CN114119094A

CN114119094A - Method and device for determining target user

Info

Publication number: CN114119094A
Application number: CN202111441483.5A
Authority: CN
Inventors: 王洵湉; 李旭锋
Original assignee: WeBank Co Ltd
Current assignee: WeBank Co Ltd
Priority date: 2021-11-30
Filing date: 2021-11-30
Publication date: 2022-03-01

Abstract

The embodiment of the invention provides a method and a device for determining a target user. The method comprises the following steps: performing characteristic analysis on user information of each historical user in the selected seed packet to determine the significant characteristics of the seed packet; determining a first algorithm tool of the seed packet from the corresponding relation between the salient features and the algorithm tools; performing model training on the first algorithm tool by taking the seed packets as samples to obtain a target model; for any existing user, determining, by the target model, whether the existing user is a target user that is in anticipation. The finally determined target users are not limited to the historical users in the seed packets. If the target user is determined according to different activities, the selected seed packet is only required to be adjusted, and the method for determining the target user is more flexible. The historical users in the seed packets are modeled through the preset first algorithm tool, so that the participation of algorithm personnel is not needed, the prediction accuracy is high, and meanwhile, the manpower is saved.

Description

Method and device for determining target user

Technical Field

The embodiment of the invention relates to the technical field of financial science and technology, in particular to a method, a device, computing equipment and a computer-readable storage medium for determining a target user.

Background

With the development of computer technology, more and more technologies are applied in the financial field, and the traditional financial industry is gradually changing to financial technology (Fintech), but due to the requirements of the financial industry on safety and real-time performance, higher requirements are also put forward on the technologies.

When the operator promotes activities, the operator often cannot push activity messages for all registered users, so that the push pertinence is not strong, and the push effect is poor. The operator will filter the target user population based on experience to push the campaign messages on a targeted basis.

The existing operator pushing scheme is that a candidate user group is determined based on experience, information is pushed to the candidate user group, users with expected effect evaluation are recorded as target user groups, and then the target user groups can be selected to push related information in subsequent similar popularization activities. For example, the operator determines that the candidate user group is users with more than 10 ten thousand deposits on the platform, pushes a message a to the users, the message a can be entitlement commodities, preferential information and the like, records user feedback, and selects the user who receives the entitlement commodities or places an order through the preferential information as a target user group. Then in the subsequent activities related to pushing the interest commodities and the discount information, the activity information is still pushed for the target activity groups.

It can be found that the target activity groups determined by the method are limited to more than 10 ten thousand users, the determined target activity groups are the same for different activities, the target activity groups are touched for multiple times, and the touch probability of other users is low. The method of determining the target activity population is not flexible.

To sum up, the embodiments of the present invention provide a method for determining a target user, so as to flexibly determine target users in different activities and improve the accuracy of determining the target users.

Disclosure of Invention

The embodiment of the invention provides a method for determining a target user, which is used for flexibly determining the target user in different activities and improving the accuracy of determining the target user.

In a first aspect, an embodiment of the present invention provides a method for determining a target user, including:

performing characteristic analysis on user information of each historical user in the selected seed packet to determine the significant characteristics of the seed packet; each historical user is a user with a result label of a set target after a historical event; the set target is either in anticipation or out of anticipation;

determining a first algorithm tool of the seed packet from the corresponding relation between the salient features and the algorithm tools;

performing model training on the first algorithm tool by taking the seed packets as samples to obtain a target model; the target model is used for determining the probability that the user meets a set target through the user information;

for any existing user, determining, by the target model, whether the existing user is a target user that is in anticipation.

The seed packets of various historical events selected by service personnel based on requirements comprise various historical users having result labels of set targets after the historical events are carried out, and as various historical users in the seed packets have the set targets, the obvious characteristics of various historical users in the seed packets can be determined according to the user information of various historical users. According to the corresponding relation between the obvious features and the algorithm tool, the first algorithm tool is conveniently determined, so that the feature extraction can be rapidly carried out on the user information in the seed packet, and thus, whether the user information is a target user can be judged for any existing user. The finally determined target users are not limited to the historical users in the seed packets. If the target user is determined according to different activities, the selected seed packet is only required to be adjusted, and the method for determining the target user is more flexible. The historical users in the seed packets are modeled through the preset first algorithm tool, so that the participation of algorithm personnel is not needed, the prediction accuracy is high, and meanwhile, the manpower is saved. The corresponding relation between each significant feature and the algorithm tool is preset, so that the operation is simple, convenient and quick, and the efficiency is improved.

