Method and system for personalized advertisement content accurate delivery based on user portrait

Abstract

The application discloses a personalized advertisement content accurate delivery method and system based on user portrait, and belongs to the technical field of intelligent promotion, which combines gender / age and other user portrait features and real-time elevator scene data, dynamically calculates advertisement playing benefit score R, realizes accurate matching of advertisement content and current audience, and improves the conversion efficiency of single advertisement. In addition, the application adopts a double evaluation mechanism of benefit score sorting and advertisement importance, dynamically generates an optimal playing scheme through an H1 value minimization algorithm, guarantees the exposure rights and interests of high-value advertisements, and can optimize the playing strategy in real time according to historical effects. The application introduces a completion proportion standard deviation monitoring, dynamically adjusts the playing order, balances the progress differences of various advertisements, and effectively avoids the long-term accumulation of some advertisements due to low benefit score.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent promotion, and particularly relates to a personalized advertisement content accurate delivery method and system based on user portrait. BACKGROUND

[0002] Since the birth of elevator advertisement in the 1990s, it has experienced three main technical development stages: static plane advertisement: mainly paper or frame poster, relying on limited tags such as building type, passenger flow period, etc. for delivery, there are problems such as extensive delivery, difficult quantification of effect, and content homogenization. Elevator LCD outside advertisement: enterprises such as Focus Media introduce dynamic video advertisement to enhance visual impact, but are still limited by short waiting time (usually <10 seconds) outside the elevator, and cannot accurately match user attributes. Elevator intelligent screen advertisement: new media such as enterprises use digital screens to support dynamic content updates, AI visual recognition, time-based delivery and interactive functions, gradually replacing traditional frame advertisements and outside elevators.

[0003] Although the intelligent screen technology has significantly improved the efficiency of advertisement delivery, there are still the following technical bottlenecks: lack of dynamic analysis of the real-time behavior of the elevator passengers, resulting in low interest prediction accuracy. The delivery mode still wastes resources: the traditional CPD (cost per day) mode is easy to cause "empty elevator playing", and the content generation and scene adaptability are poor: the advertisement content is mostly fixed templates, which cannot be adjusted in real time according to the characteristics of the elevator passengers (such as age, gender). For example, the conversion rate of mother and baby advertisements in the old community decreases by more than 40%, in order to solve the above problems, the present application provides the following technical scheme. SUMMARY

[0004] The purpose of the present application is to provide a personalized advertisement content accurate delivery method and system based on user portrait, which solves the problem of poor flexibility of advertisement delivery and low conversion rate in the prior art.

[0005] The purpose of the present application can be achieved by the following technical scheme:

[0006] The personalized advertisement content accurate delivery method based on user portrait comprises the following steps:

[0007] Step 1, confirm the audience group types of each advertisement and the corresponding portrait data;

[0008] Obtain the proportion ai of each audience group type corresponding to each advertisement;

[0009] i takes a value of 1 to n, and n represents the number of audience group types;

[0010] Step 2, at time t1 before playing the next advertisement, the image information in the elevator is collected by the image collection unit in the elevator; wherein t1 is a preset value;

[0011] analyze the gender and age of each person in the elevator;

[0012] group the people in the elevator into several person groups according to the gender and age of each person, and each person group corresponds to an audience group type;

[0013] obtain each advertisement that can be played by the display unit and record it as a to-be-played advertisement;

[0014] obtain the advertisement play benefit score R corresponding to the composition of the people in the current elevator when each to-be-played advertisement is assumed to be played;

[0015] In the third step, the to-be-played advertisement with the largest corresponding advertisement play benefit score R is used as the finally determined next played advertisement.

[0016] As a further scheme of the present application, when determining the number of audience group types, the proportion of each audience group is sorted in descending order, and then added in descending order of proportion until the comprehensive proportion exceeds a preset proportion coefficient. The audience group participating in the summation calculation is used as the final confirmed audience group type of the advertisement, and a proportion normalization process is performed.

[0017] As a further scheme of the present application, the calculation method of the advertisement play benefit score R is

[0018] Wherein, ai represents the proportion corresponding to each person group, and bi represents the number of each person group among the people in the elevator.

[0019] As a further scheme of the present application, it also includes an advertisement play scheme optimization strategy, including the following steps:

[0020] obtain the advertisement play benefit score R of each advertisement played by each display unit in a preset time range in the past;

[0021] For one display unit, sort each advertisement played by it in descending order of the sum of the corresponding advertisement play benefit scores R, denoted as ek.

