A method and apparatus for determining a recommended advertisement, and an electronic device

Abstract

The application discloses a method and device for determining recommended advertisements and electronic equipment, and relates to the technical field of big data intelligent analysis. The method comprises the following steps: obtaining behavior index data of a plurality of sample users on each sample advertisement; determining score values of each sample user on each advertisement category according to the behavior index data, so as to form a score vector corresponding to each sample user; taking the score vector corresponding to each sample user as a representative of the sample user, clustering the plurality of sample users by using a clustering algorithm, and obtaining a plurality of user groups; obtaining behavior index data of a target user on each historical advertisement; determining a target user group closest to the target user from the plurality of user groups according to the behavior index data of the target user; and determining an advertisement recommended for the target user according to an advertisement set interested by each user in the target user group. According to the scheme, the advertisement recommended for the target user is determined according to the preference of the sample user, and the accuracy of the recommended advertisement is improved.

Description

Technical Field

[0001] This application relates to the field of big data intelligent analysis technology, and in particular to a method, apparatus and electronic device for determining recommended advertisements. Background Technology

[0002] With the rapid development of the internet age, user information and behavioral data are also growing rapidly, and internet technology has become an important part of people's lives. Recommending personalized advertisements based on users' individual preferences can usually bring a better service experience.

[0003] Currently, the determination of personalized advertisements recommended to users usually involves obtaining the target user's basic personal information (such as gender, age, topics of interest, etc.) and behavioral data, determining the user profile based on the basic personal information and behavioral data, and then using pre-set advertisements corresponding to the user profile as advertisements recommended to the target user.

[0004] However, if the pre-set ads corresponding to the user profile are inaccurate, or if the method for determining the user profile is inaccurate, the recommended ads will be inaccurate, affecting the experience of the target users. Summary of the Invention

[0005] The purpose of this application is to provide a method, apparatus, and electronic device for determining recommended advertisements, so as to solve the problem of inaccuracy in existing methods for determining recommended advertisements.

[0006] To address the aforementioned technical problems, this specification provides a method for determining recommended advertisements, comprising: acquiring behavioral indicator data of multiple sample users on each sample advertisement while browsing advertisements; determining the rating values ​​of each sample user for each advertisement category based on the behavioral indicator data, and constructing a rating vector corresponding to each sample user using the rating values ​​of the sample users for each advertisement category; using the rating vectors corresponding to the sample users to represent the sample users, and clustering the multiple sample users using a clustering algorithm to obtain multiple user groups; acquiring behavioral indicator data of a target user on each historical advertisement while browsing historical advertisements; determining the target user group most similar to the target user among the multiple user groups based on the target user's behavioral indicator data; and determining the advertisements recommended to the target user based on the set of advertisements of interest to each user in the target user group.

[0007] In some embodiments, before determining the rating vector of each sample user for each advertising category based on the behavioral indicator data, the method further includes: obtaining a set of sample advertisements viewed by the plurality of sample users; and clustering the sample advertisements in the set to obtain a plurality of advertising categories.

[0008] In some embodiments, determining the rating of each sample user for each advertising category based on the behavioral indicator data includes: calculating the rating of the current sample user for the current advertising category based on the behavioral indicator data of the current sample user for all sample advertisements under the current advertising category.

[0009] In some embodiments, the rating of the current sample user for the current advertising category is calculated based on the behavioral indicator data of the current sample user for all sample advertisements under the current advertising category, including: calculating the rating of the current sample user for each sample advertisement by the following methods: calculating the rating of the current sample user for the current sample advertisement based on various behavioral indicator data of the current sample user for the current sample advertisement; and calculating the rating of the current sample user for the current advertising category based on the rating of each sample advertisement under the current category.

[0010] In some embodiments, the rating of the current sample user for the current advertising category is calculated based on the behavioral indicator data of the current sample user for all sample advertisements under the current advertising category. This includes: calculating the comprehensive value of each behavioral indicator data corresponding to the current category; calculating the comprehensive value of the current behavioral indicator data based on the current behavioral indicator data of each sample advertisement under the current category; obtaining the weights corresponding to each behavioral indicator data; and performing a weighted summation of the comprehensive values ​​of each behavioral indicator data to obtain the rating of the current sample user for the current advertising category.

[0011] In some embodiments, a clustering algorithm is used to cluster multiple sample users to obtain multiple user groups, including: randomly initializing n points from all sample users as initial cluster centers; repeatedly performing the following operations until a preset cutoff condition is reached: calculating the distance between each sample user and each cluster center, determining the cluster center with the smallest distance to the sample user, and assigning the sample user to a class represented by the distance center; and redetermining the cluster center of each class based on all sample users in the updated class.

[0012] In some embodiments, determining the target user group that is closest to the target user among the plurality of user groups based on the target user's behavioral indicator data includes: using the target user's behavioral indicator data to represent the target user, calculating the distance between the target user and each user group; and selecting the user group with the smallest distance as the target user group that is closest to the target user.

[0013] In some embodiments, the following steps are performed at predetermined intervals: acquiring behavioral indicator data of multiple sample users on each sample advertisement while browsing advertisements; determining the rating value of each sample user for each advertisement category based on the behavioral indicator data, and constructing a rating vector corresponding to each sample user based on the rating values ​​of each sample user for each advertisement category; using the rating vector corresponding to each sample user to represent the sample user, and using a clustering algorithm to cluster multiple sample users to obtain multiple user groups; when the data of a sample user reaches a predetermined threshold, or when the number of sample advertisements viewed by a sample user reaches a predetermined threshold, clearing the earliest sample advertisement behavioral indicator data so that the number of sample advertisements is within a predetermined numerical range.

