CA3138745A1 - Method and device for circling and creating target population - Google Patents

Abstract

The present invention makes public and provides a method of and a device for circling target population, and a method of and a device for creating a target population circling model, of which the method of circling target population comprises: obtaining a target brand category to be tested, marketing information, a marketing channel and a marketing effect indicator; obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set; and inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user identifications sequenced according to marketing effect indicator scores.

Description

METHOD AND DEVICE FOR CIRCLING AND CREATING TARGET POPULATION
BACKGROUND OF THE INVENTION
Technical Field [0001] The present invention relates to the field of computer information processing technology, and more particularly to a method of and a device for circling target population, and a method of and a device for creating a target population circling model.
Description of Related Art

[0002] Population application (products recommendation carried out through advertisements, short messages and information pushing, etc.) currently employed by brand owners under e-commerce environment is usually directed to potential customers according to professions, and recommendation is made mainly by directly selecting profession categories to which the brands pertain. However, employment of the method directed to potential customers according to professions is defective in the returned data being unable to reach the brands, lacks precision, and cannot satisfy the requirement of brand owners for fine operation.

[0003] In addition, taking for example the contents made public in the Chinese Patent with the publication number as CN110517070A and the patent title as Method of and Device for Circling Consumer Population, it obtains a first population collection and a second population collection through retrieval of keywords, then obtains associated information according to attribute information, and hence obtains consumer population relevant to the consumer circling after circling logics have been determined. Therefore, this patent is based on the relevancy of attributes of population collections to circle the consumer population, but attributes (such as professions, preferences, etc.) of the population tend to have deviations, thus causing imprecise subsequent circling.

Date Recue/Date Received 2021-11-12

[0004] Accordingly, there is an urgent need to search for a method of circling potential consumers that can satisfy fine operation of brands and is relatively high in precision.
SUMMARY OF THE INVENTION

[0005] In order to solve the aforementioned technical problems, the present invention provides a method of and a device for circling target population, and a method of and a device for creating a target population circling model capable of circling potentially purchasing population with respect to brands, and achieving high precision in the circling result.
Technical solutions provided by the present invention are described below.

[0006] According to the first aspect, there is provided a method of circling target population, and the method at least comprises the following steps:

[0007] obtaining a target brand category to be tested, marketing information, a marketing channel and a marketing effect indicator;

[0008] obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set; and

[0009] inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user IDs sequenced according to marketing effect indicator scores.

[0010] In a preferred embodiment, the step of obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set includes the following sub-steps:

[0011] obtaining previous marketing information of the target brand category to be tested on a Date Recue/Date Received 2021-11-12 target platform within a first preset time period, and extracting user IDs and marketing effect in the previous marketing information;

[0012] selecting a set of users with marketing effect is within a preset threshold as an initial test sample set; and

[0013] taking the initial test sample set as basis and combining therewith received label customization circling information to obtain the prediction sample set.

[0014] In a preferred embodiment, the method further comprises: pre-creating a target population circling model, wherein the step includes the following sub-steps:

[0015] extracting a first positive sample set and a first negative sample set from the target platform according to the target brand category to be tested, wherein the first positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a second preset time period and sample user features corresponding thereto, and the first negative sample set contains sample users randomly selected from a date on which users of the first positive sample set did not purchase products relevant to the target brand category within the second preset time period and sample user features corresponding thereto;

[0016] respectively training n models via preset n algorithms with sample user features in the first positive sample set and the first negative sample set as input, with marketing effect indicator scores as output, where n>2; and

[0017] determining the model with the highest accuracy rate from the n models as the target population circling model.

[0018] In a preferred embodiment, prior to inputting the prediction sample set into a pre-created target population circling model, the method further comprises real-time model validation, including:

[0019] extracting a second positive sample set and a second negative sample set from the target platform according to the target brand category to be tested, wherein the second positive sample set contains all sample users having purchased products relevant to the target brand Date Recue/Date Received 2021-11-12 category on the target platform within a third preset time period and sample user features corresponding thereto, and the second negative sample set contains sample users randomly selected from a date on which users of the second positive sample set did not purchase products relevant to the target brand category within the third preset time period and sample user features corresponding thereto;

[0020] respectively training m models via preset m algorithms with sample user features in the second positive sample set and the second negative sample set as input, with marketing effect indicator scores as output; and

[0021] determining the model with the highest accuracy rate from n+m models as the target population circling model, where m>2.

[0022] In a preferred embodiment, the sample user features include at least one of dimensional data selected from user label features of the users on the target platform, behavior features of users towards corresponding brands, purchasing behaviors of users on the target platform, and behavior features of users towards competitive products on the target platform.