[0022] Wherein k represents the order corresponding to each advertisement, and ek represents the sum of the specific advertisement play benefit scores R;

[0023] Sort according to the order from large to small according to the importance of the advertisement, and then determine the display unit for each advertisement in turn according to the order to form several sets of schemes;

[0024] For one of the sets of schemes, after all the advertisements complete the determination of the display unit, calculate the order of each advertisement in the corresponding several display units, and then calculate the sum H of the orders corresponding to each advertisement.

[0025] Calculate the sum H1 of all the corresponding order of the advertisement;

[0026] Calculate the H1 value of each scheme in turn, and determine the scheme with the minimum H1 value as the final determined advertisement playing scheme, and play the advertisement according to the advertisement playing scheme in the next period.

[0027] As a further scheme of the present application, when the number of advertisements to be played changes:

[0028] If it is reduced, delete the playing of the corresponding reduced advertisement in the advertisement playing scheme;

[0029] When the number increases, the increased advertisements are played in each display unit, and then the advertisement playing scheme optimization strategy is executed according to the playing result.

[0030] As a further scheme of the present application, when the number of advertisements to be played and the playing quantity requirement of each advertisement are determined, the playing order of each advertisement is further allocated, and the specific steps are as follows:

[0031] Calculate the advertisement playing benefit score R of each advertisement to be played in turn;

[0032] Obtain the completion ratio cj of each advertisement to be played in the current period; j takes a value from 1 to m, and m represents the number of advertisements to be played;

[0033] The completion ratio refers to the sum of the corresponding advertisement playing benefit score R of each playing in the current period / the sum of the advertisement playing benefit score R to be completed in the current period;

[0034] Calculate the completion ratio cj of each advertisement to be played before playing the new advertisement to be played, and then calculate the standard deviation S1 of the m cj values;

[0035] Calculate the standard deviation S2 of the advertisement to be played corresponding to the updated completion ratio cj after playing each advertisement to be played;

[0036] Calculate the m corresponding |S1-S2| values, and select the S2 value corresponding to the minimum |S1-S2| value, and take the advertisement to be played corresponding to the S2 value as the next played advertisement in the display unit.

[0037] As a further scheme of the present application, it also includes a method for adjusting the number of advertisements to be played in each display unit, which is as follows:

[0038] When a display unit completes the playing of the advertisements to be played in a period;

[0039] First, confirm whether there is time left in the period, if yes, add the number of advertisements to be played in the display unit;

[0040] If not:

[0041] For an advertisement, obtain the order g when its corresponding advertisement play benefit R is sorted in size each time as a to-be-played advertisement;

[0042] Obtain the average value G of the order g corresponding to each play of the advertisement in the entire cycle;

[0043] Obtain the average value Gp of the G values of all to-be-played advertisements after a to-be-played advertisement in a display unit completes a cycle;

[0044] When the average value Gp is greater than a preset value Gy, the number of to-be-played advertisements in the display unit is added.

[0045] The application also discloses a personalized advertisement content accurate delivery system based on user portraits, which is used for executing the delivery method, and comprises:

[0046] An image acquisition unit is configured to acquire scene images in the elevator and transmit the acquired scene image information to an image analysis unit;

[0047] The image analysis unit is configured to analyze the scene images and identify the number of persons, genders and age information in the scene images;

[0048] A general control unit is configured to determine the next played advertisement according to the crowd portrait data in the elevator analyzed by the image analysis unit;

[0049] A display unit is configured to play the advertisement.

[0050] The application has the following beneficial effects:

[0051] The application combines the user portrait features such as gender and age with real-time data of the elevator scene, dynamically calculates the advertisement play benefit score R, realizes accurate matching of the advertisement content and the current audience, and improves the conversion efficiency of single advertisement.

[0052] The application adopts a double evaluation mechanism of benefit score sorting and advertisement importance, dynamically generates an optimal play scheme through an H1 value minimization algorithm, guarantees the exposure rights and interests of high-value advertisements, and can also optimize the play strategy in real time according to historical effects.

[0053] The application balances the progress differences of various advertisements by dynamically adjusting the play order through the introduction of the completion proportion standard deviation monitoring (|S1-S2| minimization principle). Experimental data show that this method can reduce the advertisement progress deviation by about 37% (periodic test in weeks), and effectively avoids the long-term accumulation of some advertisements due to low benefit scores.

[0054] The application intelligently expands the advertisement carrying capacity of the low-efficiency display unit by comparing the advertisement benefit order mean Gp with the threshold Gy. The practical application shows that the strategy can increase the advertisement income of a single elevator screen body by 15-22%.

