CN113297478A

CN113297478A - Information pushing method and device based on user life cycle and electronic equipment

Info

Publication number: CN113297478A
Application number: CN202110446233.4A
Authority: CN
Inventors: 张燕
Original assignee: Shanghai Qiyue Information Technology Co Ltd
Current assignee: Shanghai Qiyue Information Technology Co Ltd
Priority date: 2021-04-25
Filing date: 2021-04-25
Publication date: 2021-08-24
Anticipated expiration: 2041-04-25
Also published as: CN113297478B

Abstract

The disclosure relates to an information pushing method and device based on a user life cycle, electronic equipment and a computer readable medium. The method comprises the following steps: acquiring user information of a user, wherein the user information comprises basic information and behavior information; generating a plurality of core indicators and a plurality of original parameters based on the user information; inputting the plurality of core indexes and the plurality of original parameters into a user life cycle model generated through long-term and short-term memory network training to obtain the current stage of the user; generating policy information and/or marketing information for the user based on the current stage. According to the information pushing method and device based on the life cycle of the user, the electronic equipment and the computer readable medium, fine and personalized marketing can be performed on the users with different life cycles, the operation cost is reduced, and the user value is improved.

Description

Information pushing method and device based on user life cycle and electronic equipment

Technical Field

The present disclosure relates to the field of computer information processing, and in particular, to an information pushing method and apparatus based on a user lifecycle, an electronic device, and a computer-readable medium.

Background

The customer life cycle refers to the period from when a customer learns about the enterprise or when the enterprise wants to develop a certain customer until the business relationship between the customer and the enterprise is completely terminated and the related matters are completely processed. The life cycle of a customer is an evolution of the life cycle of an enterprise product, but for a business enterprise, the life cycle of a customer is much more important than the life cycle of a product of the enterprise. The customer lifecycle describes the overall characteristics of the movement of customer relationships from one state (one phase) to another state (another phase).

In order to provide more accurate service to users, improve the company effect, and maximize the value of company products, more and more companies are now paying attention to service policies specified for the life cycle of users. The corresponding user life cycles are completely different for different products. For companies providing internet financial services, the life cycle of users of the companies is definitely different from the life cycle of users of a game platform or a shopping platform, and how to introduce the life cycle of the users into the field of internet financial services to provide better financial services for the users of the companies is a current hotspot problem.

The above information disclosed in this background section is only for enhancement of understanding of the background of the disclosure and therefore it may contain information that does not constitute prior art that is already known to a person of ordinary skill in the art.

Disclosure of Invention

In view of this, the present disclosure provides an information pushing method and apparatus based on a user life cycle, an electronic device, and a computer readable medium, which can perform refinement and personalized marketing for users with different life cycles, reduce operation cost, and improve user value.

Additional features and advantages of the disclosure will be set forth in the detailed description which follows, or in part will be obvious from the description, or may be learned by practice of the disclosure.

According to an aspect of the present disclosure, an information pushing method based on a user lifecycle is provided, where the method includes: acquiring user information of a user, wherein the user information comprises basic information and behavior information; generating a plurality of core indicators and a plurality of original parameters based on the user information; inputting the plurality of core indexes and the plurality of original parameters into a user life cycle model generated through long-term and short-term memory network training to obtain the current stage of the user; generating policy information and/or marketing information for the user based on the current stage.

Optionally, the method further comprises: calculating current stages of a plurality of users through the user lifecycle model; generating user guest group assessment information and/or user guest group structure information based on the current stage of the plurality of users.

Optionally, the method further comprises: acquiring a plurality of user information corresponding to a plurality of historical users; generating a plurality of groups of core indexes and a plurality of groups of original parameters based on the plurality of user information, wherein one group of core indexes and original parameters correspond to one historical user; determining user lifecycle samples and stage metrics based on the sets of core metrics and the sets of original parameters; inputting the multiple groups of core indexes, the multiple groups of original parameters and the user life cycle samples into a long-term and short-term memory network for training to generate a training result; and performing parameter optimization on the training result to generate the user life cycle model.

Optionally, generating a plurality of core metrics and a plurality of original parameters based on the user information includes: generating a repurchase rate index based on accumulated consumption data in the user information; generating a growth rate indicator based on consumption data within a current stage in the user information; generating an aggressiveness index based on the plurality of categories in the user information and the amount corresponding thereto; generating a fund proportion index based on the asset transition data in the user information.

Optionally, inputting the plurality of core indicators and the plurality of original parameters into a user lifecycle model generated through long-short term memory network training, to obtain a current stage of the user, including: inputting the plurality of core indicators and the plurality of original parameters into the user lifecycle model to generate a plurality of one-hot coded corresponding probabilities thereof; parsing the plurality of unique hot code generation life cycles; and taking the stage with the highest probability in the plurality of stages as the current stage of the user.

Optionally, determining user lifecycle samples and stage metrics based on the sets of core metrics and the sets of original parameters comprises: performing cluster calculation on the multiple groups of core indexes and the multiple groups of original parameters to generate multiple historical user categories; extracting core users in each of the plurality of historical user categories; user lifecycle samples are generated by a plurality of core users.

Optionally, determining a user life cycle sample and a stage index based on the plurality of sets of core indexes and the plurality of sets of original parameters further comprises: determining a stage index of each stage in the user life cycle based on the user information of a plurality of historical users and the user life cycle sample, wherein the stage index is used for judging the stage of the user.

Optionally, inputting the plurality of sets of core indicators, the plurality of sets of original parameters, and the user life cycle sample into a long-short term memory network for training, and generating a training result, including: converting the plurality of phases of the user lifecycle sample into a plurality of one-hot codes; inputting the multiple groups of core indexes, the multiple groups of original parameters and the multiple unique hot codes into a long-short term memory network for training to generate a training result, wherein the training result is the multiple unique hot codes and the corresponding probabilities thereof; the long-term and short-term memory network is of a six-layer structure, and the training step length is five.