[0014] In some embodiments, determining the advertisements recommended to the target user based on the set of advertisements that each user in the target user group is interested in includes: selecting the M sample advertisements with the highest ratings in the target user group as the advertisements recommended to the target user, where M is a preset natural number.

[0015] In some embodiments, determining the advertisements recommended to the target user based on the set of advertisements that each user in the target user group is interested in includes: determining the N sample users in the target user group who are closest to the target user, where N is a preset natural number; and determining the advertisements recommended to the target user based on the historical recommendation results of the N sample users.

[0016] In some embodiments, determining the advertisements recommended to the target user based on the set of advertisements that each user in the target user group is interested in includes: determining the M sample advertisements with the highest ratings in the target user group as a first advertisement set, where M is a preset natural number; obtaining the actual behavioral indicator data corresponding to each advertisement in the first advertisement set; determining the N sample users in the target user group who are closest to the target user, and using the historical recommendation results of the N sample users as a second advertisement set; obtaining the behavioral indicator prediction data corresponding to each advertisement in the second advertisement set; obtaining the intersection of the first advertisement set and the second advertisement set; calculating the behavioral indicator prediction data of the target advertisements in the intersection based on the first behavioral indicator and the behavioral indicator prediction data; and selecting a target number of advertisements from the intersection as the target advertisements recommended to the target user based on the behavioral indicator prediction data.

[0017] A second aspect of this specification provides an apparatus for determining recommended advertisements, comprising: a first acquisition unit for acquiring behavioral indicator data of multiple sample users on each sample advertisement while browsing advertisements; a first determination unit for determining the rating values ​​of each sample user for each advertisement category based on the behavioral indicator data, and constructing a rating vector corresponding to each sample user based on the rating values ​​of the sample users for each advertisement category; a user clustering unit for clustering multiple sample users using a clustering algorithm to represent the sample users with the rating vectors corresponding to the sample users; a second acquisition unit for acquiring behavioral indicator data of a target user on each historical advertisement while browsing historical advertisements; a second determination unit for determining the target user group most similar to the target user among the multiple user groups based on the behavioral indicator data of the target user; and a third determination unit for determining the advertisements recommended to the target user based on the set of advertisements of interest to each user in the target user group.

[0018] In some embodiments, the apparatus further includes: a third acquisition unit, configured to acquire a set of sample advertisements viewed by the plurality of sample users; and an advertisement clustering unit, configured to cluster the sample advertisements in the set to obtain a plurality of advertisement categories.

[0019] In some embodiments, the first determining unit includes: a first calculation subunit, configured to calculate the rating value of the current sample user for the current advertising category according to the following method: calculating the rating value of the current sample user for the current advertising category based on the behavioral indicator data of the current sample user for all sample advertisements under the current advertising category; and a construction subunit, configured to construct a rating vector corresponding to the sample user based on the rating values ​​of the sample user for each advertising category.

[0020] In some embodiments, the first calculation subunit includes: a second calculation subunit, configured to calculate the rating of the current sample user for each sample advertisement by: calculating the rating of the current sample user for the current sample advertisement based on various behavioral indicator data of the current sample user for the current sample advertisement; and a third calculation subunit, configured to calculate the rating of the current sample user for the current advertisement category based on the rating of each sample advertisement under the current category.

[0021] In some embodiments, the first calculation subunit includes: a third calculation subunit, configured to calculate the comprehensive value of each behavioral indicator data corresponding to the current category: calculate the comprehensive value of the current behavioral indicator data based on the current behavioral indicator data of each sample advertisement under the current category; a first acquisition subunit, configured to acquire the weights corresponding to each type of behavioral indicator data; and a summation subunit, configured to perform weighted summation on the comprehensive values ​​of each type of behavioral indicator data to obtain the rating value of the current sample user for the current advertisement category.

[0022] In some embodiments, the user clustering unit includes: an initialization subunit for randomly initializing n points from all sample users as initial cluster centers; a fourth calculation subunit and an update subunit for cyclically performing operations until a preset cutoff condition is reached, wherein the fourth calculation subunit is used to calculate the distance between each sample user and each cluster center, determine the cluster center with the smallest distance to the sample user, and assign the sample user to a class represented by the distance center; the update subunit is used to redetermine the cluster center of each class based on all sample users in the updated class.

[0023] In some embodiments, the second determining unit includes: a fifth calculation subunit, used to calculate the distance between the target user and each user group using the target user's behavioral indicator data to represent the target user; and a first determining subunit, used to select the user group with the smallest distance as the target user group closest to the target user.

[0024] In some embodiments, the following steps are performed at predetermined intervals: acquiring behavioral indicator data of multiple sample users on each sample advertisement while browsing advertisements; determining the rating value of each sample user for each advertisement category based on the behavioral indicator data, and constructing a rating vector corresponding to each sample user based on the rating values ​​of each sample user for each advertisement category; using the rating vector corresponding to each sample user to represent the sample user, and using a clustering algorithm to cluster multiple sample users to obtain multiple user groups; when the data of a sample user reaches a predetermined threshold, or when the number of sample advertisements viewed by a sample user reaches a predetermined threshold, clearing the earliest sample advertisement behavioral indicator data so that the number of sample advertisements is within a predetermined numerical range.