[0023] In a preferred embodiment, the algorithms include, but are not limited to, at least one of logistic regression, random forest, and xgboost.

[0024] According to the second aspect, there is provided a method of creating a target population circling model, and the method comprises the following steps:

[0025] extracting a first positive sample set and a first negative sample set from a target platform according to a target brand category to be tested, wherein the first positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a second preset time period and sample user features corresponding thereto, and the first negative sample set contains sample users randomly selected from a date on which users of the first positive sample set did not purchase products relevant to the target brand category within the second preset time period and Date Recue/Date Received 2021-11-12 sample user features corresponding thereto;

[0026] respectively training n models via preset n algorithms with sample user features in the first positive sample set and the first negative sample set as input, with marketing effect indicator scores as output, where n>2; and

[0027] determining the model with the highest accuracy rate from the n models as the target population circling model.

[0028] According to the third aspect, there is provided a device for circling target population, and the device comprises:

[0029] a first obtaining module, for obtaining a target brand category to be tested, marketing information, a marketing channel and a marketing effect indicator;

[0030] a second obtaining module, for obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set;
and

[0031] a testing module, for inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user IDs sequenced according to marketing effect indicator scores.

[0032] According to the fourth aspect, there is provided a device for creating a target population circling model, and the device at least comprises:

[0033] a first extracting module, for extracting a first positive sample set and a first negative sample set from a target platform according to a target brand category to be tested, wherein the first positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a second preset time period and sample user features corresponding thereto, and the first negative sample set contains sample users randomly selected from a date on which users of the first positive sample set did not purchase products relevant to the target brand category within the second preset Date Recue/Date Received 2021-11-12 time period and sample user features corresponding thereto;

[0034] a training module, for respectively training n models via preset n algorithms with sample user features in the first positive sample set and the first negative sample set as input, with marketing effect indicator scores as output, where n>2; and

[0035] a determining module, for determining the model with the highest accuracy rate from the n models as the target population circling model.

[0036] According to the fifth aspect, there is provided a computer system that comprises:

[0037] one or more processor(s); and

[0038] a memory, associated with the one or more processor(s), wherein the memory is employed to store a program instruction, and the program instruction performs the following operations when it is read and executed by the one or more processor(s):

[0039] obtaining a target brand category to be tested, marketing information, a marketing channel and a marketing effect indicator;

[0040] obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set; and

[0041] inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user IDs sequenced according to marketing effect indicator scores.

[0042] In comparison with prior-art technologies, the present invention achieves the following advantageous effects.

[0043] The present invention provides a method of and a device for circling target population, and a method of and a device for creating a target population circling model, of which the method of circling target population comprises: obtaining a target brand category to be Date Recue/Date Received 2021-11-12 tested, marketing information, a marketing channel and a marketing effect indicator;
obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set; and inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user IDs sequenced according to marketing effect indicator scores. The method takes as basis the previous marketing information on the target platform within a preset time period to obtain the prediction sample set, whereby it is made possible to quickly readjust the prediction sample set by readjusting the preset time period, so as to enhance marketing effect of the target population package.

[0044] In addition, when the target population circling model is being created, modeling is performed with respect to specific brands, different brand tonalities of different professions are satisfied, and precision in population is enhanced; moreover, parallel training is carried out on corresponding number of models via a plurality of algorithms, and the model with the highest accuracy rate is determined as the target population circling model, so that the optimal model is selected; furthermore, before each prediction, positive samples and negative samples can be repeatedly extracted to again create a plurality of models, models created before are compared with models created thereafter and minimum accuracy rates are automatically compared in unison, and the model with the highest accuracy rate is determined as the target population circling model, until the model is solidified.

[0045] As should be noted, it suffices for the solutions of the present application to realize anyone of the foregoing technical effects.
BRIEF DESCRIPTION OF THE DRAWINGS

Date Recue/Date Received 2021-11-12

[0046] To more clearly describe the technical solutions in the embodiments of the present invention, drawings required to illustrate the embodiments will be briefly introduced below.
Apparently, the drawings introduced below are merely directed to some embodiments of the present invention, while persons ordinarily skilled in the art may further acquire other drawings on the basis of these drawings without spending creative effort in the process.