Optionally, performing parameter tuning on the training result to generate the user lifecycle model, including: generating an evaluation function; adjusting and optimizing the multiple groups of core indexes based on the evaluation function; adjusting the stage indexes based on the evaluation function; and optimizing the user life cycle model based on the evaluation function.

Optionally, generating an evaluation function comprises: generating a cohesion function according to the occurrence frequency of each stage after the change of the historical user life cycle stage; generating a separation function according to the existing duration of each stage of the life cycle of the historical user; generating the evaluation function based on the cohesion function and the separation function.

According to an aspect of the present disclosure, an information pushing apparatus based on a user lifecycle is provided, the apparatus including: the information module is used for acquiring user information of a user, wherein the user information comprises basic information and behavior information; a parameter module for generating a plurality of core indicators and a plurality of original parameters based on the user information; the calculation module is used for inputting the plurality of core indexes and the plurality of original parameters into a user life cycle model generated through long-term and short-term memory network training to obtain the current stage of the user; a user module to generate policy information and/or marketing information for the user based on the current stage.

Optionally, the method further comprises: an enterprise module for computing current phases of a plurality of users through the user lifecycle model; generating user guest group assessment information and/or user guest group structure information based on the current stage of the plurality of users.

Optionally, the method further comprises: the history information module is used for acquiring a plurality of user information corresponding to a plurality of history users; the historical parameter module is used for generating a plurality of groups of core indexes and a plurality of groups of original parameters based on the plurality of user information, wherein one group of core indexes and original parameters correspond to one historical user; a sample index module for determining user lifecycle samples and stage indexes based on the sets of core indexes and the sets of original parameters; the model training module is used for inputting the plurality of groups of core indexes, the plurality of groups of original parameters and the user life cycle sample into a long-term and short-term memory network for training to generate a training result; and the parameter tuning module is used for performing parameter tuning on the training result to generate the user life cycle model.

Optionally, the parameter module is further configured to generate a repurchase rate index based on the accumulated consumption data in the user information; generating a growth rate indicator based on consumption data within a current stage in the user information; generating an aggressiveness index based on the plurality of categories in the user information and the amount corresponding thereto; generating a fund proportion index based on the asset transition data in the user information.

Optionally, the calculation module includes: an input unit, configured to input the plurality of core indicators and the plurality of original parameters into the user lifecycle model, and generate corresponding probabilities of a plurality of unique hot codes; the analysis unit is used for analyzing a plurality of stages of the life cycle generated by the plurality of one-hot codes; a phase unit, configured to use a phase with a highest probability among the multiple phases as a current phase of the user.

Optionally, the sample indicator module includes: the clustering unit is used for carrying out clustering calculation on the multiple groups of core indexes and the multiple groups of original parameters to generate multiple historical user categories; a core unit, configured to extract a core user in each of the plurality of historical user categories; a sample unit for generating user lifecycle samples by a plurality of core users.

Optionally, the sample index module further includes: the index unit is used for determining the stage indexes of all stages in the user life cycle based on the user information of a plurality of historical users and the user life cycle samples, and the stage indexes are used for judging the stage where the user is located by the user.

Optionally, the model training module includes: the encoding unit is used for converting the multiple stages of the user life cycle sample into multiple one-hot codes; the training unit is used for inputting the multiple groups of core indexes, the multiple groups of original parameters and the multiple unique hot codes into a long-term and short-term memory network for training to generate a training result, wherein the training result is the multiple unique hot codes and the corresponding probabilities thereof; the long-term and short-term memory network is of a six-layer structure, and the training step length is five.

Optionally, the parameter tuning module includes: a function unit for generating an evaluation function; the tuning unit is used for tuning the multiple groups of core indexes based on the evaluation function; adjusting the stage indexes based on the evaluation function; and optimizing the user life cycle model based on the evaluation function.

Optionally, the function unit is further configured to generate a cohesive function according to the number of times that each stage appears after the change of the historical user life cycle stage; generating a separation function according to the existing duration of each stage of the life cycle of the historical user; generating the evaluation function based on the cohesion function and the separation function.

According to an aspect of the present disclosure, an electronic device is provided, the electronic device including: one or more processors; storage means for storing one or more programs; when executed by one or more processors, cause the one or more processors to implement a method as above.

According to an aspect of the disclosure, a computer-readable medium is proposed, on which a computer program is stored, which program, when being executed by a processor, carries out the method as above.

According to the information pushing method and device based on the user life cycle, the electronic equipment and the computer readable medium, user information of a user is obtained, wherein the user information comprises basic information and behavior information; generating a plurality of core indicators and a plurality of original parameters based on the user information; inputting the plurality of core indexes and the plurality of original parameters into a user life cycle model generated through long-term and short-term memory network training to obtain the current stage of the user; based on the mode of generating strategy information and/or marketing information for the users in the current stage, refined and personalized marketing can be performed for the users with different life cycles, the operation cost is reduced, and the user value is improved.

According to the information pushing method and device based on the user life cycle, the electronic equipment and the computer readable medium, enterprises can be helped to analyze the current user structure, evaluate the user quality, stability, effectiveness, loyalty and the like, and subdivide client groups, so that the problem that decision judgment is wrong due to noise influence caused by information loss is avoided, and the inviscid users can be recognized in advance.

According to the information pushing method and device based on the user life cycle, the electronic equipment and the computer readable medium, enterprises can be helped to deduce future user structure changes, marginal benefit maximization is achieved, and supply and demand balance is guaranteed.