Optionally, the seed packet includes at least a first seed packet and a second seed packet; the historical users in the first sub-packet are all users with result labels meeting expectations after the historical events; the historical users in the second seed packet are all users with results which are not in accordance with expectations after the historical events;

for any existing user, determining whether the existing user is a target user that is expected through the target model, including:

for any existing user, determining a first probability that the existing user meets expectations through a target model of a first seed packet; determining a second probability that the existing user is not in agreement with expectations by a target model of a second seed packet;

determining whether the existing user is a target user that is expected according to the first probability and the second probability.

The first sub-packet containing the result labels of all the users which are expected to be met and the second sub-packet containing the result labels of all the users which are not expected to be met are selected, more user information of historical users is integrated, the probability of whether the existing users are the target users or not is determined from two dimensions, and the possible intention of the existing users can be reflected more accurately. Thereby improving the accuracy of determining the target user.

Optionally, the seed packet includes at least a first seed packet and a second seed packet;

each historical user in the first sub-packet and each historical user in the second sub-packet go through different historical events; or

The result labels of the historical users in the first sub-packet are in accordance with the expectation, and the result labels of the historical users in the second sub-packet are not in accordance with the expectation.

And various types of seed packages are selected according to requirements, more information is integrated, and the accuracy of determining the target user is improved.

Optionally, determining the salient features of the seed packet includes:

and carrying out feature analysis on the user information of each historical user in the seed packet based on any selected feature mining tool, and determining the significant features of the seed packet.

And more choices are given to business personnel, and the method is not limited to the characteristic mining tool preset by the system. The service personnel can select the feature mining tool as required. The feature mining process is more flexible, and the efficiency is improved.

Optionally, the method further comprises:

and if the corresponding relation between the salient features and the algorithm tool does not contain the salient features of the seed packet, returning to the step of determining the salient features of the seed packet, and selecting the features meeting set conditions from the remaining features as the salient features until the corresponding relation between the salient features and the algorithm tool contains the salient features of the seed packet.

And if the corresponding relation between the significant features and the algorithm tool does not contain the significant features of the seed packets, updating the significant features until the target algorithm tool is determined, so that manual selection operation is not needed, the labor is saved, and the method is convenient and quick.

Optionally, with the seed packet as a sample, performing model training on the first algorithm tool to obtain a target model, including:

performing model training on the first algorithm tool by taking the seed packet as a sample to obtain a first optional model;

performing model training on one or more second algorithm tools selected from an algorithm tool library by taking the seed packets as samples to obtain one or more second selectable models;

inputting user information of each test user in the test set and a result label of each test user into the first selectable model and the second selectable model, and determining the accuracy of the first selectable model and the second selectable model; and selecting and obtaining the target model based on the accuracy of the first selectable model and the second selectable model.

The target model can be determined by not only modeling based on a preset first algorithm tool, but also freely selecting a second algorithm tool to model and comparing the accuracy of each obtained model. More choices are provided for business personnel, algorithm tools can be selected quickly, accuracy can be verified quickly, and a target model can be found based on comparison of the accuracy. The accuracy and flexibility of determining the target model are improved.

Optionally, after determining whether the existing user is a target user that meets expectations, the method further includes:

and determining push information for the target user according to the historical behavior of the expected historical user, wherein the result label in the seed packet conforms to the historical behavior of the target user, and pushing the push information to the target user.

This may increase the likelihood that the existing user will respond to the historical behavior.

In a second aspect, an embodiment of the present invention further provides an apparatus for determining a target user, including:

a determination unit configured to:

a processing unit to:

the processing unit is specifically configured to:

Optionally, the determining unit is specifically configured to:

Optionally, the determining unit is further configured to:

Optionally, the processing unit is specifically configured to:

Optionally, the processing unit is further configured to:

In a third aspect, an embodiment of the present invention further provides a computing device, including:

a memory for storing a computer program;

and the processor is used for calling the computer program stored in the memory and executing the method for determining the target user listed in any mode according to the obtained program.