[0025] In some embodiments, the third determining unit includes: a second determining subunit, configured to determine the M sample advertisements with the highest ratings in the target user group as a first advertisement set, where M is a preset natural number; a second obtaining subunit, configured to obtain actual behavioral indicator data corresponding to each advertisement in the first advertisement set; a third determining subunit, configured to determine the N sample users closest to the target user in the target user group, and use the historical recommendation results of the N sample users as a second advertisement set; a third obtaining subunit, configured to obtain behavioral indicator prediction data corresponding to each advertisement in the second advertisement set; a fourth obtaining subunit, configured to obtain the intersection of the first advertisement set and the second advertisement set; a sixth calculation subunit, configured to calculate the behavioral indicator prediction data of the target advertisements in the intersection based on the first behavioral indicator and the behavioral indicator prediction data; and a filtering subunit, configured to select a target number of advertisements from the intersection as target advertisements recommended to the target user based on the behavioral indicator prediction data.

[0026] A third aspect of this specification provides an electronic device, comprising: a memory and a processor, wherein the processor and the memory are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to implement the steps of the method described in any of the first aspects.

[0027] A fourth aspect of this specification provides a computer storage medium storing computer program instructions that, when executed by a processor, implement the steps of the method described in any of the first aspects.

[0028] A fifth aspect of this specification provides a computer program product comprising a computer program that, when executed by a processor, implements the steps of the method described in any of the first aspects.

[0029] The method, apparatus, and electronic product for determining recommended advertisements provided in this specification cluster sample users based on behavioral indicator data of sample users towards sample advertisements, resulting in multiple user groups. The target user group closest to the target user is identified, and then the advertisements recommended to the target user are determined based on the set of advertisements of interest to each user within that target user group. In other words, the advertisements recommended to the target user are determined based on the preferences of the sample users, eliminating the need to define a target user profile. This avoids the problem of inaccurate recommended advertisements caused by inaccurate user profile information. Clustering sample users into multiple user groups using behavioral indicator data of sample users towards sample advertisements, i.e., dividing user groups based on user behavior rather than static user indicator data (such as age, selected preferences, etc.), results in more accurate user group segmentation. Furthermore, the advertisements recommended to the target user are determined based on the preferences of the user group closest to the target user, thus making the recommended advertisements more accurate. Therefore, this solution improves the accuracy of recommended advertisements. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0031] Figure 1 A flowchart illustrating a method for determining recommended advertisements provided in this specification is shown;

[0032] Figure 2 A flowchart illustrating another method for determining recommended advertisements provided in this specification is shown;

[0033] Figure 3 A flowchart is shown for a method to determine the advertisements recommended to the target users based on the set of advertisements that each user in the target user group is interested in;

[0034] Figure 4 A schematic block diagram of a device for determining recommended advertisements provided in this specification is shown;

[0035] Figure 5 A schematic block diagram of an electronic device provided in this specification is shown. Detailed Implementation

[0036] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0037] The acquisition, storage, use, and processing of data in this application all comply with the relevant provisions of national laws and regulations.

[0038] This specification provides a method for determining recommended advertisements, such as... Figure 1 As shown, it includes the following steps:

[0039] S10: Obtain behavioral metrics data of multiple sample users when browsing advertisements for each sample advertisement.

[0040] Sample ads are pre-placed advertisements. The purpose of these ads, besides their inherent function, is to gather user behavior metrics. The types of sample ads can vary widely.

[0041] Behavioral metrics data refers to multiple indicators of user behavior when browsing advertisements. For example, for advertisements displayed on electronic media, multiple metrics data may include several of the following: whether they click on the ad link, the duration of browsing the ad page, the duration of browsing the content linked by the ad, whether they forward the ad link, and whether they choose to block the ad; for physical advertising media such as billboards, multiple metrics data may include several of the following: gaze duration, facial expression after gazing, whether they discuss the ad with a companion after gazing and the duration of the discussion, and whether they take photos; for voice advertisements, multiple metrics data may include several of the following: whether they listen to the ad content, their facial expression after listening to the ad, whether they discuss the ad with a companion after listening and the duration of the discussion, and whether they imitate the ad slogan, etc.

[0042] For advertisements displayed on electronic media, behavioral indicator data can be obtained by pre-setting tracking points in appropriate locations within the background program of the advertisement display and collecting behavioral indicator data through these tracking points. For physical advertising media such as billboards and voice advertising media, cameras and voice collectors can be set up in the surrounding area to determine behavioral indicator data through the collected user video and voice.

[0043] S20: Based on the behavioral indicator data, determine the rating values ​​of each sample user for each advertising category, and construct the rating vector corresponding to the sample user based on the rating values ​​of the sample users for each advertising category.

[0044] Before step S20, the category identifier of each sample advertisement can be obtained first.

[0045] In some embodiments, sample ads may be categorized according to predefined criteria before being deployed, in which case the category identifier of each ad can be obtained directly.

[0046] In some embodiments, the ad category can be the result of performing a clustering classification algorithm based on at least one of the following information: keywords and / or sentences in the ad copy, keyframes, key content in the image, and ad display style (e.g., ink painting style, comic style, retro style, etc.).

[0047] For example, first, obtain a set of sample advertisements viewed by the multiple sample users; then, cluster the sample advertisements in the set to obtain multiple advertisement categories. By clustering only the sample advertisements viewed by the sample users to obtain advertisement categories, the category to which the advertisement belongs can be adjusted according to the user's preferences, thus ensuring that the final recommended advertisements change with the changing trends of public preferences.

