Intelligent advertisement putting platform based on data driving

Abstract

The invention belongs to the technical field of advertisement putting, and particularly relates to an intelligent advertisement putting platform based on data driving, which is characterized in that the basic audience degree, the effective conversion degree and the behavior deviation degree of each advertisement category relative to a specified user at the current login time point are evaluated; the method comprises the following steps: screening each suitable advertisement category of target application software relative to a specified user at present, marking each advertisement contained in the target application software as each suitable advertisement, predicting the attraction degree of each suitable advertisement content to the specified user to screen each optimal advertisement, deeply analyzing each advertisement space of the target application software to effectively convert corresponding advertisement visual elements, and displaying the advertisement visual elements. And in combination with the advertiser budget corresponding to each optimal advertisement, determining the optimal advertisement of the specified user corresponding to each advertisement position of the current target application software so as to carry out advertisement putting work, and evaluating the advertisement putting effect of the specified user of the target application software on the current day, thereby realizing complete, accurate and effective advertisement putting. Therefore, the advertisement putting strategy optimization and the effect improvement are facilitated.

Description

Technical Field

[0001] The present invention belongs to the technical field of advertisement delivery, and in particular relates to an advertisement intelligent delivery platform based on data-driven. Background Art

[0002] With the rapid development of Internet technology, the global information flow has increased exponentially. Consumers are exposed to a flood of information every day, which undoubtedly intensifies the competition in the advertising market. In the current complex and dynamically changing market environment, the past methods that relied on subjective experience, large-scale indiscriminate coverage, and unchanging delivery strategies can no longer accurately connect with the target audience and optimize the effectiveness of advertising delivery. Therefore, exploring how to build a precise advertising delivery strategy has become a crucial issue.

[0003] In the prior art, there are also some related solutions involving advertising delivery, for example, a method and platform for advertising delivery with Chinese patent publication number CN109919641B, which obtains text information of a web page, parses the text information of the web page, adds a main scene tag to the web page according to the result of the analysis, determines that the application scene tag of the advertisement in the database is compatible with the main scene tag of the web page, associates the main scene tag of the web page with the application scene tag of the advertisement, and delivers the advertisement to the web page for display to achieve effective advertising delivery.

[0004] Another Chinese patent publication number CN102789619A is a method and an advertising platform device for targeted advertising delivery. It receives a request for advertising encapsulation of media content initiated by a value-added service device through an advertising platform device, determines the advertising position corresponding to the service identifier of the media content, queries the advertising content and target users that match the advertising position, groups the target users according to the combination of the advertising position and the advertising content, and groups the target users with the same advertising content in the same advertising position into one group, encapsulates the media content for advertising, and sends the encapsulated media content and the corresponding target user group to the value-added service device. There is no need to modify the existing network process, and personalized targeted advertising can be achieved according to user preferences.

[0005] Although the above scheme proposes some solutions related to advertising delivery, the existing technology still has the following limitations, specifically: 1. The existing technology focuses on formulating advertising delivery strategies based on the basic user portrait to screen user-adaptive advertising content, specifically covering basic characterization parameters such as user age, preference, and consumption ability. It lacks an integrated screening mechanism for time, advertising category and advertising content. It lacks sufficient consideration of the differences in users' acceptance and preference for advertising categories at different time points, and the different advertising content and user behavior motivations under the same advertising category. It is difficult to accurately match the dynamic changes of advertising delivery and user needs in multiple dimensions.

[0006] 2. The existing technology for formulating advertising delivery strategies focuses on the mapping and comparison between ad slot attributes and advertising content materials, but ignores the analysis of visual elements involved in the effective conversion of ad slots. It fails to fully understand that placing specific ads in ad slots can prompt users to click and convert. In addition to the ad content, the ad cover image plays a key visual appeal role. However, the lack of understanding and utilization of user visual preferences makes it difficult for advertising to accurately reach the target user group, reducing the conversion efficiency of advertising. Summary of the invention

[0007] In order to overcome the shortcomings of the background technology, the embodiments of the present invention provide a data-driven intelligent advertising delivery platform, which can effectively solve the problems involved in the above-mentioned background technology.