According to the information pushing method and device based on the life cycle of the user, the electronic equipment and the computer readable medium, the property condition of the user can be known in time, and proper customer care is given based on long-term customer relationship. Such as user short term fund turnover difficulties or other difficulties, providing more favorable borrowing or other assistance to address user short term difficulties based on user historical performance.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

Drawings

The above and other objects, features and advantages of the present disclosure will become more apparent by describing in detail exemplary embodiments thereof with reference to the attached drawings. The drawings described below are merely some embodiments of the present disclosure, and other drawings may be derived from those drawings by those of ordinary skill in the art without inventive effort.

Fig. 1 is a system block diagram illustrating an information push method and apparatus based on a user lifecycle according to an exemplary embodiment.

Fig. 2 is a flowchart illustrating an information push method based on a user lifecycle according to an exemplary embodiment.

Fig. 3 is a flowchart illustrating an information push method based on a user lifecycle according to another exemplary embodiment.

Fig. 4 is a block diagram illustrating an information push apparatus based on a user lifecycle according to an exemplary embodiment.

Fig. 5 is a block diagram illustrating an information push apparatus based on a user lifecycle according to another exemplary embodiment.

FIG. 6 is a block diagram illustrating an electronic device in accordance with an example embodiment.

FIG. 7 is a block diagram illustrating a computer-readable medium in accordance with an example embodiment.

Detailed Description

Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of example embodiments to those skilled in the art. The same reference numerals denote the same or similar parts in the drawings, and thus, a repetitive description thereof will be omitted.

Furthermore, the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the disclosure. One skilled in the relevant art will recognize, however, that the subject matter of the present disclosure can be practiced without one or more of the specific details, or with other methods, components, devices, steps, and so forth. In other instances, well-known methods, devices, implementations, or operations have not been shown or described in detail to avoid obscuring aspects of the disclosure.

The block diagrams shown in the figures are functional entities only and do not necessarily correspond to physically separate entities. I.e. these functional entities may be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor means and/or microcontroller means.

The flow charts shown in the drawings are merely illustrative and do not necessarily include all of the contents and operations/steps, nor do they necessarily have to be performed in the order described. For example, some operations/steps may be decomposed, and some operations/steps may be combined or partially combined, so that the actual execution sequence may be changed according to the actual situation.

It will be understood that, although the terms first, second, third, etc. may be used herein to describe various components, these components should not be limited by these terms. These terms are used to distinguish one element from another. Thus, a first component discussed below may be termed a second component without departing from the teachings of the disclosed concept. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

It is to be understood by those skilled in the art that the drawings are merely schematic representations of exemplary embodiments, and that the blocks or processes shown in the drawings are not necessarily required to practice the present disclosure and are, therefore, not intended to limit the scope of the present disclosure.

The technical abbreviations involved in this disclosure are explained as follows:

life cycle of the user: simply the entire process from contacting the product to leaving the product. The life cycle is widely applied in different industries and fields. As an important index for the company to recognize the user portrait, the method is also a core basic work for the enterprise to improve the user value, and if the life cycle is not reasonably identified and divided, the enterprise can wrongly understand the group composition of the clients, so that the method cannot correctly guide the operation of the company, but influences the normal operation of the company.

K-Means: in the space, K cluster centers are randomly determined, data sample points to be classified are distributed to the nearest clusters according to the nearest principle, then the coordinate average value of all the points in each cluster is calculated, a new cluster center is determined again according to the calculation result, and iteration is continuously performed according to the method until the moving distance of the last cluster center is smaller than a given value or the clustering frequency meets the requirement.

A neural network model: is a model that contains input, output and computational functions. The input can be analogized to dendrites of each neuron, the output can be analogized to axons of each neuron, and the calculation can be analogized to individual nuclei [ ]. It is also understood to be a black box that can be used to model any function, given x, the desired function y can be calculated by the neural network, provided there are enough x, y training samples.

Activation function: the function is a very important function in each model of the whole neural network, and the main purpose of the function is to make up the defect of insufficient interpretation capability of a linear model and add a nonlinear factor into the neural network. Since each function of the neural network is differentiable, the activation function chosen must also ensure that its inputs and outputs are also differentiable.

Loss function: the method is a very key function in neural network learning, and guides the convergence direction of a model through the difference size of a real value and a predicted value. The common algorithm is: mean Squared Error (MSE) and cross entropy.

Gradient reduction: the method has the main function that the minimized loss function and the model parameter value are solved, the gradient descent method is used for gradually iterating and solving, when the gradient vector is equal to 0, the loss function reaches a minimum value, and then the gradient descent algorithm stops iterating and calculating.

LSTM: by reserving errors and carrying out reverse transmission along time and network layers, the method has the greatest advantage that the errors are kept on a more constant level, so that the cyclic network can carry out learning of relatively more time steps (more than 1000 time steps), a channel for establishing remote causal connection is further opened, and the problem of gradient disappearance in the cyclic network can be solved through the structure.

As shown in fig. 1, the system architecture 10 may include

terminal devices

101, 102, 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the

terminal devices

101, 102, 103 and the server 105. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.

The user may use the

terminal devices

101, 102, 103 to interact with the server 105 via the network 104 to receive or send messages or the like. The

terminal devices

101, 102, 103 may have various communication client applications installed thereon, such as a financial services application, a shopping application, a web browser application, an instant messaging tool, a mailbox client, social platform software, and the like.

The

terminal devices

101, 102, 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop portable computers, desktop computers, and the like.

The server 105 may be a server that provides various services, such as a background management server that supports financial services websites browsed by the user using the

terminal apparatuses

101, 102, and 103. The background management server may analyze the received user data, and feed back the processing result to the administrator of the financial service website and/or the

terminal device

101, 102, 103.