In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, where a computer-executable program is stored, where the computer-executable program is configured to cause a computer to execute a method for determining a target user, which is listed in any of the above manners.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

Fig. 1 is a schematic diagram of a system architecture according to an embodiment of the present invention;

fig. 2 is a schematic flowchart of a possible method for determining a target user according to an embodiment of the present invention;

FIG. 3 is a schematic interface diagram of a possible intelligent modeling system 300 according to an embodiment of the present invention;

fig. 4 is a schematic diagram of a specific application scenario provided in the embodiment of the present invention;

fig. 5 is a flowchart illustrating a method for determining a target user according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of a computer device according to an embodiment of the present invention.

Detailed Description

To make the objects, embodiments and advantages of the present application clearer, the following description of exemplary embodiments of the present application will clearly and completely describe the exemplary embodiments of the present application with reference to the accompanying drawings in the exemplary embodiments of the present application, and it is to be understood that the described exemplary embodiments are only a part of the embodiments of the present application, and not all of the embodiments.

All other embodiments, which can be derived by a person skilled in the art from the exemplary embodiments described herein without inventive step, are intended to be within the scope of the claims appended hereto. In addition, while the disclosure herein has been presented in terms of one or more exemplary examples, it should be appreciated that aspects of the disclosure may be implemented solely as a complete embodiment.

It should be noted that the brief descriptions of the terms in the present application are only for the convenience of understanding the embodiments described below, and are not intended to limit the embodiments of the present application. These terms should be understood in their ordinary and customary meaning unless otherwise indicated.

The terms "first," "second," "third," and the like in the description and claims of this application and in the above-described drawings are used for distinguishing between similar or analogous objects or entities and are not necessarily intended to limit the order or sequence of any particular one, Unless otherwise indicated. It is to be understood that the terms so used are interchangeable under appropriate circumstances such that the embodiments described herein are, for example, capable of operation in sequences other than those illustrated or otherwise described herein.

Furthermore, the terms "comprises" and "comprising," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a product or device that comprises a list of elements is not necessarily limited to those elements explicitly listed, but may include other elements not expressly listed or inherent to such product or device.

Fig. 1 illustrates an exemplary system architecture, which may be a server 100, including a processor 110, a communication interface 120, and a memory 130, to which embodiments of the present invention are applicable.

The communication interface 120 is used for communicating with a terminal device, and transceiving information transmitted by the terminal device to implement communication.

The processor 110 is a control center of the server 100, connects various parts of the entire server 100 using various interfaces and routes, performs various functions of the server 100 and processes data by operating or executing software programs and/or modules stored in the memory 130 and calling data stored in the memory 130. Alternatively, processor 110 may include one or more processing units.

The memory 130 may be used to store software programs and modules, and the processor 110 executes various functional applications and data processing by operating the software programs and modules stored in the memory 130. The memory 130 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to a business process, and the like. Further, the memory 130 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device.

It should be noted that the structure shown in fig. 1 is only an example, and the embodiment of the present invention is not limited thereto.

The server shown in fig. 1 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a Network service, cloud communication, a middleware service, a domain name service, a security service, a Content Delivery Network (CDN), a big data and artificial intelligence platform, and the like.

An embodiment of the present invention provides a possible method for determining a target user, as shown in fig. 2, including:

step 201, performing characteristic analysis on user information of each historical user in the selected seed packet to determine the significant characteristics of the seed packet; each historical user is a user with a result label of a set target after a historical event; the set target is either in anticipation or out of anticipation.

Step 202, determining a first algorithm tool of the seed packet from the corresponding relationship between the salient features and the algorithm tools.

Step 203, performing model training on the first algorithm tool by taking the seed packet as a sample to obtain a target model; the target model is used for determining the probability that the user meets the set target through the user information.

Step 204, for any existing user, determining whether the existing user is a target user meeting expectations through the target model.

Fig. 3 shows an interface schematic diagram of a possible intelligent modeling system 300, which includes a seed packet 310, a feature mining tool library 320, and an algorithm tool library 330. When business personnel need to select a target user for a certain activity to push a message, a seed packet, a feature mining tool, an algorithm tool and the like can be selected according to requirements in an intelligent modeling system. And finally, modeling the data in the seed packets selected by the service personnel based on the algorithm tool selected by the service personnel or the algorithm tool preset by the system, wherein the obtained model can be used for determining the target user.