[0008] The purpose of the present invention can be achieved through the following technical solutions: an advertising intelligent delivery platform based on data-driven, including: a user login response module, an advertising category screening module, a preferred advertising screening module, an advertising delivery determination module, a delivery effectiveness evaluation module and a cloud database.

[0009] The user login response module is connected to the advertisement category screening module, the advertisement category screening module is connected to the optimal advertisement screening module, the optimal advertisement screening module is connected to the advertisement placement determination module, the advertisement placement determination module is connected to the placement effectiveness evaluation module, and the cloud database is respectively connected to the advertisement category screening module, the optimal advertisement screening module, and the placement effectiveness evaluation module.

[0010] The user login response module is used to respond to the login operation of the specified user of the target application software, and determine the click price of each advertising position, the reference conversion rate of each advertising category and the reference audience group type of the target application software corresponding to the current login time point of the specified user.

[0011] The advertising category screening module is used to construct a basic portrait of the designated user and analyze the authorized behavior, evaluate the basic audience level, effective conversion level and behavioral bias of each advertising category relative to the designated user at the current login time, and thereby screen the appropriate advertising categories for the target application software relative to the designated user.

[0012] The optimal advertisement screening module is used to extract the advertisements contained in the various suitable advertisement categories currently for the designated user in the cooperative advertisement library of the target application software, organize and mark them as various suitable advertisements, predict the attractiveness of the content of each suitable advertisement to the designated user, and screen the various optimal advertisements currently for the designated user of the target application software.

[0013] The advertisement placement determination module is used to analyze the visual elements of advertisements corresponding to the effective conversion of each advertisement position of the target application software, match and analyze each optimal advertisement with each advertisement position, determine the best advertisement for each advertisement position of the current target application software corresponding to the designated user, and thereby carry out advertisement placement work on the target application software page of the designated user.

[0014] The delivery effectiveness evaluation module is used to collect the interaction data of the designated users on the day for the advertisements delivered in the target application software, and evaluate the delivery effectiveness of the advertisements delivered to the designated users of the target application software on the day.

[0015] The cloud database is used to store the occupational categories corresponding to each preset age range and each basic consumption ability level, store the corresponding image element set and the relevant keyword set for each advertising category, store the preset suitability coefficient benchmark threshold of the software advertising category suitability standard specification, store the preset content attractiveness benchmark threshold of the software advertising optimal investment standard specification, and store the user benchmark response interaction threshold.

[0016] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention screens the appropriate advertising categories for the target application software relative to the specified user by evaluating the basic audience level, effective conversion level and behavioral bias level of each advertising category relative to the specified user at the current login time point, and uses this as a starting point to ensure the reliability of subsequent analysis of the appropriate advertising content, thereby greatly improving the accuracy and effectiveness of advertising delivery.

[0017] (2) The present invention combines the recent behavioral tendencies and current behavioral motivations of designated users to predict the attractiveness of each suitable advertising content to designated users. This not only helps advertisers to understand in advance the degree of compatibility between advertising content and user interests, thereby more accurately selecting and optimizing suitable advertising plans and avoiding the waste of resources caused by blind advertising, but also can significantly improve the pertinence and effectiveness of advertising, and maximize the attention and response willingness of designated users to advertising.

[0018] (3) The present invention deeply analyzes the advertising visual elements corresponding to each advertising position of the target application software when achieving effective conversion, explores the key influence of the advertising visual level on the conversion effect, combines the advertiser budget corresponding to each preferred advertisement, and determines the optimal advertisement for each advertising position of the current target application software corresponding to the designated user, which helps to significantly improve the expected revenue and actual conversion effect of advertising placement, and provides a valuable, scientific and reliable basis and solution for advertising operation decision-making.