[0048] The rating values ​​represent the user's level of attention to each advertising category.

[0049] For example, if the sample users include A, B, and C, and the sample ad categories include X, Y, and Z, and the attention given to each sample user to each sample ad category is shown in Table 1 below, then the scoring vector can be represented as: Among them, a1, b1, c1, a2, b2, c2, a3, b3, c3, etc. represent the score values.

[0050] Table 1

[0051] A a1 b1 c1 B a2 b2 c2 C a3 b3 c3

[0052] The method for determining recommended advertisements provided in this manual is data-driven. The quality of the collected data determines the performance of the entire system. To ensure the accuracy and reliability of subsequent calculation results, a formula can be used first. The data is standardized by scaling it to unit variance to standardize the features, where x represents the behavioral indicator data before standardization. * The standardized behavioral indicator data is represented by μ, which represents the average value of the same behavioral indicator data, and σ, which represents the variance of the same behavioral indicator data. Here, the same behavioral indicator data is the same as the unstandardized behavioral indicator data x.

[0053] In some embodiments, such as Figure 2 As shown, S20 may include S21: Calculate the current sample user's rating for the current advertising category according to the following method: Calculate the current sample user's rating for the current advertising category based on the behavioral indicator data of the current sample user for all sample advertisements under the current advertising category.

[0054] Taking Table 1 as an example, if the current sample user is A and the current sample ad category is X, the score of sample user A for sample ad category X is calculated based on the behavioral index data of sample user A for all sample ads under category X.

[0055] Table 2

[0056] A p1 p2 p3 p4 A q1 q2 q3 q4 A w1 w2 w3 w4

[0057] In some embodiments, S21 may calculate the rating values ​​of the current sample users for each sample advertisement through the following steps S211 and S212:

[0058] S211: Calculate the rating of the current sample user for the current sample advertisement based on the various behavioral indicators of the current sample user for the current sample advertisement.

[0059] S212: Calculate the rating of the current sample user for the current advertising category based on the rating of each sample advertisement in the current category.

[0060] The data in Table 2 corresponds to AX in Table 1 (i.e., the data in the second row and second column). There are four types of behavioral indicator data: p1, q1, w1, p2, q2, w2, p3, q3, w3, etc., which represent the values ​​of the corresponding indicators. Each row in Table 2 represents the data corresponding to one sample advertisement.

[0061] Taking Table 2 as an example, S211 refers to first calculating the rating value 1 based on p1, p2, p3, and p4; then calculating the rating value 2 based on q1, q2, q3, and q4; and finally calculating the rating value 3 based on w1, w2, w3, and w4. S212 refers to calculating the rating value of sample user A for the current ad category based on the rating values ​​1, 2, and 3. For example, a weighted average of the rating values ​​1, 2, and 3 is taken to obtain the rating value for the ad category.

[0062] In some embodiments, S21 may calculate the comprehensive value of each behavioral indicator data corresponding to the current category in steps S213 and S214 respectively:

[0063] S213: Calculate the comprehensive value of the current behavior indicator data based on the current behavior indicator data of each sample advertisement in the current category.

[0064] S214: Obtain the weights corresponding to various behavioral indicator data.

[0065] S215: The weighted sum of the comprehensive values ​​of various behavioral indicator data is used to obtain the rating value of the current sample users for the current advertising category.

[0066] Taking Table 2 as an example, we can first calculate the comprehensive value 1 based on p1, q1, and w1, calculate the comprehensive value 2 based on p2, q2, and w2, and calculate the comprehensive value 3 based on p3, q3, and w3. S214 refers to calculating the rating of sample user A for the current advertising category based on comprehensive value 1, comprehensive value 2, and comprehensive value 3.

[0067] S30: Using the rating vector as a representative sample user, a clustering algorithm is used to cluster the sample users to obtain multiple user groups.

[0068] For example, first, randomly initialize n points from all sample users as initial cluster centers; then, repeatedly execute the following operations S31 and S32 until the preset cutoff condition is reached:

[0069] S31: Calculate the distance between each sample user and each cluster center, determine the cluster center with the smallest distance to the sample user, and assign the sample user to the class represented by the distance center.

[0070] S32: Based on all sample users in each category after the update, redetermine the cluster centers for each category.

[0071] Clustering algorithms are used to cluster sample users. Since the clustering is based on rating vectors, which are the ratings of sample users for each category of sample ads, sample users in the same cluster have more commonalities in terms of ads they are interested in. Therefore, the accuracy of recommending ads to target users based on the clustering results is higher.

[0072] S40: Obtain behavioral metrics data of target users when browsing historical ads.

[0073] The historical ads retrieved in S40 can be within a predetermined time span, such as historical ads viewed within one month.

[0074] The historical ads here can be sample ads that have already been placed, or newly placed ads that are different from the sample ads.

[0075] S50: Based on the behavioral indicator data of the target user, determine the target user group that is most similar to the target user among the multiple user groups.

[0076] S50 can use the behavioral indicator data of the target user to represent the target user, calculate the distance between the target user and each user group, and take the user group with the smallest distance as the target user group that is closest to the target user.