The intelligent system can store a plurality of seed packets of the same historical event with different set targets, the same set target of different historical events or different historical events and different set targets, wherein the historical users in each seed packet are the same set target of the same historical event. For example, in seed packet 1, a user who successfully promoted in 5-month account opening recommendation activity (the historical event is 5-month account opening recommendation activity, and the set target is according to expectation), in seed packet 2, a user who successfully promoted in 4-month drowsy client bailing activity (the historical event is 4-month drowsy client bailing activity, and the set target is according to expectation), in seed packet 3, a user who unsuccessfully promoted in 5-month order call charge activity (the historical event is 5-month order call charge activity, and the set target is not according to expectation), and in seed packet 4, a user who successfully promoted in 5-month order call charge activity (the historical event is 5-month order call charge activity, and the set target is according to expectation). The historical users which accord with the same historical event and the same set target are gathered in the seed packet, so that service personnel can select the seed packet conveniently based on needs, and feature extraction is carried out according to the user information of the historical users in the seed packet, namely the obvious features of the historical users in the seed packet are determined.

When the service personnel selects the seed packet, one seed packet can be selected, and a plurality of seed packets can also be selected. For example, seed packages 1 and 3 are selected. Wherein the seed packet 1 is a seed packet with a set target meeting the expectation, namely a positive seed packet; the seed packet 3 is a packet of seeds that is targeted to be undesirable, i.e., negative seeds. And performing feature extraction on the positive seed packet and the negative seed packet, and analyzing the significant features of the historical users which are in line with the expectation and the significant features of the historical users which are not in line with the expectation so as to guide the subsequent modeling process.

Each seed packet contains user information of each historical user participating in the historical event, such as user ID, gender, age, province, school calendar and the like. As shown in table 1.

TABLE 1

User ID	Sex	Age (age)	Province of labor	Study calendar
					A	Woman	23	Shanghai province	This section
B	For male	25	Guangdong (Chinese character of Guangdong)	This section
					C	Woman	23	Shanghai province	Master's soldier
D	Woman	25	Guangdong (Chinese character of Guangdong)	Doctor (Rooibos)

The system performs characteristic analysis on the user information of the historical users in the selected sub-packets to determine the significant characteristics of the sub-packets.

A method for determining the salient features is described by taking the example of performing feature analysis on the user information of the historical users in the seed bag 1. A plurality of feature mining tools, such as LR (Linear Regression), FM (factor decomposition Machine), GBDT (Gradient Boosting Decision Tree), DT (Decision Tree), and the like, are set in a feature mining tool library at an interface of the intelligent modeling system. The service personnel can select the feature digging tool at the interface of the intelligent modeling system, and the feature digging tool can also be preset for each seed packet by the intelligent modeling system. The embodiments of the present invention are not limited in this regard.

For example, the feature mining tool selected by the business personnel is a decision tree, and the feature mining tool has strong interpretability. And determining the information value of any characteristic of each historical user in the seed packet by using the characteristic mining tool, and then selecting one or more characteristics as the significant characteristics according to the information values of the multiple characteristics. Specifically, the information value of any feature is obtained by the following method, including: and aiming at any characteristic value in the characteristics, calculating a first distribution proportion of the characteristic value in the seed packet and a second distribution proportion of the characteristic value in the overall users, and determining the information value of the characteristics according to the first distribution proportion and the second distribution proportion of a plurality of characteristic values in the characteristics.

For example, if the user information in the seed packet 1 is shown in table 1, the feature "age" includes two feature values of 23 years and 25 years. Then, the distribution ratio of the 23 year old users in the seed packet and the distribution ratio of the 23 year old users in the overall users are calculated first (the overall users may be registered users of the platform or selected users of some type, such as more than 10 ten thousands of deposited users, which is not limited in the embodiment of the present invention), and the distribution ratio of the 25 year old users in the seed packet and the distribution ratio of the 25 year old users in the overall users are calculated (only two ages are calculated because there are only two characteristic values of 23 years and 25 years in the users in the seed packet 1).