[0055] The application realizes the optimal balance of precise delivery, dynamic optimization and resource utilization in the elevator advertisement scene through intelligent decision-making in the time-space dual dimension, and is especially suitable for advertisement value mining in high liquidity closed spaces. BRIEF DESCRIPTION OF DRAWINGS

[0056] The application will be further described below with reference to the drawings.

[0057] Figure 1 is a flowchart of the personalized advertisement content precise delivery method based on user portraits according to an embodiment of the application. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0059] Embodiment one

[0060] The personalized advertisement content precise delivery method based on user portraits, as shown in Figure 1 includes the following steps:

[0061] Step one, confirm the audience group types of each advertisement and the portrait data of each audience group type;

[0062] Obtain the proportion ai of each audience group type corresponding to each advertisement;

[0063] i takes the value of 1 to n, and n represents the number of audience group types;

[0064] Since the composition of the audience groups of many products is relatively complex, in addition to the main several audience groups, there are many other audience groups with small proportions, therefore, in order to simplify the calculation, the proportions of each audience group can be sorted in descending order, and then added in descending order of proportion until the comprehensive proportion exceeds the preset proportion coefficient, the audience groups participating in the summation calculation are regarded as the final confirmed audience group types of the advertisement, and the proportions of each audience group type of the final confirmed advertisement are proportionally expanded so that the sum of the proportions of each audience group type of the final confirmed advertisement is 1;

[0065] The audience group portrait data here includes gender and age.

[0066] Specifically, the audience portrait of the product corresponding to the advertisement can be obtained through purchase records, questionnaire surveys and the like. This method of obtaining the audience portrait is a commonly used known technology in research, and thus will not be specifically described and limited herein.

[0067] Secondly, at a time t1 before playing the next advertisement, image information in the elevator is collected by an image collection unit in the elevator, wherein t1 is a preset value.

[0068] The image analysis unit analyzes the image information collected by the image collection unit to obtain the gender and age of each person in the elevator.

[0069] According to the gender and age of each person in the elevator, the persons in the elevator are classified into several person groups, and each person group corresponds to an audience group type.

[0070] The display unit can play each advertisement. In order to make the subsequent description clear, these advertisements are marked as to-be-played advertisements.

[0071] The advertisement play benefit score R corresponding to the composition of the persons in the current elevator when each to-be-played advertisement is played is obtained.

[0072] The calculation method of the advertisement play benefit score R is

[0073] Wherein, ai represents the proportion of each person group, and bi represents the number of each person group in the persons in the elevator.

[0074] After calculation according to the above method, the greater the advertisement play benefit score R of a to-be-played advertisement, the higher the effect and benefit of the play.

[0075] Thirdly, the to-be-played advertisement with the largest corresponding advertisement play benefit score R is used as the finally determined next played advertisement. This method can make full use of the flow of people and maximize the benefit of the advertisement play.

[0076] Fourthly, the advertisement play benefit scores R of the advertisements played by each display unit in a preset time range in the past are obtained.

[0077] For a display unit, each advertisement played by the display unit is sorted in descending order of the sum of the corresponding advertisement play benefit scores R, denoted as ek.

[0078] Wherein, k represents the order of each advertisement, and ek represents the sum of the specific advertisement play benefit scores R.

[0079] In the fifth step, each advertisement is ranked according to the importance score from high to low, and then the display units for each advertisement are determined in sequence according to the ranking;

[0080] According to different selections, multiple sets of schemes are formed;

[0081] For one of the sets of schemes, after the display units for all the advertisements are determined, the order of each advertisement in the corresponding display units is calculated, and then the sum H of the corresponding orders of each advertisement is calculated;

[0082] The sum H1 of the sums H of the corresponding orders of all the advertisements is calculated;

[0083] The H1 values corresponding to each set of schemes are calculated in sequence, and the set of scheme with the minimum H1 value is determined as the final determined advertisement playing scheme, and the advertisement is played according to the advertisement playing scheme in the next period.

[0084] The importance score of the advertisement can be determined according to the income of the advertisement, and the higher the income, the higher the corresponding score.

[0085] When the number of advertisements to be played changes:

[0086] If the number is reduced, the playing of the corresponding reduced advertisements in the advertisement playing scheme is deleted;

[0087] When the number is increased, the increased advertisements are played in each display unit (i.e., they are added to the to-be-played advertisements in each display unit), and then the operations in the fourth and fifth steps are performed according to the playing results.