[0077] When calculating the distance between the target user and each user group, the distance between the target user and the center of the user group can be taken as the distance between the target user and the user group; alternatively, the minimum distance between the target user and each user in the user group can be taken as the distance between the target user and the user group; or the average of the minimum and maximum distances between the target user and each user in the user group can be taken as the distance between the target user and the user group.

[0078] S60: Based on the set of advertisements that each user in the target user group is interested in, determine the advertisements recommended to the target users.

[0079] Once the ads targeted for recommendation are determined, they can be displayed to users through fixed ads or carousels.

[0080] In some embodiments, S60 may select the M sample advertisements with the highest ratings from the target user group as advertisements recommended to the target user, where M is a preset natural number. The ratings can be calculated using the method described above.

[0081] In some embodiments, S60 may first determine the N sample users in the target user group that are closest to the target user, where N is a preset natural number; then, based on the historical recommendation results of the N sample users, determine the advertisements recommended to the target user. For example, the historical recommendation results of the N sample users may be directly used as the advertisements recommended to the target user, or a portion of the recommendation results may be further filtered from the historical recommendation results of the N sample users as the advertisements recommended to the target user.

[0082] In some embodiments, S60 may further combine the ads most preferred by the target user group as a whole with the historical recommendation results of sample users most similar to the target user to jointly determine the ads recommended to the target user. Accordingly, such as Figure 3 As shown, S60 may include the following steps S61-S67.

[0083] S61: Select the M sample ads with the highest ratings from the target user group as the first ad set, where M is a preset natural number.

[0084] S62: Obtain the actual behavioral indicator data corresponding to each advertisement in the first advertisement set.

[0085] The actual behavioral indicator data here refers to the actual collected behavioral indicator data, not the predicted data.

[0086] S63: Identify the N sample users in the target user group that are closest to the target user, and use the historical recommendation results of the N sample users as the second advertising set.

[0087] S64: Obtain the behavioral indicator prediction data corresponding to each advertisement in the second advertisement set.

[0088] When determining which ads to recommend to a user, behavioral metrics for each ad are calculated using actually collected data—that is, predicted data—rather than the actual collected data. Therefore, the behavioral metric predicted data obtained by S64 is predicted data, not actually collected data.

[0089] S65: Obtain the intersection of the first set of advertisements and the second set of advertisements.

[0090] S66: Calculate the predicted behavioral indicators for the target advertisement in the intersection based on actual behavioral indicator data and predicted behavioral indicator data.

[0091] For example, the weight of the actual behavior indicator data can be predetermined as η1, and the weight of the predicted behavior indicator data can be predetermined as η2. Then the predicted behavior indicator data of the target advertisement in the intersection is: D = D1 × η1 + D2 × η2, where D1 is the actual behavior indicator data and D2 is the predicted behavior indicator data.

[0092] S67: Based on the predicted data of the behavioral indicators, select a number of advertisements from the intersection as target advertisements to recommend to the target users.

[0093] In some embodiments, S10, S20, and S30 are executed at predetermined time intervals. That is, the user group is updated at predetermined time intervals, so that the preferences of the user group change in line with the current preferences of the general public.

[0094] When the data of sample users reaches a predetermined threshold, or when the number of sample ads viewed by sample users reaches a predetermined threshold, the earliest sample ad behavior indicator data is cleared so that the number of sample ads remains within a predetermined range.

[0095] The method, apparatus, and electronic product for determining recommended advertisements provided in this specification cluster sample users based on behavioral indicator data of sample users towards sample advertisements, resulting in multiple user groups. The target user group closest to the target user is identified, and then the advertisements recommended to the target user are determined based on the set of advertisements of interest to each user within that target user group. In other words, the advertisements recommended to the target user are determined based on the preferences of the sample users, eliminating the need to define a target user profile. This avoids the problem of inaccurate recommended advertisements caused by inaccurate user profile information. Clustering sample users into multiple user groups using behavioral indicator data of sample users towards sample advertisements, i.e., dividing user groups based on user behavior rather than static user indicator data (such as age, selected preferences, etc.), results in more accurate user group segmentation. Furthermore, the advertisements recommended to the target user are determined based on the preferences of the user group closest to the target user, thus making the recommended advertisements more accurate. Therefore, this solution improves the accuracy of recommended advertisements.

[0096] This system acquires basic user information, transaction data, product data, and user behavior data from the application platform. Based on this user information, it uses clustering algorithms to analyze user preferences. By mining user behavior, it discovers users' personalized needs and interests, and displays potentially interesting ad placements and links to users in a highly targeted manner. Simultaneously, it assists mobile banks in building digital payment capabilities, redirecting user traffic back to the bank and maximizing advertising revenue.

[0097] This specification provides a device for determining recommended advertisements, which can be used to achieve... Figure 1 The method shown. (As illustrated) Figure 4 As shown, the device includes a first acquisition unit 10, a first determination unit 20, a user clustering unit 30, a second acquisition unit 40, a second determination unit 50, and a third determination unit 60.

[0098] The first acquisition unit 10 is used to acquire behavioral indicator data of multiple sample users when browsing advertisements.

[0099] The first determining unit 20 is used to determine the rating values ​​of each sample user for each advertising category based on the behavioral indicator data, and to construct the rating vector corresponding to the sample user based on the rating values ​​of the sample users for each advertising category.

[0100] User clustering unit 30 is used to represent sample users with the rating vectors corresponding to the sample users, and to cluster multiple sample users using a clustering algorithm to obtain multiple user groups.

[0101] The second acquisition unit 40 is used to acquire behavioral indicator data of target users on each historical advertisement when browsing historical advertisements.