The IV (Information Value) of the feature "age" is obtained by using the formula (1). Wherein i is a characteristic value.

IV ═ Σ (seed packet distribution i-total user distribution i) × ln (seed packet distribution i/total user distribution i) (1)

The information values of the features such as the province, the age and the like can be obtained in the same way. The larger the information value, the more prominent the feature is in the seed package 1 as compared to the overall user, which reflects the uniqueness of the user in the seed package. And sorting the information values, and selecting one or more features with the largest information values as the significant features.

If the information value sequence of each characteristic is determined as follows: age, province, school calendar, sex. The first one may be selected as the salient feature or the first 3 may be selected as the salient features without limitation.

Next, an algorithm tool is determined according to the selected salient features. The embodiment of the invention provides three modes for determining an algorithm tool.

In the first mode, a service person selects a second algorithm tool in an algorithm tool library at an interface of the intelligent modeling system according to the requirement.

The algorithm tool library at the interface of the intelligent modeling system is provided with various algorithm tools, such as LR (Linear Regression), FM (factor Machine), GBDT (Gradient Boosting Decision Tree), DT (Decision Tree), and the like. The service personnel can freely choose.

For example, in the previous step "age" was determined as a significant feature, in which the service person chose FM as the second algorithmic tool.

And secondly, the system presets the corresponding relation between the salient features and the algorithm tool, and the algorithm tool can be directly determined after the salient features are determined.

For example, "age" is determined as a significant feature in the previous step, and in this step, the first algorithm tool corresponding to the significant feature of the seed package 1 is determined to be LR in the correspondence relationship between the significant feature and the algorithm tools.

In the corresponding relation between the salient features and the algorithm tool, if the salient features determined at this time do not exist, the step of determining the salient features of the seed packet is returned, and the features meeting the set conditions are selected from the remaining features as the salient features until the corresponding relation between the salient features and the algorithm tool contains the salient features of the seed packet. For example, when it is determined that there is no significant feature "age" in the correspondence of the significant features to the algorithmic tools, return is made to re-determine the significant features: sorting the information values of the features determined in the previous step: age, province, school calendar, gender, the second ranked feature "province" is selected as the salient feature. Until the salient feature is included in its correspondence with the algorithmic means.

And thirdly, using an algorithm tool preset by the system and an algorithm tool selected by the service personnel for training the seed packet, and selecting the optimal algorithm tool.

Before describing the equation three, how to model by an algorithm tool needs to be described first, so as to obtain a target model.

The modeling mode is that the selected seed packet is used as a sample, the user information of each historical user in the seed packet and the result labels of each historical user are input into an algorithm tool, and model training is carried out until a target model is obtained.

For example, for the historical users in seed bag 1, the LR algorithm tool is used for model training. The model formula is formula (2). Wherein xi is the value of each feature, ω i is the weight corresponding to the value of each feature, i.e. the variable to be calculated in the model, and y' is the predicted value to be calculated.

h(x)＝ω1x1+ω2x2+…+ωnxn+b (2)

y’＝1/(1+e-h(x)) (3)

Accordingly, the loss function is:

y_ii.e. the real value, y 'of the ith sample'_iNamely, the predicted value obtained by substituting the characteristic value of the ith sample into the formulas (2) and (3). And substituting the user information of all the historical users in the seed packet 1 into the formulas (2), (3) and (4) to calculate the ω i, so that the loss function L is minimum. The manner of updating ω i may employ a gradient descent method, i.e.:

wherein, ω is_tFor the t-th iteration, the corresponding ω value, α, is the learning rate, i.e., the learning sample speed, and is set empirically, here to be 0.01, ω₀The value is assigned randomly. I.e. first to ω₀And (4) randomly assigning values, substituting all samples, namely the steps (1) and (2) into the formulas (1) and (2) and (3), calculating the value L, substituting the value into the formula (4) according to the value, updating omega, and repeating the iteration until the value L is basically not updated, so that the model training is finished. ω obtained at this time is a weight corresponding to the value of each feature in the formula (1), and can reflect the importance of the value of each feature.