The server 105 may, for example, obtain user information of the user, the user information including basic information and behavior information; the server 105 may generate a plurality of core metrics and a plurality of raw parameters, e.g., based on the user information; the server 105 may, for example, input the plurality of core indicators and the plurality of original parameters into a user lifecycle model generated through long-short term memory network training, to obtain a current stage of the user; server 105 may generate policy information and/or marketing information for the user, e.g., based on the current stage.

Server 105 may also calculate the current phase of multiple users, e.g., through the user lifecycle model; generating user guest group assessment information and/or user guest group structure information based on the current stage of the plurality of users.

The server 105 may be a single entity server, or may be composed of a plurality of servers, for example, it should be noted that the information pushing method based on the user lifecycle provided by the embodiment of the present disclosure may be executed by the server 105, and accordingly, an information pushing apparatus based on the user lifecycle may be disposed in the server 105. And the web page end provided for the user to browse the financial service platform is generally positioned in the

terminal equipment

101, 102 and 103.

Fig. 2 is a flowchart illustrating an information push method based on a user lifecycle according to an exemplary embodiment. The information pushing method 20 based on the user life cycle at least includes steps S202 to S208.

As shown in fig. 2, in S202, user information of a user is acquired, the user information including basic information and behavior information. The basic information may include the gender, age, phone number, job occupation, income occupation, etc. registered by the user on a financial service platform. The behavior information may include operation information of the user on a certain financial service platform, and may include multiple logins and time thereof, registration time, time and times of borrowing operation and repayment operation, and the like.

In S204, a plurality of core metrics and a plurality of original parameters are generated based on the user information. A repurchase rate indicator may be generated, for example, based on cumulative consumption data in the user information; generating a growth rate indicator based on consumption data within a current stage in the user information; generating an aggressiveness index based on the plurality of categories in the user information and the amount corresponding thereto; generating a fund proportion index based on the asset transition data in the user information. The original parameters can be related contents such as credit amount, account amount, remaining available amount and the like.

Wherein, the repurchase rate can represent the historical accumulated consumption participation of the user:

wherein Mi is the ith consumption amount, R_t-1The cumulative fund repurchase rate after the last consumption is Te [1, ∞);

the growth rate may represent a trend representing the user's recent consumption:

wherein, Δ m_t-1For the capital growth rate of the last day, rho is momentum, rho belongs to [0,1 ]]；

The aggressiveness index may represent the current user's integrated enthusiasm:

wherein, for the "advance payment amount":

wherein f (p) epsilon (0, infinity), w_pThe default value is 0.5 for the weight coefficient of the advance payment amount, wherein for the acquired marketing amount:

wherein f (w) e (-infinity, 0), w_wThe default value is 0.3 for the acquired marketing amount weight coefficient

Wherein, for "consumption":

wherein f (i) e (0, ∞), w_iThe default value is 1 for the consumption weight coefficient

Wherein, for "repayment":

wherein f (r) ∈ (-infinity, 0)]，w_rThe default value is 0.5 for repayment weight coefficient

Secondly, calculating the enthusiasm index of the user on the day

Finally, obtaining a weighted user enthusiasm index according to the user enthusiasm index of the past day

The fund history ratio can represent the current user asset history level:

wherein R is_tRepresents the current historical proportion of capital, sigma M_t-1Represents the sum of the total assets over the past days of change (t-1 days).

In S206, the plurality of core indicators and the plurality of original parameters are input into a user lifecycle model generated by long-short term memory network training, so as to obtain a current stage of the user.

In One embodiment, the plurality of core metrics and the plurality of original parameters may be input into the user lifecycle model, generating their corresponding probabilities for a plurality of One-Hot codes (One-Hot); parsing the plurality of unique hot code generation life cycles; and taking the stage with the highest probability in the plurality of stages as the current stage of the user.

In one embodiment, the user lifecycle includes a plurality of phases: induction period, growth period, maturation period, dormancy period, and abortion period. Each stage corresponds to a unique hot code. Through calculation of the user lifecycle model, each one-hot code and its corresponding probability are input. The probability represents the likelihood that the user is in the current stage. The phase in which the user is currently located can be determined from the maximum probability.

More specifically, any prediction sample can be predicted by calling a model predictor, np.

In S208, policy information and/or marketing information is generated for the user based on the current stage. In one embodiment, wherein the user lead-in period: and a user acquisition stage, wherein the potential user flow in the market is converted into the user at home. Growth period: register for login and activate, and have begun to experience the relevant service or function of the product. And (3) mature period: deeply using the functions or services of the product, contributing more active time, advertising revenue or payment, and the like. A dormant period: mature users who do not produce value behavior for a period of time. And (3) loss period: users that have not logged in and accessed for more than a period of time.

According to different stages of the user, different user strategies are allocated to the user, so that the user can obtain better user experience on the platform. For example, for a user in a sleep period, preferential information can be pushed to the user so as to promote the user to perform activities on the platform.

According to the information pushing method based on the life cycle of the user, the user information of the user is obtained, and the user information comprises basic information and behavior information; generating a plurality of core indicators and a plurality of original parameters based on the user information; inputting the plurality of core indexes and the plurality of original parameters into a user life cycle model generated through long-term and short-term memory network training to obtain the current stage of the user; based on the mode of generating strategy information and/or marketing information for the users in the current stage, refined and personalized marketing can be performed for the users with different life cycles, the operation cost is reduced, and the user value is improved.

In one embodiment, further comprising: calculating current stages of a plurality of users through the user lifecycle model; generating user guest group assessment information and/or user guest group structure information based on the current stage of the plurality of users. User data of all users in the current platform can be acquired regularly, and the stages of the users are determined based on the user life cycle model.