[0102] The second determining unit 50 is used to determine the target user group that is most similar to the target user among the multiple user groups based on the target user's behavioral indicator data.

[0103] The third determining unit 60 is used to determine the advertisements recommended to the target users based on the set of advertisements that each user in the target user group is interested in.

[0104] In some embodiments, the apparatus further includes: a third acquisition unit, configured to acquire a set of sample advertisements viewed by the plurality of sample users; and an advertisement clustering unit, configured to cluster the sample advertisements in the set to obtain a plurality of advertisement categories.

[0105] In some embodiments, the first determining unit includes: a first calculation subunit, configured to calculate the rating value of the current sample user for the current advertising category according to the following method: calculating the rating value of the current sample user for the current advertising category based on the behavioral indicator data of the current sample user for all sample advertisements under the current advertising category; and a construction subunit, configured to construct a rating vector corresponding to the sample user based on the rating values ​​of the sample user for each advertising category.

[0106] In some embodiments, the first calculation subunit includes: a second calculation subunit, configured to calculate the rating of the current sample user for each sample advertisement by: calculating the rating of the current sample user for the current sample advertisement based on various behavioral indicator data of the current sample user for the current sample advertisement; and a third calculation subunit, configured to calculate the rating of the current sample user for the current advertisement category based on the rating of each sample advertisement under the current category.

[0107] In some embodiments, the first calculation subunit includes: a third calculation subunit, configured to calculate the comprehensive value of each behavioral indicator data corresponding to the current category: calculate the comprehensive value of the current behavioral indicator data based on the current behavioral indicator data of each sample advertisement under the current category; a first acquisition subunit, configured to acquire the weights corresponding to each type of behavioral indicator data; and a summation subunit, configured to perform weighted summation on the comprehensive values ​​of each type of behavioral indicator data to obtain the rating value of the current sample user for the current advertisement category.

[0108] In some embodiments, the user clustering unit includes: an initialization subunit for randomly initializing n points from all sample users as initial cluster centers; a fourth calculation subunit and an update subunit for cyclically performing operations until a preset cutoff condition is reached, wherein the fourth calculation subunit is used to calculate the distance between each sample user and each cluster center, determine the cluster center with the smallest distance to the sample user, and assign the sample user to a class represented by the distance center; the update subunit is used to redetermine the cluster center of each class based on all sample users in the updated class.

[0109] In some embodiments, the second determining unit includes: a fifth calculation subunit, used to calculate the distance between the target user and each user group using the target user's behavioral indicator data to represent the target user; and a first determining subunit, used to select the user group with the smallest distance as the target user group closest to the target user.

[0110] In some embodiments, the following steps are performed at predetermined intervals: acquiring behavioral indicator data of multiple sample users on each sample advertisement while browsing advertisements; determining the rating value of each sample user for each advertisement category based on the behavioral indicator data, and constructing a rating vector corresponding to each sample user based on the rating values ​​of each sample user for each advertisement category; using the rating vector corresponding to each sample user to represent the sample user, and using a clustering algorithm to cluster multiple sample users to obtain multiple user groups; when the data of a sample user reaches a predetermined threshold, or when the number of sample advertisements viewed by a sample user reaches a predetermined threshold, clearing the earliest sample advertisement behavioral indicator data so that the number of sample advertisements is within a predetermined numerical range.

[0111] In some embodiments, the third determining unit includes: a second determining subunit, configured to determine the M sample advertisements with the highest ratings in the target user group as a first advertisement set, where M is a preset natural number; a second obtaining subunit, configured to obtain actual behavioral indicator data corresponding to each advertisement in the first advertisement set; a third determining subunit, configured to determine the N sample users closest to the target user in the target user group, and use the historical recommendation results of the N sample users as a second advertisement set; a third obtaining subunit, configured to obtain behavioral indicator prediction data corresponding to each advertisement in the second advertisement set; a fourth obtaining subunit, configured to obtain the intersection of the first advertisement set and the second advertisement set; a sixth calculation subunit, configured to calculate the behavioral indicator prediction data of the target advertisements in the intersection based on the first behavioral indicator and the behavioral indicator prediction data; and a filtering subunit, configured to select a target number of advertisements from the intersection as target advertisements recommended to the target user based on the behavioral indicator prediction data.

[0112] The description and beneficial effects of the device for determining the above-mentioned recommended advertisement can be found in the description and beneficial effects of the method section, and will not be repeated here.

[0113] This invention also provides an electronic device, such as... Figure 5 As shown, the electronic device may include a processor 501 and a memory 502, wherein the processor 501 and the memory 502 may be connected via a bus or other means. Figure 5 Taking the example of a connection between China and Israel via a bus.

[0114] Processor 501 can be a central processing unit (CPU). Processor 501 can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations of the above types of chips.

[0115] Memory 502, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the method for determining recommended advertisements in this embodiment of the invention (e.g., Figure 4 The first acquisition unit 10, the first determination unit 20, the user clustering unit 30, the second acquisition unit 40, the second determination unit 50, and the third determination unit 60 are shown. The processor 501 executes various functional applications and data classification by running non-transitory software programs, instructions, and modules stored in the memory 502, thereby implementing the method for determining recommended advertisements in the above method embodiments.

[0116] Memory 502 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by the processor 501, etc. Furthermore, memory 502 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 502 may optionally include memory remotely located relative to processor 501, and these remote memories may be connected to processor 501 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0117] The one or more modules are stored in the memory 502, and when executed by the processor 501, the method for determining the recommended advertisements described above is performed.