[0047] Fig. 1 is a flowchart illustrating a method of circling target population provided by Embodiment 1 of the present invention;

[0048] Fig. 2 is a flowchart illustrating a method of creating a target population circling model provided by Embodiment 2 of the present invention;

[0049] Fig. 3 is a view schematically illustrating the structure of a device for circling target population provided by Embodiment 3 of the present invention;

[0050] Fig. 4 is a view schematically illustrating the structure of a device for creating a target population circling model provided by Embodiment 4 of the present invention;
and

[0051] Fig. 5 is a view illustrating the framework of a computer system provided by Embodiment 5 of the present invention.
DETAILED DESCRIPTION OF THE INVENTION

[0052] To make more lucid and clear the objectives, technical solutions and advantages of the present invention, the technical solutions in the embodiments of the present invention will be clearly and comprehensively described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the embodiments as described are merely partial, rather than the entire, embodiments of the present invention.
Any other embodiments makeable by persons ordinarily skilled in the art on the basis of Date Recue/Date Received 2021-11-12 the embodiments in the present invention without creative effort shall all fall within the protection scope of the present invention.

[0053] In view of the problem that it is impossible to be so exact as to brands when potential consumers are circled at a current e-commerce platform to recommend products, embodiments of the present invention provide a method of and a device for circling target population, and a method of and a device for creating a target population circling model, whereby it is made possible to perform directed recommendations with respect to brand categories, and the precision in circling potential target population is therefore enhanced.

[0054] The method of and device for circling target population, the method of and device for creating a target population circling model as well as a system therefor are described in greater detail below with reference to specific embodiments.
Embodiment 1

[0055] With reference to what is shown in Fig. 1, this embodiment provides a method of circling target population, and the method at least comprises the following steps.

[0056] Si - obtaining a target brand category to be tested, marketing information, a marketing channel and a marketing effect indicator.

[0057] The method of circling target population in this embodiment is directed to products of categories designated under a brand, rather than the traditional circling of population directed to products of designated categories of all brands in a profession, so the method of circling target population in this embodiment can be so exact as to brands, whereby, during specific execution, it is not required for brand owners and operating personnel to participate in label circling, as it is only required for the business side to determine such Date Recue/Date Received 2021-11-12 information as brands and categories then population can be circled, and finally is obtained the circling result directed to designated brand categories.

[0058] Marketing information includes marketing orders and marketing periods;
according to the budget of the activity, different marketing periods also require different marketing orders. The marketing channel includes advertisement channel, and short message/PUSH
channel; once marketing channels are different, user IDs are also different, the advertisement channel usually makes use of unique devices to represent user IDs, while the short message channel and the push channel usually make use of member codes to represent user IDs. The marketing effect indicator means that it is needed to obtain appraisal effect after the method is executed; when the marketing channel is advertisement, the marketing effect indicator will be click-through rate, when the marketing channel is short message/PUSH, the marketing effect indicator will be conversion rate.

[0059] Exemplarily, the current item is as follows: brand category is Dyson vacuum cleaner, marketing order is 100 thousand people, marketing period is 3 days, marketing channel is advertisement, and marketing effect indicator is click-through rate.

[0060] S2 - obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set; this step specifically including:

[0061] S21 - obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period, and extracting user IDs and marketing effect in the previous marketing information;

[0062] S22 ¨ selecting a set of users with marketing effect is within a preset threshold as an initial test sample set; and Date Recue/Date Received 2021-11-12

[0063] S23 - taking the initial test sample set as basis and combining therewith received label customization circling information to obtain the prediction sample set.

[0064] The target platform can be any of such platforms as Suning or JD.com, while no restriction is made thereto in the present application. The previous marketing information is any preset marketing information within the first preset time period before the current marketing. Accordingly, on a designated target platform, on the precondition that the brand category has been determined, the marketing information within the first preset time period is determined and knowable, but new previous marketing information can be obtained through readjustment of the first preset time period, so, when the method is executed, a set of users with better previous marketing effect is selected as the initial test sample set according to experience or through many rounds of readjustment, namely to select a set of users with marketing effect is within a preset threshold as the initial test sample set. In comparison with randomly taking the target platform or a set of users of a certain product as the initial test sample set, this method achieves better marketing effect of the target population package.

[0065] After the initial test sample set has been determined, the initial test sample set is taken as basis to be combined with received label customization circling information to obtain the prediction sample set. The operating personnel bases on the current marketing item requirement to perform customized label circling on the basis of the initial test sample set and according to experience to finally obtain the prediction sample set.

[0066] S3 - inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user IDs sequenced according to marketing effect indicator scores. Specifically, step S3 specifically includes the following.

Date Recue/Date Received 2021-11-12

[0067] S31 ¨ inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes user IDs and corresponding marketing effect indicator scores. When the prediction is successfully converted, the marketing effect indicator scores 100, and when the prediction is not successfully converted, the marketing effect indicator scores 0.

[0068] In a preferred embodiment, after the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator have been obtained, they can be pre-stored in a datasheet, and it suffices to input the information in the datasheet into a pre-trained model during test.

