation model according to the user characteristics of each first class of users after screening and the user characteristics of each second class of users after screening, comprising the following steps:

7. An information pushing apparatus, comprising:

and the information pushing module is used for selecting at least one target user from the candidate users according to the probability of each candidate user responding to the pushed information and indicating the operator platform to push the information of the target user.

8. The method of claim 1, further comprising a model training module to:

according to the importance of each user characteristic and/or the correlation between each user characteristic and the user category, screening the user characteristics of each first type of user, and screening the user characteristics of each second type of user;

and training the pre-constructed transformation estimation model according to the user characteristics of each first class of users after screening and the user characteristics of each second class of users after screening.

9. An electronic device, comprising:

at least one processor; and

the memory stores a computer program executable by the at least one processor, the computer program being executable by the at least one processor to enable the at least one processor to perform the information push method of any one of claims 1-6.

10. A computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions for causing a processor to implement the information pushing method according to any one of claims 1 to 6 when executed.