[0019] (4) The present invention helps to instantly understand the implementation of advertising strategies for specific user groups on the same day by evaluating the effectiveness of advertising delivery for designated users of target application software on the same day, so as to quickly discover potential problems and deficiencies, and provide highly targeted and timely data support for the adjustment and optimization of subsequent advertising delivery plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0021] Figure 1 It is a schematic diagram of module connection of the present invention.

[0022] Figure 2 A logical schematic diagram of the present invention for screening the various suitable advertising categories of target application software currently relative to a designated user.

[0023] Figure 3 A logical schematic diagram of the present invention for screening the currently preferred advertisements of the target application software relative to the designated user. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] Reference Figure 1 As shown, the present invention provides an advertising intelligent delivery platform based on data-driven, including: a user login response module, an advertising category screening module, a preferred advertisement screening module, an advertisement delivery determination module, a delivery effectiveness evaluation module and a cloud database.

[0026] The user login response module is connected to the advertisement category screening module, the advertisement category screening module is connected to the optimal advertisement screening module, the optimal advertisement screening module is connected to the advertisement placement determination module, the advertisement placement determination module is connected to the placement effectiveness evaluation module, and the cloud database is respectively connected to the advertisement category screening module, the optimal advertisement screening module, and the placement effectiveness evaluation module.

[0027] The user login response module is used to respond to the login operation of the specified user of the target application software, and determine the click price of each advertising position, the reference conversion rate of each advertising category and the reference audience group type of the target application software corresponding to the current login time point of the specified user.

[0028] It should be noted that the price per click of each ad spot corresponding to the current login time point of the specified user in the above-mentioned target application software is evaluated based on the conversion performance of each ad spot within the historical preset period. The specific determination process is: extract the exposure and click volume of each ad spot in each preset time period within the historical preset period of the target application software, perform a ratio analysis between the exposure volume and the preset reference exposure volume, perform a ratio analysis between the click volume and the preset reference click volume, and accumulate the results of the two ratio analyses to obtain the value coefficient of each ad spot in each preset time period within the historical preset period of the target application software, according to the value coefficient interval and price per click corresponding to each value level set for the ad spot of the target application software, obtain the price per click corresponding to the value level of each ad spot in each preset time period within the historical preset period of the target application software, find the preset time period corresponding to the current login time point of the specified user, and thereby determine the price per click of each ad spot corresponding to the current login time point of the specified user in the target application software.

[0029] Reference Figure 2 As shown, the advertising category screening module is used to construct a basic portrait and perform authorization behavior analysis on the designated user, evaluate the basic audience level, effective conversion level and behavioral bias level of each advertising category relative to the designated user at the current login time point, and thereby screen the appropriate advertising categories for the target application software relative to the designated user.

[0030] Specifically, the evaluation of the basic audience level of each advertising category relative to the designated user at the current login time point includes: extracting the registration information of the designated user in the target application software, including age, gender and occupation.

[0031] According to the preset age ranges and basic consumption ability levels stored in the cloud database, the occupational categories are divided accordingly, and the preset age range and basic consumption ability level of the specified user are determined. Combined with the gender of the specified user, the basic portrait of the specified user is constructed and matched with its group type. The advertising category with the reference audience group type of the specified user matching group type at the current login time point is obtained, and the basic audience level of the advertising category relative to the specified user at the current login time point is set to , and set the basic audience level of other advertising categories relative to the specified user to , , thereby obtaining the basic audience level of each advertising category relative to the specified user at the current login time.

[0032] It should be noted that the specific process of determining the reference audience group type for each advertising category corresponding to the current login time point of the specified user in the above-mentioned target application software is as follows: extracting the conversion user registration information corresponding to each effective conversion of each advertising category within the preset period of the overall user history of the target application software, which is consistent with the process of building a basic portrait of the specified user and matching its group type, obtaining the conversion user group type corresponding to each effective conversion of each advertising category within the preset period of the overall user history of the target application software, extracting the conversion user group type corresponding to each effective conversion of each advertising category within the preset time period of the current login time point within the preset period of the overall user history of the target application software, and screening the user group type with the largest number of effective conversions as the reference audience group type, thereby determining the reference audience group type for each advertising category corresponding to the current login time point of the target application software.