The distribution proportion of users in all stages in the current platform is analyzed through a statistical analysis method, user customer group evaluation information is generated according to different user proportion, and the user structure in a period of time in the future can be predicted. Based on the above analysis, the business strategy of the company can be adjusted to enable the user structure to achieve optimal distribution and the storage platform to operate well.

According to the information pushing method based on the user life cycle, an enterprise can be helped to analyze the current user structure, the user quality, stability, effectiveness, loyalty and the like are evaluated, and a client group is subdivided, so that the error in decision judgment caused by noise influence caused by information loss is avoided, and the inviscid user is recognized in advance;

according to the information pushing method based on the user life cycle, enterprises can be helped to deduce future user structure changes, so that marginal benefit maximization is achieved, and supply and demand balance is guaranteed;

according to the information pushing method based on the life cycle of the user, the assets of the user can be known in time, and proper customer care is given based on long-term customer relations. Such as user short term fund turnover difficulties or other difficulties, providing more favorable borrowing or other assistance to address user short term difficulties based on user historical performance.

It should be clearly understood that this disclosure describes how to make and use particular examples, but the principles of this disclosure are not limited to any details of these examples. Rather, these principles can be applied to many other embodiments based on the teachings of the present disclosure.

Fig. 3 is a flowchart illustrating an information push method based on a user lifecycle according to another exemplary embodiment. The process 30 shown in FIG. 3 is a detailed description of "generating a user lifecycle model through long-short term memory network training".

As shown in fig. 3, in S302, a plurality of user information corresponding to a plurality of history users is acquired.

In S304, a plurality of sets of core indicators and a plurality of sets of original parameters are generated based on the plurality of user information, wherein one set of core indicators and original parameters corresponds to one historical user.

In S306, user lifecycle samples and stage metrics are determined based on the sets of core metrics and the sets of raw parameters. Cluster calculations may be performed on the sets of core metrics and the sets of raw parameters to generate a plurality of historical user categories; extracting core users in each of the plurality of historical user categories; user lifecycle samples are generated by a plurality of core users. Phase indicators for various phases in the user lifecycle that the user determines the phase the user is in may also be determined based on user information for a plurality of historical users and the user lifecycle samples.

More specifically, the phase indexes of each phase in the user life cycle may be determined according to the characteristics of the core user, and the phase indexes may specifically include a value range of the core index. In general, the historical user and the core user may be cluster compared to determine the stage of the user lifecycle in which the historical user is located. In some cases, when the historical users and the core users cannot be clustered well by means of clustering, the threshold ranges of the historical users and the core indicators can be compared to determine the corresponding stages of the historical users and the core users.

In S308, the multiple sets of core indicators, the multiple sets of original parameters, and the user life cycle samples are input into a long-short term memory network for training, and a training result is generated. The multiple phases of the user lifecycle sample can be converted into multiple one-hot codes, for example; inputting the multiple groups of core indexes, the multiple groups of original parameters and the multiple unique hot codes into a long-short term memory network for training to generate a training result, wherein the training result is the multiple unique hot codes and the corresponding probabilities thereof; the long-term and short-term memory network is of a six-layer structure, and the training step length is five.

The LSTM-Long Short-Term Memory network (Long Short-Term Memory) is provided with Long and Short Memory units, wherein each Long and Short Memory unit is provided with three NAND gates which are an input gate, a forgetting gate and an output gate, and when input information at the time t or a hidden state of t-1 enters the three gates respectively, whether the hidden state is abandoned, updated and read is determined according to a certain rule or algorithm. However, the method is a supervised model, and samples with significant characteristics need to be screened out through a certain technical means to perform supervised learning.

The method comprises the steps of firstly selecting a user life cycle sample with remarkable characteristics by self-defining a core index and using an unsupervised algorithm K-means clustering war to an original sample, complementarily mapping data with unobvious residual characteristics to corresponding user life cycle stages in a self-defined rule mode, and then constructing a model by a supervised LSTM data sample based on clustering and rule mapping to realize identification and prediction of the user life cycle.

More specifically, the core index, the original credit amount, the current account amount, the remaining available amount and the like can be added into the characteristic engineering used for model training. After the data is normalized, model training is performed.

The method can adopt the firm _ generator of the keras to carry out model training, customize the DataLoader class of the data loader to carry out data loading, and utilize the generator and the like to process a data layer. In the training process, the training time step is set to 5, the sample trains a plurality of very similar indexes and 5 One-hot codes as the input of the model, and the output of the 5 One-hot codes is used as the final output of the model.

And MSE can be used as a loss function, ADAM is used as an optimizer, the model is set to be six layers, and model construction is finally completed. And then respectively substituting the model into a data generator of the DataLoader, the batch size and the steps _ per _ epoch to carry out model training, thereby realizing the optimal loss value and accuracy.

In S310, the training results are parameter-tuned to generate the user lifecycle model. An evaluation function may be generated, for example; adjusting and optimizing the multiple groups of core indexes based on the evaluation function; adjusting the stage indexes based on the evaluation function; and optimizing the user life cycle model based on the evaluation function.

Wherein generating an evaluation function comprises: generating a cohesion function according to the occurrence frequency of each stage after the change of the historical user life cycle stage; generating a separation function according to the existing duration of each stage of the life cycle of the historical user; generating the evaluation function based on the cohesion function and the separation function.

More specifically, the quality of the model can be evaluated by evaluating the degree of cohesion and separation between the results.

Defining a cohesion function:

wherein F (x) represents the degree of cohesion and takes the value of (0, 1)]，x₀To x₄Respectively representing the occurrence frequency of each life cycle stage after the life cycle stage of the user changes, wherein the larger the value is, the higher the cohesion is, and the more equal the occurrence frequency of each stage is.