For example, x1 is 18 years old, if so, x1 is 1, if not, x1 is 0; x2 is 19 years old, if yes, x2 is 1, if no, x2 is 0; … … x6 is 23 years old, if yes, x6 is 1, if no, x6 is 0; … … x8 is 25 years old, if yes, x8 is 1, if no, x8 is 0; … … x9 is female, if yes, x9 is 1, if no, x9 is 0; x10 is the guangdong, if so, x10 is 1, and if not, x10 is 0 … …, and the value of each feature can be represented by 1 or 0. The above is merely an example, xi may also represent a range, such as whether 18-30 years old, etc., and embodiments of the present invention are not limited in this respect.

First, ω in the formula (2) is preset_tFor each historical user in the seed packet 1 shown in table 1, the value of a is substituted into the formula (2) for a to obtain the predicted value y 'in (3)'₁(ii) a Substituting the value of B into the formula (2) to obtain the predicted value y 'in the formula (3)'₂(ii) a Substituting the value of C into formula (2) to obtain the predicted value y 'in (3)'₃(ii) a Substituting the value of D into formula (2) to obtain the predicted value y 'in (3)'₄Substituting the obtained predicted value into a formula (4) to obtain a loss function L1; update ω according to equation (5)_tThe above steps are repeated to obtain a loss function L2 … … until the loss function value is no longer updated, at which point ω is_tThe values of (a) are the parameters of the trained target model. One possible determined model parameter is shown in table 2.

TABLE 2

Next, the third mode mentioned above will be described.

And modeling by using the seed packets as samples and adopting a first algorithm tool preset by the system to obtain a first optional model. The modeling approach is referred to above.

And selecting one or more second algorithm tools from the algorithm tool library by the service personnel for model training by taking the seed packet as a sample to obtain one or more second selectable models.

And inputting the user information of each test user in the test set and the result label of each test user into the first selectable model and the second selectable model, and determining the accuracy of the first selectable model and the second selectable model. And determining the model with the accuracy rate meeting the preset threshold value or the highest accuracy rate as the final target model.

One specific application scenario is shown in fig. 4. Comprises the following steps:

step 401, the system determines salient features.

And step 402, searching for an algorithm tool corresponding to the salient features in the corresponding relation between the salient features and the algorithm tool.

Step 403, whether the salient features exist or not; if yes, go to step 404; if not, returning to step 401;

step 404, determining a first algorithm tool;

step 405, the system pops up a request to the service personnel: whether to use the first algorithm tool;

and 406, performing model training by using the first algorithm tool to obtain a first optional model.

Step 407, the service personnel selects one or more second algorithm tools from the algorithm tool library.

Step 408, the system pops up a request to the service personnel: whether to use the first alternative model.

Step 409, one or more second selectable models are obtained.

And step 410, determining a target model in the first optional model and/or the second optional model according to the accuracy.

And step 411, model application.

After the system determines the significant features, searching for the algorithm tool corresponding to the significant features in the corresponding relationship between the significant features and the algorithm tool, if the significant features do not exist, returning to re-determine the significant features until the significant features exist in the corresponding relationship between the significant features and the algorithm tool, and then determining the first algorithm tool corresponding to the significant features. At this time, the system can pop up a request to the service personnel: whether the first algorithm tool is used or not, if so, the first algorithm tool is used for model training to obtain a first selectable model; and if the business personnel select no, popping up an algorithm tool selection interface for the business personnel to select an algorithm tool. The business person may select one or more second algorithm tools. The system derives one or more second selectable models according to one or more second algorithmic tools. And determining the target model according to the accuracy of the second optional model. And finally applying the model.

After the system obtains the first optional model, the system may also pop a request to the service personnel: if the first selectable model is not used, the model is applied if the business person selection is yes, and if the business person selection is not, step 407 is entered, and the business person selects one or more second algorithm tools again. The system derives one or more second selectable models according to one or more second algorithmic tools. And determining the target model according to the accuracy of the first optional model and the second optional model. And finally applying the model.

Optionally, a plurality of salient features may be determined, a plurality of first algorithm tools may be determined in the correspondence between the salient features and the algorithm tools, so that a plurality of target models may be obtained, and the accuracy of the plurality of target models may be tested, thereby determining a final target model.