[0069] S32 ¨ sequencing the user IDs in the target population package according to marketing effect indicator scores, preferably, in a descending order.

[0070] After such descending sequencing, it suffices to select header population that conforms to the marketing quantity as the circled population, whereby flexibility of population circling can be enhanced.

[0071] Exemplarily, the population package of Dyson vacuum cleaner output from this embodiment contains the top 100 thousand device models sequenced in a descending order according to click-through rate.

[0072] Moreover, the method further comprises step SO ¨ pre-creating a target population circling model, which includes the following sub-steps.

[0073] SO1 - extracting a first positive sample set and a first negative sample set from the target Date Recue/Date Received 2021-11-12 platform according to the target brand category to be tested, wherein the first positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a second preset time period and sample user features corresponding thereto, and the first negative sample set contains sample users randomly selected from a date on which users of the first positive sample set did not purchase products relevant to the target brand category within the second preset time period and sample user features corresponding thereto.

[0074] The sample user features include at least one of dimensional data selected from user label features of the users on the target platform, behavior features of users towards corresponding brands, purchasing behaviors of users on the target platform, and behavior features of users towards competitive products on the target platform.

[0075] Exemplarily, the user label features are user labels labeled onto the users according to user behaviors and basic information etc., and specific feature dimensions include age, gender, membership class, purchasing power, credit rating, loyalty, city level to which the member pertains, member registration duration, member value grading (from the perspective of conversion, or the perspective of flow rate).

[0076] The behavior features of users towards corresponding brands are behaviors of users towards brands to be predicted, and the feature dimensions are, for example, purchasing dimension, adding to shopping cart dimension, listing as favorite dimension, browsing dimension, appraising dimension, complaining dimension, etc. Specifically, purchasing dimension includes purchasing interval, purchasing number of days, purchasing amount, purchasing frequency, customer unit price, returning frequency, etc.; adding to shopping cart dimension includes the number of additions to shopping cart; listing as favorite dimension includes number of such listings; browsing dimension includes browsing numbers, browsing number of days, number of commodities browsed, etc.;
appraising dimension includes average star level of commodities appraised, and number of Date Recue/Date Received 2021-11-12 commodities appraised, etc.; and complaining dimension includes complaint numbers.

[0077] The purchasing behavior features of users on the target platform are purchasing behaviors of various categories of users on the target platform, and the feature dimensions include purchasing interval, purchasing number of days, purchasing amount, purchasing frequency dimension, and the categories involved include air-conditioners, refrigerators and washing machines, brown appliances, digital products, computers, communication products, small home appliances, kitchen and bathing products, general goods, personal care articles and household cleaning appliances, maternal and infant products, beauty cosmetics, imported healthcare products and fresh foods, prepared alcoholic and nonalcoholic beverages, Medicare business department, grain, oil and leisure foods, alcohols, fresh foods, houseware and furnishings, health nourishing goods, home decorations, **buying ¨together, **Youfang rentals, etc.

[0078] The behavior features of users towards competitive products on the target platform are behaviors of users towards competitive brands of brands to be predicted, and feature dimensions included therein are purchasing dimension, adding to shopping cart dimension, listing as favorite dimension, browsing dimension, etc. Specific dimension contents thereof are similar to those of user label features of users on the target platform, while no repetition is made in this context.

[0079] Preferably, the first positive sample set and the first negative sample set are equivalent in magnitude, so as to enhance confidence of the model.

[0080] SO2 ¨ respectively training n models via preset n algorithms with sample user features in the first positive sample set and the first negative sample set as input, with marketing effect indicator scores as output, where n>2. The algorithms include, but are not limited to, at least one of logistic regression, random forest, and xgboost.

Date Recue/Date Received 2021-11-12

[0081] Specifically, the first sample set and the second sample set are combined into a single sample set, from which 70% data is randomly extracted to serve as a training set, and the remaining random 30% data serves as a testing set, of which the data of the training set is used to train models, and the testing set is used to appraise models.

[0082] In a preferred embodiment, when the number of samples is few, the mode of cross validation can be employed to enhance the precision and the recall rate of the model. At the same time, the random forest algorithm can also be combined with manual work to debug kernel parameters (the number of trees and depths of trees), to thereby output an optimal model.

[0083] S03 ¨ determining the model with the highest accuracy rate from the n models as the target population circling model.

[0084] On the basis of the above, training of the target population circling model is completed.