[0088] Embodiment Two

[0089] Because in actual operation, the method in Embodiment One can improve the delivery effect of some advertisements, but has a negative impact on the delivery effect of other advertisements, therefore, based on Embodiment One, this embodiment further allocates the playing order of each advertisement when the number of to-be-played advertisements and the playing quantity requirement of each advertisement are determined, and the specific steps are as follows:

[0090] According to the method in the second step of Embodiment One, the advertisement playing benefit score R of each to-be-played advertisement is calculated in sequence;

[0091] The completion ratio cj of each to-be-played advertisement in the current period is obtained; where j takes a value from 1 to m, and m represents the number of to-be-played advertisements;

[0092] The period can be one day, one week, or one month, etc.

[0093] The completion ratio refers to the sum of the corresponding advertisement play benefit score R of each play of the to-be-played advertisement in the current period / the sum of the advertisement play benefit score R to be completed in the current period;

[0094] The completion ratio cj of each to-be-played advertisement is calculated before a new to-be-played advertisement is played, and then the standard deviation S1 of the m cj values is calculated;

[0095] The standard deviation S2 of the to-be-played advertisement corresponding to the updated completion ratio cj after playing each to-be-played advertisement is calculated;

[0096] The m corresponding |S1-S2| values (different updated values corresponding to playing different to-be-played advertisements) are calculated, and the to-be-played advertisement corresponding to the S2 value with the minimum |S1-S2| value is selected and played as the next advertisement played by the display unit.

[0097] The present application confirms the next played advertisement through the completion ratio of each advertisement and the real-time updated advertisement play benefit score, so as to ensure that the completion ratio progress of each to-be-played advertisement is similar, thereby ensuring that the to-be-played advertisement with a larger sum of advertisement play benefit score R to be completed can be played under the condition of a larger single advertisement play benefit score R as much as possible, thereby improving the utilization efficiency of human flow resources under the premise of ensuring the play requirements of all advertisements.

[0098] Embodiment Three

[0099] On the basis of embodiment two, the present embodiment further proposes a method for adjusting the number of to-be-played advertisements of each display unit, which is specifically as follows:

[0100] According to the method in embodiment two, when the to-be-played advertisements in a display unit complete the play in a period;

[0101] First, it is confirmed whether there is time left in the period, if yes, the number of to-be-played advertisements is added in the display unit;

[0102] If not:

[0103] For an advertisement, the order g of its corresponding advertisement play benefit R when sorted in size is obtained when it is played as a to-be-played advertisement each time;

[0104] The average value G of the order g corresponding to each play of the advertisement in the entire period is obtained;

[0105] The average value Gp of the G values of all to-be-played advertisements after the play of to-be-played advertisements in a display unit in a period is obtained;

[0106] When the average value Gp is greater than the preset value Gy, it indicates that the utilization efficiency of the corresponding display unit is low, and the number of to-be-played advertisements in the display unit can be added.

[0107] On the contrary, when the average value Gp is not greater than the preset value Gy, it indicates that the utilization efficiency of the corresponding display unit is high, and the number of to-be-played advertisements in the display unit is not adjusted.

[0108] In this way, the income brought by the display unit playing advertisements can be further improved, and the return on investment of advertising can be improved.

[0109] Embodiment Four

[0110] The application also discloses a system for executing the above-mentioned personalized advertisement content accurate placement method based on a user portrait, and the system comprises:

[0111] An image acquisition unit is configured to acquire a scene image in the elevator and transmit the acquired scene image information to the image analysis unit.

[0112] An image analysis unit is configured to analyze the scene image and identify the number of persons, the gender and the age information in the scene image.

[0113] A total control unit is configured to determine the next played advertisement according to the crowd portrait data analyzed by the image analysis unit.

[0114] A display unit is configured to play the advertisement.

[0115] The above content is merely an example and description of the application, and those skilled in the art can make various modifications or supplements or use similar ways to replace the described specific embodiments, as long as the modifications or supplements or replacements do not deviate from the application or exceed the scope defined by the present application, and all of them shall belong to the protection scope of the application.