Alternatively, a plurality of salient features may be determined, and an algorithm tool corresponding to a combination of the plurality of salient features is stored in the correspondence relationship between the salient features and the algorithm tool, so that one or more corresponding algorithm tools may be determined.

Therefore, business personnel can quickly verify the effect based on multiple models, update the models or adjust parameters in time, and find the optimal target model after comparing the effects of the multiple models, so that the accuracy of determining the target user is improved.

The process of determining the target model for the seed packet 1 is described above, and the process may also be performed for the seed packet 3 selected by the service person, so as to obtain the respective target models.

For any existing user, the user information (e.g., woman, 23 years old, etc.) of the existing user is input into the target model of seed packet 1, resulting in a first probability that the existing user will fit the expectation. For example, the first probability of 0.6 is output.

User information (e.g., female, 23 years old, etc.) for the existing user is entered into the target model for seed packet 2, resulting in a second probability that the existing user is not in compliance with expectations. For example, the second probability of 0.2 is output.

Determining whether the existing user is a target user that is expected according to the first probability and the second probability. Subtracting the second probability from the first probability to obtain the probability of the user meeting the expectation as 0.4.

And if the probability of meeting the expectation is greater than a preset threshold value, the existing user is the target user.

Optionally, one or more seed packets meeting expectations can be selected to perform the method flow; one or more seed packages which are not in accordance with the expectation can be selected to carry out the method flow; the method may further include obtaining one or more target models by selecting one or more seed packets that meet expectations, and then obtaining one or more first probabilities that the existing user meets expectations, obtaining one or more target models by selecting one or more seed packets that do not meet expectations, and then obtaining one or more second probabilities that the existing user does not meet expectations, and subtracting the sum of the second probabilities from the sum of the first probabilities, and then obtaining a probability that the existing user meets expectations. The embodiments of the present invention are not limited in this regard. For example, if a service person selects the seed packet 3, the system may extract significant features from users in the seed packet 3, determine a first algorithm tool, and model historical users in the seed packet 3 by using the first algorithm tool, and since all the historical users in the seed packet 3 are users who are not in line with expectations, the obtained target model may be used to determine the probability that any existing user is not in line with expectations, and further know the probability that the existing user is in line with expectations, i.e., determine whether the existing user is a target user capable of pushing messages. For example, if it is determined that any existing user is not expected with a probability of 0.2, it is determined that it is expected with a probability of 1-0.2 ═ 0.8.

For example, after two seed packets that are expected and two seed packets that are not expected are selected for model training to obtain 4 target models, the first probabilities of an existing user are calculated to be 0.8 and 0.5, respectively, and the second probabilities are 0.4 and 0.3, respectively, so that the expected probability of the existing user is (0.8+0.5) - (0.4+0.3) ═ 0.6.

Optionally, the determined target user may be guided to the next push action. That is, after determining whether the existing user is the target user that meets the expectation, the method further includes:

For example, it is also stored in the seed bag 1 that the historical behavior performed on the historical user in the historical event is sending a preferential short message, and it is also stored in the seed bag 2 that the historical behavior performed on the historical user in the historical event is giving a three-month video interest. For an existing user, the first probability determined with the object model of seed bag 1 is 0.6 and the first probability determined with the object model of seed bag 2 is 0.2. Then the historical behavior of seed packet 1 may be suggested for this existing user: and sending the preferential short message.

Since the existing user is more consistent with the user information of the historical user in the seed packet 1, the historical behavior of the seed packet 1 is implemented for the existing user, so that the possibility that the existing user responds to the historical behavior can be improved.

Based on the same technical concept, fig. 5 exemplarily shows a structure of a target user determining device according to an embodiment of the present invention, where the structure can perform a process of determining a target user.

As shown in fig. 5, the apparatus specifically includes:

a determining unit 501, configured to:

a processing unit 502 for:

the processing unit 502 is specifically configured to:

Optionally, the determining unit 501 is specifically configured to:

Optionally, the determining unit 501 is further configured to:

Optionally, the processing unit 502 is specifically configured to:

Optionally, the processing unit 502 is further configured to:

Based on the same technical concept, the embodiment of the present application provides a computer device, as shown in fig. 6, including at least one processor 601 and a memory 602 connected to the at least one processor, where a specific connection medium between the processor 601 and the memory 602 is not limited in the embodiment of the present application, and the processor 601 and the memory 602 are connected through a bus in fig. 6 as an example. The bus may be divided into an address bus, a data bus, a control bus, etc.