[0085] In a preferred embodiment, when the model is not solidified, during test, prior to inputting the prediction sample set into a pre-created target population circling model, the method further comprises real-time model validation, including:

[0086] Sal - extracting a second positive sample set and a second negative sample set from the target platform according to the target brand category to be tested, wherein the second positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a third preset time period and sample user features corresponding thereto, and the second negative sample set contains sample users randomly selected from a date on which users of the second positive sample set did not purchase products relevant to the target brand category within the third preset time period and sample user features corresponding thereto;

[0087] 5a2 - respectively training m models via preset m algorithms with sample user features in the second positive sample set and the second negative sample set as input, with Date Recue/Date Received 2021-11-12 marketing effect indicator scores as output, where m>2; and

[0088] Sa3 - determining the model with the highest accuracy rate from n+m models as the target population circling model.

[0089] Of course, after the model has been solidified, it would be unnecessary to use the second positive sample set and the second negative sample set, as described in steps Sa1-Sa3, to again train or validate to select the optimal model, as it suffices to directly solidify the model.

[0090] In a preferred embodiment, before the previous marketing information and the sample user features are obtained, underlying data cleaning is performed on the obtained source data, and the underlying data cleaning rule includes processing underlying abnormal data.
Cleaning of the business layer during such processing includes removing refund orders, removing scalped orders (by scalper users), removing public cards (if one user ID
corresponds to orders of very large magnitude, the business system will judge it as a public member card), and removing abnormally browsed data (one member code or device corresponds to a pv number of very large magnitude, the business system will judge it as a suspicious crawler user). Cleaning of the data layer includes converting discrete features of string format into tht, double type, and performing one-hot coding on classification features.

[0091] The method takes as basis the previous marketing information on the target platform within a preset time period to obtain the prediction sample set, whereby it is made possible to quickly readjust the prediction sample set by readjusting the preset time period, so as to enhance marketing effect of the target population package.

[0092] In addition, when a target population circling model is being created, parallel training is carried out on corresponding number of models via a plurality of algorithms, and the model with the highest accuracy rate is determined as the target population circling model, so that Date Recue/Date Received 2021-11-12 the optimal model is selected; furthermore, before each prediction, positive samples and negative samples can be repeatedly extracted to again create a plurality of models, models created before are compared with models created thereafter and minimum accuracy rates are automatically compared in unison, and the model with the highest accuracy rate is determined as the target population circling model, until the model is solidified.
Embodiment 2

[0093] With reference to what is shown in Fig. 2, this embodiment provides a method of creating a target population circling model, which method comprises the following steps:

[0094] S101 - extracting a first positive sample set and a first negative sample set from a target platform according to a target brand category to be tested, wherein the first positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a second preset time period and sample user features corresponding thereto, and the first negative sample set contains sample users randomly selected from a date on which users of the first positive sample set did not purchase products relevant to the target brand category within the second preset time period and sample user features corresponding thereto;

[0095] S102 - respectively training n models via preset n algorithms with sample user features in the first positive sample set and the first negative sample set as input, with marketing effect indicator scores as output, where n>2; and

[0096] S103 - determining the model with the highest accuracy rate from the n models as the target population circling model.

[0097] The sample user features include at least one of dimensional data selected from user label features of the users on the target platform, behavior features of users towards

[0098] corresponding brands, purchasing behaviors of users on the target platform, and behavior features of users towards competitive products on the target platform.

Date Recue/Date Received 2021-11-12

[0099] The algorithms include, but are not limited to, at least one of logistic regression, random forest, and xgboost.

[0100] The specific realization process of and effects possessed by the method of creating a target population circling model in this embodiment can be inferred from the relevant descriptions in Embodiment 1, while no repetition is made in this context.
Embodiment 3

[0101] To execute the method of circling target population in the aforementioned Embodiment 1, this embodiment provides a corresponding device for circling target population, as shown in Fig. 3, the device at least comprises:

[0102] a first obtaining module, for obtaining a target brand category to be tested, marketing information, a marketing channel and a marketing effect indicator;

[0103] a second obtaining module, for obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set;
and

[0104] a testing module, for inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user IDs sequenced according to marketing effect indicator scores.

[0105] Further, the second obtaining module includes:

[0106] a first obtaining unit, for obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period, and extracting user IDs and marketing effect in the previous marketing information;

[0107] a second obtaining unit, for selecting a set of users with marketing effect is within a preset threshold as an initial test sample set; and Date Recue/Date Received 2021-11-12

[0108] a first processing unit, for taking the initial test sample set as basis and combining therewith received label customization circling information to obtain the prediction sample set.