Defining a separation function:

wherein F (y) represents the degree of separation and takes the value of [0, 1%]，y₀To y₄The method respectively represents whether five stages of the germination stage, the growth stage, the maturation stage, the decline stage and the loss stage of the life cycle of a user exist, and the larger the value is, the more comprehensive the life cycle stage is represented.

Final evaluation function:

wherein F (i) represents the final evaluation function and takes the value of (0, 1)]，f_xiA value representing the cohesion function of the ith user, f_yiAnd (3) representing the value of the separation function of the ith user, wherein the larger the value is, the better the recognition effect of the life cycle stage is represented.

In addition, for better understanding of the effect of the subsequent test, the following two detailed evaluation functions can be designed:

f_x(i)＝avg(f_xi)

f_y(i)＝avg(f_yi) Wherein f is_x(i),f_y(i) Respectively representing the cohesion degree and separation degree of the whole body, and the effect is better when the value is larger. And respectively optimizing the core index, the stage index and the user life cycle model through the functions.

Those skilled in the art will appreciate that all or part of the steps implementing the above embodiments are implemented as computer programs executed by a CPU. When executed by the CPU, performs the functions defined by the above-described methods provided by the present disclosure. The program may be stored in a computer readable storage medium, which may be a read-only memory, a magnetic or optical disk, or the like.

Furthermore, it should be noted that the above-mentioned figures are only schematic illustrations of the processes involved in the methods according to exemplary embodiments of the present disclosure, and are not intended to be limiting. It will be readily understood that the processes shown in the above figures are not intended to indicate or limit the chronological order of the processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, e.g., in multiple modules.

The following are embodiments of the disclosed apparatus that may be used to perform embodiments of the disclosed methods. For details not disclosed in the embodiments of the apparatus of the present disclosure, refer to the embodiments of the method of the present disclosure.

Fig. 4 is a block diagram illustrating an information push apparatus based on a user lifecycle according to another exemplary embodiment. As shown in fig. 4, the information pushing apparatus 40 based on the user life cycle includes: information module 402, parameter module 404, calculation module 406, user module 408, and enterprise module 410.

The information module 402 is configured to obtain user information of a user, where the user information includes basic information and behavior information;

the parameter module 404 is configured to generate a plurality of core metrics and a plurality of original parameters based on the user information; the parameter module 404 is further configured to generate a repurchase rate index based on the accumulated consumption data in the user information; generating a growth rate indicator based on consumption data within a current stage in the user information; generating an aggressiveness index based on the plurality of categories in the user information and the amount corresponding thereto; generating a fund proportion index based on the asset transition data in the user information.

The calculation module 406 is configured to input the plurality of core indicators and the plurality of original parameters into a user life cycle model generated through long-term and short-term memory network training, so as to obtain a current stage of the user; the calculation module 406 includes: an input unit, configured to input the plurality of core indicators and the plurality of original parameters into the user lifecycle model, and generate corresponding probabilities of a plurality of unique hot codes; the analysis unit is used for analyzing a plurality of stages of the life cycle generated by the plurality of one-hot codes; a phase unit, configured to use a phase with a highest probability among the multiple phases as a current phase of the user.

User module 408 is configured to generate policy information and/or marketing information for the user based on the current stage.

The enterprise module 410 is used for calculating the current stages of a plurality of users through the user life cycle model; generating user guest group assessment information and/or user guest group structure information based on the current stage of the plurality of users.

Fig. 5 is a block diagram illustrating an information push apparatus based on a user lifecycle according to an exemplary embodiment. As shown in fig. 5, the information pushing apparatus 50 based on the user life cycle includes: a historical information module 502, a historical parameter module 504, a sample index module 506, a model training module 508, and a parameter tuning module 510.

The history information module 502 is configured to obtain a plurality of user information corresponding to a plurality of history users;

the historical parameter module 504 is configured to generate a plurality of sets of core indicators and a plurality of sets of original parameters based on the plurality of user information, where one set of core indicators and original parameters corresponds to one historical user;

a sample index module 506 for determining user lifecycle samples and stage indexes based on the sets of core indexes and the sets of original parameters; the sample metric module 506 includes: the clustering unit is used for carrying out clustering calculation on the multiple groups of core indexes and the multiple groups of original parameters to generate multiple historical user categories; a core unit, configured to extract a core user in each of the plurality of historical user categories; a sample unit for generating user lifecycle samples by a plurality of core users; the index unit is used for determining the stage indexes of all stages in the user life cycle based on the user information of a plurality of historical users and the user life cycle samples, and the stage indexes are used for judging the stage where the user is located by the user.

The model training module 508 is configured to input the multiple sets of core indicators, the multiple sets of original parameters, and the user lifecycle samples into a long-short term memory network for training, so as to generate a training result; the model training module 508 includes: the encoding unit is used for converting the multiple stages of the user life cycle sample into multiple one-hot codes; the training unit is used for inputting the multiple groups of core indexes, the multiple groups of original parameters and the multiple unique hot codes into a long-term and short-term memory network for training to generate a training result, wherein the training result is the multiple unique hot codes and the corresponding probabilities thereof; the long-term and short-term memory network is of a six-layer structure, and the training step length is five.

The parameter tuning module 510 is configured to perform parameter tuning on the training result to generate the user life cycle model. The parameter tuning module 510 includes: a function unit for generating an evaluation function; the function unit is also used for generating a cohesive function according to the occurrence frequency of each stage after the change of the life cycle stage of the historical user; generating a separation function according to the existing duration of each stage of the life cycle of the historical user; generating the evaluation function based on the cohesion function and the separation function. The tuning unit is used for tuning the multiple groups of core indexes based on the evaluation function; adjusting the stage indexes based on the evaluation function; and optimizing the user life cycle model based on the evaluation function.