Claims

1. A method for personalized advertisement content accurate delivery based on user portrait, characterized in that, It comprises the following steps: First, confirm the audience group types of each advertisement and the corresponding portrait data; Obtain the proportion ai of each audience group type corresponding to each advertisement; i takes the value of 1 to n, n represents the number of audience group types; Second, at time t1 before playing the next advertisement, collect image information inside the elevator through the image acquisition unit inside the elevator; Where t1 is a preset value; Analyze and obtain the gender and age of each person inside the elevator; According to the gender and age of each person inside the elevator, divide the people inside the elevator into several personal groups, and each personal group corresponds to an audience group type; Obtain each advertisement that the display unit can play and mark it as a to-be-played advertisement; Obtain the advertisement play benefit score R corresponding to the composition of the current elevator when each to-be-played advertisement is played; Third, the to-be-played advertisement with the maximum corresponding advertisement play benefit score R is used as the final determined next played advertisement; After the third step, the advertisement play scheme optimization strategy is also included, comprising the following steps: Obtain the advertisement play benefit score R of each advertisement played by each display unit in the past preset time range; For a display unit, sort each advertisement it plays in descending order of the sum of the corresponding advertisement play benefit score R, denoted as ek; Where k represents the order of each advertisement, and ek represents the sum of the specific advertisement play benefit score R; Sort the advertisements in descending order of importance, and then determine the display unit for each advertisement in turn according to the order to form several schemes; For one of the schemes, after all the advertisements have determined the display unit, calculate the order of each advertisement in the corresponding several display units, and then calculate the sum of the orders H corresponding to each advertisement; Calculate the sum H1 of the sum of the orders H corresponding to all advertisements; Calculate the H1 value corresponding to each scheme in turn, and determine the scheme with the minimum H1 as the final determined advertisement play scheme, and play the advertisements according to the advertisement play scheme in the next cycle. 2.The method of claim 1, wherein, When determining the number of audience group types, sort the proportions of each audience group in descending order, and then add them in descending order of proportion until the total proportion exceeds the preset proportion coefficient. The audience group participating in the summation calculation is the final confirmed audience group type of the advertisement, and the proportion is normalized. 3.The method of claim 1, wherein, The calculation method of the advertisement playing benefit score R is ; Where ai represents the proportion corresponding to each personal group, and bi represents the number of each personal group among the people inside the elevator. 4.The method of claim 1, wherein, When the number of advertisements to be played changes: If it is reduced on the basis of the existing, delete the play of the corresponding reduced advertisements in the advertisement play scheme; When the number increases, the increased advertisements are played in each display unit, and then the advertisement play scheme optimization strategy is executed according to the play result. 5.The user portrait-based personalized advertisement content accurate delivery method according to claim 1, characterized in that, When the number of to-be-played advertisements and the play quantity requirement of each advertisement are determined, further distribute the play order of each advertisement, and the specific steps are as follows: Calculate the advertisement play benefit score R of each to-be-played advertisement in turn; Obtain the completion ratio cj of each to-be-played advertisement in the current cycle; j takes the value of 1 to m, and m represents the number of to-be-played advertisements; The completion ratio refers to the sum of the corresponding advertisement play benefit score R of each play of the to-be-played advertisement in the current period / the sum of the advertisement play benefit score R to be completed in the current period; Calculate the completion ratio cj of each to-be-played advertisement before playing a new to-be-played advertisement, and then calculate the standard deviation S1 of the m cj values; After playing each to-be-played advertisement, update the completion ratio cj, and then calculate the standard deviation S2 of the to-be-played advertisement corresponding to the updated completion ratio cj; Calculate the m corresponding |S1-S2| values, and select the to-be-played advertisement corresponding to the minimum S2 value of the corresponding |S1-S2| value, and take the to-be-played advertisement as the next played advertisement of the display unit. 6.The method of claim 5, wherein, Also includes a method for adjusting the number of to-be-played advertisements of each display unit, as follows: When a display unit completes the play of to-be-played advertisements in a period; First, confirm whether there is time left in the period, if yes, add the number of to-be-played advertisements in the display unit; If not, for an advertisement, get the corresponding advertisement play benefit R when it is played as a to-be-played advertisement each time, and sort them according to size to get the order g; Get the average value G of the corresponding order g of each play of the advertisement in the entire period; Get the average value Gp of the G values of all to-be-played advertisements after a display unit completes the play of to-be-played advertisements in a period; When the average value Gp is greater than the preset value Gy, add the number of to-be-played advertisements in the display unit.

7. A personalized advertisement content accurate delivery system based on user portrait, characterized in that, The system is used to execute the delivery method of any one of claims 1 to 6, and the system comprises: An image acquisition unit is configured to acquire a scene image in an elevator and transmit the acquired scene image information to an image analysis unit; The image analysis unit is configured to analyze the scene image and identify the number of people, gender, and age information therein; A master control unit is configured to determine the next played advertisement according to the crowd portrait data of the elevator analyzed by the image analysis unit; A display unit is configured to play the advertisement.

Images