[0109] The device further comprises a model training module for pre-creating a target population circling model, and the model training module specifically includes:

[0110] a first extracting unit, for extracting a first positive sample set and a first negative sample set from the target platform according to the target brand category to be tested, wherein the first positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a second preset time period and sample user features corresponding thereto, and the first negative sample set contains sample users randomly selected from a date on which users of the first positive sample set did not purchase products relevant to the target brand category within the second preset time period and sample user features corresponding thereto;

[0111] a first training unit, for respectively training n models via preset n algorithms with sample user features in the first positive sample set and the first negative sample set as input, with marketing effect indicator scores as output, where n>2; and

[0112] a first determining unit, for determining the model with the highest accuracy rate from the n models as the target population circling model.

[0113] The device further comprises a real-time model validating module, and the module includes:

[0114] a second extracting unit, for extracting a second positive sample set and a second negative sample set from the target platform according to the target brand category to be tested, wherein the second positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a third preset time period and sample user features corresponding thereto, and the second negative sample set contains sample users randomly selected from a date on which users of the second positive sample set did not purchase products relevant to the target brand category Date Recue/Date Received 2021-11-12 within the third preset time period and sample user features corresponding thereto;

[0115] a second training unit, for respectively training m models via preset m algorithms with sample user features in the second positive sample set and the second negative sample set as input, with marketing effect indicator scores as output, where m>2; and

[0116] a second determining unit, for determining the model with the highest accuracy rate from n+m models as the target population circling model.

[0117] The sample user features include at least one of dimensional data selected from user label features of the users on the target platform, behavior features of users towards corresponding brands, purchasing behaviors of users on the target platform, and behavior features of users towards competitive products on the target platform. The algorithms include, but are not limited to, at least one of logistic regression, random forest, and xgboost.

[0118] As should be noted, when the device for circling target population provided by this embodiment triggers the business of circling target population, the division into the aforementioned various functional modules is merely by way of example, while it is possible, in actual application, to base on requirements to assign the functions to different functional modules for completion, that is to say, to divide the internal structure of the device into different functional modules to complete the entire or partial functions described above. In addition, the device for circling target population provided by this embodiment pertains to the same conception as the method of circling target population provided by Embodiment 1, that is to say, the device is based on the method ¨
see the corresponding method embodiment for its specific realization process, while no repetition will be made in this context.
Embodiment 4

[0119] To execute the method of creating a target population circling model in the aforementioned Embodiment 2, this embodiment provides a corresponding device for Date Recue/Date Received 2021-11-12 creating a target population circling model, as shown in Fig. 4, the device at least comprises:

[0120] a first extracting module, for extracting a first positive sample set and a first negative sample set from a target platform according to a target brand category to be tested, wherein the first positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a second preset time period and sample user features corresponding thereto, and the first negative sample set contains sample users randomly selected from a date on which users of the first positive sample set did not purchase products relevant to the target brand category within the second preset time period and sample user features corresponding thereto;

[0121] a training module, for respectively training n models via preset n algorithms with sample user features in the first positive sample set and the first negative sample set as input, with marketing effect indicator scores as output, where n>2; and

[0122] a determining module, for determining the model with the highest accuracy rate from the n models as the target population circling model.

[0123] As should be noted, when the device for creating a target population circling model provided by this embodiment triggers the business of creating a target population circling model, the division into the aforementioned various functional modules is merely by way of example, while it is possible, in actual application, to base on requirements to assign the functions to different functional modules for completion, that is to say, to divide the internal structure of the device into different functional modules to complete the entire or partial functions described above. In addition, the device for creating a target population circling model provided by this embodiment pertains to the same conception as the method of creating a target population circling model provided by Embodiment 2, that is to say, the device is based on the method ¨ see the corresponding method embodiment for its specific realization process, while no repetition will be made in this context.
Embodiment 5 Date Recue/Date Received 2021-11-12

[0124] Corresponding to the aforementioned methods and devices, this embodiment provides a computer system that comprises:

[0125] one or more processor(s); and

[0126] a memory, associated with the one or more processor(s), wherein the memory is employed to store a program instruction, and the program instruction performs the following operations when it is read and executed by the one or more processor(s):

[0127] obtaining a target brand category to be tested, marketing information, a marketing channel and a marketing effect indicator;

[0128] obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set; and

[0129] inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user IDs sequenced according to marketing effect indicator scores.

[0130] Fig. 5 exemplarily illustrates the framework of the computer system that can specifically include a processor 1510, a video display adapter 1511, a magnetic disk driver 1512, an input/output interface 1513, a network interface 1514, and a memory 1520. The processor 1510, the video display adapter 1511, the magnetic disk driver 1512, the input/output interface 1513, the network interface 1514, and the memory 1520 can be communicably connected with one another via a communication bus 1530.