In the embodiment of the present application, the memory 602 stores instructions executable by the at least one processor 601, and the at least one processor 601 may execute the steps of the method for determining the target user by executing the instructions stored in the memory 602.

The processor 601 is a control center of the computer device, and may connect various parts of the computer device by using various interfaces and lines, and perform the determination of the target user by executing or executing the instructions stored in the memory 602 and calling the data stored in the memory 602. Optionally, the processor 601 may include one or more processing units, and the processor 601 may integrate an application processor and a modem processor, wherein the application processor mainly handles an operating system, a user interface, an application program, and the like, and the modem processor mainly handles wireless communication. It will be appreciated that the modem processor described above may not be integrated into the processor 601. In some embodiments, the processor 601 and the memory 602 may be implemented on the same chip, or in some embodiments, they may be implemented separately on separate chips.

The processor 601 may be a general-purpose processor, such as a Central Processing Unit (CPU), a digital signal processor, an Application Specific Integrated Circuit (ASIC), a field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof, configured to implement or perform the methods, steps, and logic blocks disclosed in the embodiments of the present Application. A general purpose processor may be a microprocessor or any conventional processor or the like. The steps of a method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware processor, or may be implemented by a combination of hardware and software modules in a processor.

The memory 602, which is a non-volatile computer-readable storage medium, may be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. The Memory 602 may include at least one type of storage medium, and may include, for example, a flash Memory, a hard disk, a multimedia card, a card-type Memory, a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Programmable Read Only Memory (PROM), a Read Only Memory (ROM), a charge Erasable Programmable Read Only Memory (EEPROM), a magnetic Memory, a magnetic disk, an optical disk, and so on. The memory 602 is any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to such. The memory 602 in the embodiments of the present application may also be circuitry or any other device capable of performing a storage function for storing program instructions and/or data.

Based on the same technical concept, embodiments of the present invention further provide a computer-readable storage medium, where a computer-executable program is stored, and the computer-executable program is configured to enable a computer to execute the method for determining a target user listed in any of the above manners.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.

Claims

1. A method for identifying a target user, comprising:

2. The method of claim 1, wherein the seed packets include one or more first seed packets and one or more second seed packets; the historical users in the first sub-packet are all users with result labels meeting expectations after the historical events; the historical users in the second seed packet are all users with results which are not in accordance with expectations after the historical events;

for any existing user, determining one or more first probabilities that the existing user meets expectations through one or more target models corresponding to the one or more first seed packages; determining one or more second probabilities that the existing user is not in agreement with expectations through one or more objective models corresponding to the one or more second seed packages;

determining whether the existing user is a target user that is expected according to the one or more first probabilities and the one or more second probabilities.

3. The method of claim 1, wherein the seed packet comprises at least a first seed packet and a second seed packet;

4. The method of claim 1, wherein performing a feature analysis on the user information of each historical user in the selected seed packet to determine the salient features of the seed packet comprises:

determining the information value of any characteristic of each historical user in the seed packet based on the selected characteristic mining tool;

selecting one or more characteristics as significant characteristics according to the information values of the characteristics;

wherein, the information value of any characteristic is obtained by the following method, including:

for any characteristic value in the characteristics, calculating a first distribution proportion of the characteristic value in a seed packet and a second distribution proportion of the characteristic value in the overall users;

and determining the information value of the characteristic according to the first distribution proportion and the second distribution proportion of a plurality of characteristic values in the characteristic.

5. The method of claim 1, further comprising:

6. The method of claim 1, wherein model training the first algorithm tool using the seed packet as a sample to obtain a target model comprises:

7. The method of any of claims 1-6, wherein after determining whether the existing user is an anticipated target user, further comprising:

8. An apparatus for determining a target user, comprising:

a determination unit configured to:

a processing unit to:

9. A computing device, comprising:

a memory for storing a computer program;

a processor for calling a computer program stored in said memory, for executing the method of any one of claims 1 to 7 in accordance with the obtained program.

10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer-executable program for causing a computer to execute the method of any one of claims 1 to 7.