[0118] For details regarding the electronic device described above, please refer to the relevant descriptions and effects in the above embodiments for further understanding; they will not be repeated here.

[0119] This specification provides a computer storage medium storing computer program instructions that, when executed by a processor, implement the steps of the method for determining the recommended advertisement described above.

[0120] This specification provides a computer program product comprising a computer program that, when executed by a processor, implements the steps of the method for determining the recommended advertisement described above.

[0121] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include combinations of the above types of memory.

[0122] The various embodiments in this specification are described in a progressive manner. For the same or similar parts between the various embodiments, please refer to each other. The focus of each embodiment is to describe the differences from other embodiments.

[0123] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions.

[0124] For ease of description, the above devices are described separately by function as various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0125] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute certain parts of the methods of various embodiments of this application.

[0126] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc.

[0127] This application can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0128] Although this application has been described through embodiments, those skilled in the art will know that this application has many modifications and variations without departing from the spirit of this application, and it is intended that the appended claims cover such modifications and variations without departing from the spirit of this application.

Claims

1. A method for determining recommended advertisements, characterized in that, include: The system acquires behavioral metrics data of multiple sample users when they browse advertisements; the sample advertisements are those used to pre-acquire the behavioral metrics data of the sample users; the behavioral metrics data refer to multiple metrics data of the sample users' behavior when browsing advertisements. Based on the behavioral indicator data, the rating values ​​of each sample user for each advertising category are determined, and the rating values ​​of the sample users for each advertising category constitute the rating vector corresponding to the sample user. Using the rating vectors corresponding to sample users to represent sample users, a clustering algorithm is used to cluster multiple sample users to obtain multiple user groups; Obtain behavioral metrics data of target users when browsing historical ads; the target users are the users of the ads to be recommended. Based on the behavioral indicator data of the target users, determine the target user group that is most similar to the target users among the multiple user groups; Based on the set of advertisements that each user in the target user group is interested in, determine the advertisements to be recommended to the target users; Based on the set of advertisements that each user in the target user group is interested in, the advertisements recommended to the target users are determined, including: The M sample ads with the highest ratings among the target user group are identified as the first ad set, where M is a preset natural number; Obtain the actual behavioral indicator data corresponding to each advertisement in the first advertisement set; the actual behavioral indicator data is the actual collected behavioral indicator data. Identify the N sample users closest to the target user group, and use the historical recommendation results of the N sample users as the second advertising set; Obtain behavioral indicator prediction data for each advertisement in the second advertisement set; the behavioral indicator prediction data is the predicted data of user behavior indicators for each advertisement obtained by further calculation of the actual collected data, and is not the actual collected data. Obtain the intersection of the first ad set and the second ad set; Based on the actual behavioral indicator data and the predicted behavioral indicator data, the predicted behavioral indicator data of the target advertisement in the intersection is calculated according to the following formula: D=D1×η1+D2×η2, where D is the predicted behavioral indicator data of the target advertisement in the intersection, D1 is the actual behavioral indicator data, and D2 is the predicted behavioral indicator data corresponding to the second advertisement set. Based on the behavioral indicator prediction data D, select the target number of advertisements from the intersection as the target advertisements recommended to the target users.

2. The method according to claim 1, characterized in that, Before determining the rating vector for each advertising category for each sample user based on the aforementioned behavioral indicator data, the process also includes: Obtain a set of sample advertisements viewed by the multiple sample users; Clustering of sample advertisements in the set yields multiple advertisement categories.

3. The method according to claim 1, characterized in that, Based on the behavioral indicator data, determine the rating values ​​of each sample user for each advertising category, including: The rating of the current sample user for the current ad category is calculated using the following method: The rating of the current sample user for the current ad category is calculated based on the behavioral metrics data of the current sample user for all sample ads under the current ad category.

4. The method according to claim 3, characterized in that, Based on the behavioral metrics data of the current sample users towards all sample ads in the current ad category, calculate the current sample user's rating for the current ad category, including: The ratings of current sample users for each sample advertisement are calculated using the following method: based on various behavioral metrics data of current sample users for the current sample advertisement, the ratings of current sample users for the current sample advertisement are calculated. Calculate the rating of the current user for the current ad category based on the ratings of each sample ad in the current category.

5. The method according to claim 3, characterized in that, Based on the behavioral metrics data of the current sample users towards all sample ads in the current ad category, calculate the current sample user's rating for the current ad category, including: Calculate the composite value of each behavioral indicator data corresponding to the current category: Calculate the composite value of the current behavioral indicator data based on the current behavioral indicator data of each sample advertisement under the current category; Obtain the weights corresponding to various behavioral indicator data; The weighted sum of the comprehensive values ​​of various behavioral indicators is used to obtain the rating value of the current sample users for the current advertising category.

6. The method according to claim 1, characterized in that, Clustering algorithms were used to cluster multiple sample users, resulting in multiple user groups, including: Randomly initialize n points from all sample users as initial cluster centers; Repeat the following operation until the preset cutoff condition is met: Calculate the distance between each sample user and each cluster center, determine the cluster center with the smallest distance to the sample user, and assign the sample user to the class represented by the cluster center; Based on all sample users in each category after the update, the cluster centers for each category are redefined.