[0131] The processor 1510 can be embodied as a general CXU (Central Xrocessing Unit), a microprocessor, an ASIC (AXXlication SXecific Integrated Circuit), or one or more integrated circuit(s) for executing relevant program(s) to realize the technical solutions provided by the present application.

Date Recue/Date Received 2021-11-12

[0132] The memory 1520 can be embodied in such a form as an ROM (Read Only Memory), an RAM (Random Access Memory), a static storage device, or a dynamic storage device.
The memory 1520 can store an operating system 1521 for controlling the running of a computer system 1500, and a basic input/output system (BIOS) for controlling lower-level operations of the computer system 1500. In addition, the memory 1520 can also store a web browser 1523, a data storage management system 1524, and an icon font processing system 1525, etc. The icon font processing system 1525 can be an application program that specifically realizes the aforementioned various step operations in the embodiments of the present application. To sum it up, when the technical solutions provided by the present application are to be realized via software or firmware, the relevant program codes are stored in the memory 1520, and invoked and executed by the processor 1510.

[0133] The input/output interface 1513 is employed to connect with an input/output module to realize input and output of information. The input/output module can be equipped in the device as a component part (not shown in the drawings), and can also be externally connected with the device to provide corresponding functions. The input means can include a keyboard, a mouse, a touch screen, a microphone, and various sensors etc., and the output means can include a display screen, a loudspeaker, a vibrator, an indicator light etc.

[0134] The network interface 1514 is employed to connect to a communication module (not shown in the drawings) to realize intercommunication between the current device and other devices. The communication module can realize communication in a wired mode (via USB, network cable, for example) or in a wireless mode (via mobile network, WIFI, Bluetooth, etc.).

[0135] The bus 1530 includes a passageway transmitting information between various component parts of the device (such as the processor 1510, the video display adapter 1511, the magnetic disk driver 1512, the input/output interface 1513, the network interface 1514, and the memory 1520).

Date Recue/Date Received 2021-11-12

[0136] Additionally, the computer system 1500 may further obtain information of specific collection conditions from a virtual resource object collection condition information database 1541 for judgment on conditions, and so on.

[0137] As should be noted, although merely the processor 1510, the video display adapter 1511, the magnetic disk driver 1512, the input/output interface 1513, the network interface 1514, the memory 1520, and the bus 1530 are illustrated for the aforementioned device, the device may further include other component parts prerequisite for realizing normal running during specific implementation. In addition, as can be understood by persons skilled in the art, the aforementioned device may as well only include component parts necessary for realizing the solutions of the present application, without including the entire component parts as illustrated.

[0138] As can be known through the description to the aforementioned embodiments, it is clearly learnt by person skilled in the art that the present application can be realized through software plus a general hardware platform. Based on such understanding, the technical solutions of the present application, or the contributions made thereby over the state of the art, can be essentially embodied in the form of a software product, and such a computer software product can be stored in a storage medium, such as an ROM/RAM, a magnetic disk, an optical disk etc., and includes plural instructions enabling a computer equipment (such as a personal computer, a cloud server, or a network device etc.) to execute the methods described in various embodiments or some sections of the embodiments of the present application.

[0139] The various embodiments are progressively described in the Description, identical or similar sections among the various embodiments can be inferred from one another, and each embodiment stresses what is different from other embodiments.
Particularly, with respect to the system or system embodiment, since it is essentially similar to the method Date Recue/Date Received 2021-11-12 embodiment, its description is relatively simple, and the relevant sections thereof can be inferred from the corresponding sections of the method embodiment. The system or system embodiment as described above is merely exemplary in nature, units therein described as separate parts can be or may not be physically separate, parts displayed as units can be or may not be physical units, that is to say, they can be located in a single site, or distributed over a plurality of network units. It is possible to base on practical requirements to select partial modules or the entire modules to realize the objectives of the embodied solutions.
It is understandable and implementable by persons ordinarily skilled in the art without spending creative effort in the process.

[0140] Although preferred embodiments in the embodiments of the present invention have been described so far, it is still possible for persons skilled in the art to make additional modifications and amendments to these embodiments upon learning of the basic inventive conception. Accordingly, the attached Claims are meant to cover the preferred embodiments and all modifications and amendments that fall within the scope of the embodiments of the present invention.

[0141] Apparently, persons skilled in the art may make various alterations and modifications to the present invention without departing from the spirit and scope of the present invention.
Thusly, should such amendments and modifications to the present invention fall within the scope of the Claims of the present invention and the equivalent technology, the present invention is also construed to cover such amendments and modifications.
Date Recue/Date Received 2021-11-12

Claims (10)

What is claimed is:

1. A method of circling target population, characterized in comprising at least the following steps:
obtaining a target brand category to be tested, marketing information, a marketing channel and a marketing effect indicator;
obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set; and inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user identifications (hereinafter referred to as "IDs") sequenced according to marketing effect indicator scores.