According to the information pushing device based on the user life cycle, user information of a user is obtained, wherein the user information comprises basic information and behavior information; generating a plurality of core indicators and a plurality of original parameters based on the user information; inputting the plurality of core indexes and the plurality of original parameters into a user life cycle model generated through long-term and short-term memory network training to obtain the current stage of the user; based on the mode of generating strategy information and/or marketing information for the users in the current stage, refined and personalized marketing can be performed for the users with different life cycles, the operation cost is reduced, and the user value is improved.

An electronic device 600 according to this embodiment of the disclosure is described below with reference to fig. 6. The electronic device 600 shown in fig. 6 is only an example and should not bring any limitations to the function and scope of use of the embodiments of the present disclosure.

As shown in fig. 6, the electronic device 600 is embodied in the form of a general purpose computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 that connects the various system components (including the storage unit 620 and the processing unit 610), a display unit 640, and the like.

Wherein the storage unit stores program code that is executable by the processing unit 610 such that the processing unit 610 performs steps in accordance with various exemplary embodiments of the present disclosure in the present specification. For example, the processing unit 610 may perform the steps shown in fig. 2 and 3.

The storage unit 620 may include readable media in the form of volatile memory units, such as a random access memory unit (RAM)6201 and/or a cache memory unit 6202, and may further include a read-only memory unit (ROM) 6203.

The memory unit 620 may also include a program/utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.

Bus 630 may be one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

The electronic device 600 may also communicate with one or more external devices 600' (e.g., keyboard, pointing device, bluetooth device, etc.), such that a user can communicate with devices with which the electronic device 600 interacts, and/or any device (e.g., router, modem, etc.) with which the electronic device 600 can communicate with one or more other computing devices. Such communication may occur via an input/output (I/O) interface 650. Also, the electronic device 600 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network such as the Internet) via the network adapter 660. The network adapter 660 may communicate with other modules of the electronic device 600 via the bus 630. It should be appreciated that although not shown in the figures, other hardware and/or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein may be implemented by software, or by software in combination with necessary hardware. Therefore, as shown in fig. 7, the technical solution according to the embodiment of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a usb disk, a removable hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which may be a personal computer, a server, or a network device, etc.) to execute the above method according to the embodiment of the present disclosure.

The software product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, 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 portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The computer readable storage medium may include a propagated data signal with readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A readable storage medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a readable storage medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Program code for carrying out operations for the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., through the internet using an internet service provider).

The computer readable medium carries one or more programs which, when executed by a device, cause the computer readable medium to perform the functions of: acquiring user information of a user, wherein the user information comprises basic information and behavior information; generating a plurality of core indicators and a plurality of original parameters based on the user information; inputting the plurality of core indexes and the plurality of original parameters into a user life cycle model generated through long-term and short-term memory network training to obtain the current stage of the user; generating policy information and/or marketing information for the user based on the current stage.

Those skilled in the art will appreciate that the modules described above may be distributed in the apparatus according to the description of the embodiments, or may be modified accordingly in one or more apparatuses unique from the embodiments. The modules of the above embodiments may be combined into one module, or further split into multiple sub-modules.

Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein may be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a usb disk, a removable hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which may be a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.

Exemplary embodiments of the present disclosure are specifically illustrated and described above. It is to be understood that the present disclosure is not limited to the precise arrangements, instrumentalities, or instrumentalities described herein; on the contrary, the disclosure is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. An information push method based on a user life cycle is characterized by comprising the following steps:

acquiring user information of a user, wherein the user information comprises basic information and behavior information;

generating a plurality of core indicators and a plurality of original parameters based on the user information;

inputting the plurality of core indexes and the plurality of original parameters into a user life cycle model generated through long-term and short-term memory network training to obtain the current stage of the user;

generating policy information and/or marketing information for the user based on the current stage.

2. The information pushing method of claim 1, further comprising:

calculating current stages of a plurality of users through the user lifecycle model;

generating user guest group assessment information and/or user guest group structure information based on the current stage of the plurality of users.

3. The information pushing method of claim 1, further comprising:

acquiring a plurality of user information corresponding to a plurality of historical users;

generating a plurality of groups of core indexes and a plurality of groups of original parameters based on the plurality of user information, wherein one group of core indexes and original parameters correspond to one historical user;

determining user lifecycle samples and stage metrics based on the sets of core metrics and the sets of original parameters;

inputting the multiple groups of core indexes, the multiple groups of original parameters and the user life cycle samples into a long-term and short-term memory network for training to generate a training result;

and performing parameter optimization on the training result to generate the user life cycle model.

4. The information pushing method of claim 1, wherein generating a plurality of core metrics and a plurality of raw parameters based on the user information comprises:

generating a repurchase rate index based on accumulated consumption data in the user information;

generating a growth rate indicator based on consumption data within a current stage in the user information;

generating an aggressiveness index based on the plurality of categories in the user information and the amount corresponding thereto;

generating a fund proportion index based on the asset transition data in the user information.

5. The information pushing method of claim 1, wherein inputting the plurality of core indexes and the plurality of original parameters into a user life cycle model generated by long-short term memory network training to obtain a current stage of the user, comprises:

inputting the plurality of core indicators and the plurality of original parameters into the user lifecycle model to generate a plurality of one-hot coded corresponding probabilities thereof;

parsing the plurality of unique hot code generation life cycles;

and taking the stage with the highest probability in the plurality of stages as the current stage of the user.

6. The information push method of claim 3, wherein determining user lifecycle samples and stage metrics based on the sets of core metrics and the sets of raw parameters comprises:

performing cluster calculation on the multiple groups of core indexes and the multiple groups of original parameters to generate multiple historical user categories;

extracting core users in each of the plurality of historical user categories;

user lifecycle samples are generated by a plurality of core users.