7. The method according to claim 1, characterized in that, Based on the behavioral metrics data of the target users, identify the target user group among the multiple user groups that is most similar to the target users, including: Using the behavioral metrics data of the target users to represent the target users, calculate the distance between the target users and each user group; The user group with the smallest distance is considered the target user group that is closest to the target user.

8. The method according to claim 1, characterized in that, The process is performed at predetermined intervals as follows: acquiring behavioral indicator data of multiple sample users on each sample advertisement while browsing the advertisement; determining the rating value of each sample user for each advertisement category based on the behavioral indicator data, and constructing a rating vector corresponding to each sample user based on the rating values ​​of each sample user for each advertisement category; using the rating vector corresponding to each sample user to represent the sample user, and using a clustering algorithm to cluster multiple sample users to obtain multiple user groups. When the data of sample users reaches a predetermined threshold, or when the number of sample ads viewed by sample users reaches a predetermined threshold, the earliest sample ad behavior indicator data is cleared so that the number of sample ads remains within a predetermined range.

9. A device for determining recommended advertisements, characterized in that, include: The first acquisition unit is used to acquire behavioral indicator data of multiple sample users when browsing advertisements; the sample advertisements are advertisements for which behavioral indicator data of sample users is acquired in advance; the behavioral indicator data refers to multiple indicator data of the sample users' behavior when browsing advertisements. The first determining unit is used to determine the rating values ​​of each sample user for each advertising category based on the behavioral indicator data, and to construct the rating vector corresponding to the sample user based on the rating values ​​of the sample users for each advertising category. The user clustering unit is used to represent the sample user with the rating vector corresponding to the sample user, and to cluster multiple sample users using a clustering algorithm to obtain multiple user groups; The second acquisition unit is used to acquire behavioral indicator data of the target user on each historical advertisement when browsing historical advertisements; the target user is the user of the advertisement to be recommended. The second determining unit is used to determine the target user group that is most similar to the target user among the multiple user groups based on the target user's behavioral indicator data; The third determining unit is used to determine the advertisements recommended to the target users based on the set of advertisements that each user in the target user group is interested in; The third determining unit includes: The second determining subunit is used to determine the M sample advertisements with the highest ratings among the target user group as the first advertisement set, where M is a preset natural number; The second acquisition subunit is used to acquire the actual behavior indicator data corresponding to each advertisement in the first advertisement set; the actual behavior indicator data is the actual collected behavior indicator data. The third determining subunit is used to determine the N sample users that are closest to the target user in the target user group, and to use the historical recommendation results of the N sample users as the second advertising set; The third acquisition subunit is used to acquire the behavioral indicator prediction data corresponding to each advertisement in the second advertisement set; the behavioral indicator prediction data is the prediction data of user behavior indicator data for each advertisement obtained by further calculation of the actual collected data, and is not the actual collected data. The fourth acquisition subunit is used to obtain the intersection of the first ad set and the second ad set; The sixth calculation subunit is used to calculate the behavioral indicator prediction data of the target advertisement in the intersection based on the actual behavioral indicator data and the behavioral indicator prediction data, according to the following formula: D=D1×η1+D2×η2, where D is the behavioral indicator prediction data of the target advertisement in the intersection, D1 is the actual behavioral indicator data, and D2 is the behavioral indicator prediction data corresponding to the second advertisement set. The filtering subunit is used to select a number of advertisements from the intersection of the behavioral indicator prediction data D as target advertisements to recommend to the target user.

10. The apparatus according to claim 9, characterized in that, The device further includes: The third acquisition unit is used to acquire a set of sample advertisements viewed by the multiple sample users; An advertising clustering unit is used to cluster sample advertisements in the set to obtain multiple advertising categories.

11. The apparatus according to claim 10, characterized in that, The first determining unit includes: The first calculation subunit is used to calculate the rating of the current sample user for the current advertising category according to the following method: calculate the rating of the current sample user for the current advertising category based on the behavioral indicator data of the current sample user for all sample advertisements under the current advertising category; Construct sub-units to form the rating vector corresponding to the sample users based on the rating values ​​of the sample users for each advertising category.

12. The apparatus according to claim 11, characterized in that, The first computing subunit includes: The second calculation subunit is used to calculate the rating of the current sample user for each sample advertisement by the following method: calculating the rating of the current sample user for the current sample advertisement based on various behavioral indicator data of the current sample user for the current sample advertisement; The third calculation subunit is used to calculate the rating of the current sample user for the current advertising category based on the rating of each sample advertisement in the current category.

13. The apparatus according to claim 11, characterized in that, The first computing subunit includes: The third calculation subunit is used to calculate the comprehensive value of each behavioral indicator data corresponding to the current category: calculate the comprehensive value of the current behavioral indicator data based on the current behavioral indicator data of each sample advertisement under the current category; The first acquisition subunit is used to acquire the weights corresponding to various behavioral indicator data; The summation subunit is used to perform a weighted summation of the comprehensive values ​​of various behavioral indicator data to obtain the rating value of the current sample user for the current advertising category.

14. An electronic device, characterized in that, include: A memory and a processor, the processor and the memory being communicatively connected to each other, the memory storing computer instructions, the processor executing the computer instructions to implement the steps of the method according to any one of claims 1 to 8.

15. A computer storage medium, characterized in that, The computer storage medium stores computer program instructions, which, when executed by a processor, implement the steps of the method according to any one of claims 1 to 8.

16. A computer program product, characterized in that, It includes a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1 to 8.

Images