2. The method according to Claim 1, characterized in that the step of obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set includes the following sub-steps:
obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period, and extracting user IDs and marketing effect in the previous marketing information;

selecting a set of users with marketing effect is within a preset threshold as an initial test sample set; and taking the initial test sample set as basis and combining therewith received label customization circling information to obtain the prediction sample set.

3. The method according to Claim 1, characterized in further comprising: pre-creating a target population circling model, wherein the step includes the following sub-steps:
extracting a first positive sample set and a first negative sample set from the target platform according to the target brand category to be tested, wherein the first positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a second preset time period and sample user features corresponding thereto, and the first negative sample set contains sample users randomly selected from a date on which users of the first positive sample set did not purchase products relevant to the target brand category within the second preset time period and sample user features corresponding thereto;
respectively training n models via preset n algorithms with sample user features in the first positive sample set and the first negative sample set as input, with marketing effect indicator scores as output, where n>2; and determining the model with the highest accuracy rate from the n models as the target population circling model.

4. The method according to Claim 3, characterized in that, prior to inputting the prediction sample set into a pre-created target population circling model, the method further comprises real-time model validation, including:

extracting a second positive sample set and a second negative sample set from the target platform according to the target brand category to be tested, wherein the second positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a third preset time period and sample user features corresponding thereto, and the second negative sample set contains sample users randomly selected from a date on which users of the second positive sample set did not purchase products relevant to the target brand category within the third preset time period and sample user features corresponding thereto;
respectively training m models via preset m algorithms with sample user features in the second positive sample set and the second negative sample set as input, with marketing effect indicator scores as output; and determining the model with the highest accuracy rate from n+m models as the target population circling model, where m>2.

5. The method according to Claim 3 or 4, characterized in that the sample user features include at least one of dimensional data selected from user label features of the users on the target platform, behavior features of users towards corresponding brands, purchasing behaviors of users on the target platform, and behavior features of users towards competitive products on the target platform.

6. The method according to Claim 5, characterized in that the algorithms include, but are not limited to, at least one of logistic regression, random forest, and xgboost.

7. A method of creating a target population circling model, characterized in comprising the following steps:
extracting a first positive sample set and a first negative sample set from a target platform according to a target brand category to be tested, wherein the first positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a second preset time period and sample user features corresponding thereto, and the first negative sample set contains sample users randomly selected from a date on which users of the first positive sample set did not purchase products relevant to the target brand category within the second preset time period and sample user features corresponding thereto;
respectively training n models via preset n algorithms with sample user features in the first positive sample set and the first negative sample set as input, with marketing effect indicator scores as output, where n>2; and determining the model with the highest accuracy rate from the n models as the target population circling model.

8. A device for circling target population, characterized in comprising:
a first obtaining module, for obtaining a target brand category to be tested, marketing information, a marketing channel and a marketing effect indicator;
a second obtaining module, for obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set; and a testing module, for inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user IDs sequenced according to marketing effect indicator scores.

9. A device for creating a target population circling model, characterized in at least comprising:
a first extracting module, for extracting a first positive sample set and a first negative sample set from a target platform according to a target brand category to be tested, wherein the first positive sample set contains all sample users having purchased products relevant to the target brand category on the target platform within a second preset time period and sample user features corresponding thereto, and the first negative sample set contains sample users randomly selected from a date on which users of the first positive sample set did not purchase products relevant to the target brand category within the second preset time period and sample user features corresponding thereto;
a training module, for respectively training n models via preset n algorithms with sample user features in the first positive sample set and the first negative sample set as input, with marketing effect indicator scores as output, vvhere n>2; and a determining module, for determining the model with the highest accuracy rate from the n models as the target population circling model.

10. A computer system, characterized in comprising:
one or more processor(s); and a memory, associated with the one or more processor(s), wherein the memory is employed to store a program instruction, and the program instruction performs the following operations when it is read and executed by the one or more processor(s):
obtaining a target brand category to be tested, marketing information, a marketing channel and a marketing effect indicator;
obtaining previous marketing information of the target brand category to be tested on a target platform within a first preset time period and serving with the label customization circling information as a prediction sample set; and inputting the prediction sample set, the target brand category to be tested, the marketing information, the marketing channel and the marketing effect indicator into a pre-created target population circling model to obtain a target population package, wherein the target population package includes a plurality of user IDs sequenced according to marketing effect indicator scores.