7. The information push method of claim 6, wherein determining user lifecycle samples and stage metrics based on the sets of core metrics and the sets of raw parameters further comprises:

determining a stage index of each stage in the user life cycle based on the user information of a plurality of historical users and the user life cycle sample, wherein the stage index is used for judging the stage of the user.

8. The information push method according to claim 3, wherein inputting the plurality of sets of core indicators, the plurality of sets of original parameters, and the user life cycle samples into a long-short term memory network for training, and generating a training result comprises:

converting the plurality of phases of the user lifecycle sample into a plurality of one-hot codes;

inputting the multiple groups of core indexes, the multiple groups of original parameters and the multiple unique hot codes into a long-short term memory network for training to generate a training result, wherein the training result is the multiple unique hot codes and the corresponding probabilities thereof;

the long-term and short-term memory network is of a six-layer structure, and the training step length is five.

9. The information push method of claim 3, wherein performing parameter tuning on the training results to generate the user lifecycle model comprises:

generating an evaluation function;

adjusting and optimizing the multiple groups of core indexes based on the evaluation function;

adjusting the stage indexes based on the evaluation function;

and optimizing the user life cycle model based on the evaluation function.

10. The information pushing method of claim 9, wherein generating an evaluation function comprises:

generating a cohesion function according to the occurrence frequency of each stage after the change of the historical user life cycle stage;

generating a separation function according to the existing duration of each stage of the life cycle of the historical user;

generating the evaluation function based on the cohesion function and the separation function.

11. An information pushing apparatus based on user life cycle, comprising:

the information module is used for acquiring user information of a user, wherein the user information comprises basic information and behavior information;

a parameter module for generating a plurality of core indicators and a plurality of original parameters based on the user information;

the calculation module is used for inputting the plurality of core indexes and the plurality of original parameters into a user life cycle model generated through long-term and short-term memory network training to obtain the current stage of the user;

a user module to generate policy information and/or marketing information for the user based on the current stage.

12. The information push apparatus according to claim 11, further comprising:

an enterprise module for computing current phases of a plurality of users through the user lifecycle model; generating user guest group assessment information and/or user guest group structure information based on the current stage of the plurality of users.

13. The information push apparatus according to claim 11, further comprising:

the history information module is used for acquiring a plurality of user information corresponding to a plurality of history users;

the historical parameter module is used for generating a plurality of groups of core indexes and a plurality of groups of original parameters based on the plurality of user information, wherein one group of core indexes and original parameters correspond to one historical user;

a sample index module for determining user lifecycle samples and stage indexes based on the sets of core indexes and the sets of original parameters;

the model training module is used for inputting the plurality of groups of core indexes, the plurality of groups of original parameters and the user life cycle sample into a long-term and short-term memory network for training to generate a training result;

and the parameter tuning module is used for performing parameter tuning on the training result to generate the user life cycle model.

14. The information push apparatus according to claim 11, wherein the parameter module is further configured to

Generating a repurchase rate index based on accumulated consumption data in the user information; generating a growth rate indicator based on consumption data within a current stage in the user information; generating an aggressiveness index based on the plurality of categories in the user information and the amount corresponding thereto; generating a fund proportion index based on the asset transition data in the user information.

15. The information push apparatus according to claim 11, wherein the computing module includes:

an input unit, configured to input the plurality of core indicators and the plurality of original parameters into the user lifecycle model, and generate corresponding probabilities of a plurality of unique hot codes;

the analysis unit is used for analyzing a plurality of stages of the life cycle generated by the plurality of one-hot codes;

a phase unit, configured to use a phase with a highest probability among the multiple phases as a current phase of the user.

16. The information pushing apparatus of claim 13, wherein the sample metric module comprises:

the clustering unit is used for carrying out clustering calculation on the multiple groups of core indexes and the multiple groups of original parameters to generate multiple historical user categories;

a core unit, configured to extract a core user in each of the plurality of historical user categories;

a sample unit for generating user lifecycle samples by a plurality of core users.

17. The information push apparatus of claim 16, wherein the sample metric module further comprises:

the index unit is used for determining the stage indexes of all stages in the user life cycle based on the user information of a plurality of historical users and the user life cycle samples, and the stage indexes are used for judging the stage where the user is located by the user.

18. The information pushing apparatus of claim 13, wherein the model training module comprises:

the encoding unit is used for converting the multiple stages of the user life cycle sample into multiple one-hot codes;

the training unit is used for inputting the multiple groups of core indexes, the multiple groups of original parameters and the multiple unique hot codes into a long-term and short-term memory network for training to generate a training result, wherein the training result is the multiple unique hot codes and the corresponding probabilities thereof; the long-term and short-term memory network is of a six-layer structure, and the training step length is five.

19. The information pushing apparatus of claim 13, wherein the parameter tuning module comprises:

a function unit for generating an evaluation function;

the tuning unit is used for tuning the multiple groups of core indexes based on the evaluation function; adjusting the stage indexes based on the evaluation function; and optimizing the user life cycle model based on the evaluation function.

20. The information pushing apparatus of claim 19, wherein the function unit is further configured to

Generating a cohesion function according to the occurrence frequency of each stage after the change of the historical user life cycle stage; generating a separation function according to the existing duration of each stage of the life cycle of the historical user; generating the evaluation function based on the cohesion function and the separation function.

21. An electronic device, comprising:

one or more processors;

storage means for storing one or more programs;

when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-10.

22. A computer-readable medium, on which a computer program is stored, which, when being executed by a processor, carries out the method according to any one of claims 1-10.