[0033] Specifically, the evaluation of the effective conversion degree of each advertising category relative to the specified user at the current login time point includes: extracting the recent behavior information of the specified user in the target application software, including the high-frequency access content set and the deep participation content set in the historical preset period, and the effective conversion number for each advertising category in the historical preset period. and the time points of each effective conversion, is the number of each advertising category, .

[0034] Divide into preset time periods, count the number of effective conversions for each advertising category in each preset time period within the preset historical period of the specified user, and extract the number of effective conversions for each advertising category in the preset time period at the current login time point , calculate the effective conversion rate of a specified user at the current login time relative to each advertising category , ,in For a single user of the target application software within a preset historical period, The average effective conversion number of each advertising category can be obtained by using the effective conversion number of each advertising category within the preset historical period of the overall users of the target application software as mentioned below and performing a ratio analysis with the cumulative number of registered users of the target application software.

[0035] It should be noted that the specific determination principle of the reference conversion rate of each advertising category corresponding to the current login time point of the specified user in the above-mentioned target application software is similar to the analysis principle of the effective conversion rate of each advertising category of the specified user at the current login time point. The specific determination process is: extract the effective conversion times of each advertising category within the preset historical period of the overall user of the target application software , count the effective conversion times of each advertising category in each preset period within the preset historical period of the overall user of the target application software, and extract the effective conversion times of each advertising category in the preset period at the current login time point of the specified user , according to the formula Analyze the reference conversion rate of each advertising category of the target application software for the current login time point of the specified user, where The total number of users of the target application software in the previous historical preset period before this historical preset period for the first The number of effective conversions for each advertising category.

[0036] Compare the effective conversion rate of each advertising category for the specified user at the current login time with the reference conversion rate of each advertising category for the target application software at the current login time of the specified user For comparison, the formula Analyze the effective conversion level of each advertising category relative to the specified user at the current login time.

[0037] It should be noted that the basis for introducing the reference conversion rate of the target application software corresponding to each advertising category at the current login time point in the analysis process of the effective conversion degree of each advertising category relative to the designated user at the current login time point is that the reference conversion rate of the target application software corresponding to each advertising category at the current login time point of the designated user is actually equivalent to the reference conversion rate of the target application software for each advertising category at the current login time point of the overall user, reflecting the popularity of each advertising category among the vast user group at the current login time point. As a basic reference indicator, it shows the potential conversion ability of each advertising category itself. The comparison of the effective conversion rate of each advertising category for the designated user at the current login time point and the reference conversion rate of each advertising category for the target application software at the current login time point of the designated user can more convincingly present the preference of the designated user for each advertising category relative to the overall user level at the current login time point, and can also more clearly illustrate the effective conversion status of each advertising category relative to the designated user at the current login time point.

[0038] Specifically, the evaluation of the degree of behavioral inclination of each advertising category relative to the designated user at the current login time point includes: extracting the high-frequency access content set and the deep-involvement content set within a historical preset period from the recent behavioral information of the designated user in the target application software, identifying and extracting the image elements and keywords in the high-frequency access content set and the deep-involvement content set, obtaining each image element and each keyword of the recent behavioral inclination of the designated user in the target application software, comparing with the image element set and the keyword set corresponding to each advertising category stored in the cloud database, and obtaining the cumulative number of image elements involved in each advertising category of the recent behavioral inclination of the designated user in the target application software. and the number of keywords involved , according to the formula Analyze the degree of behavioral bias of each advertising category relative to a specified user at the current login time.

[0039] Specifically, the screening of the currently suitable advertising categories for the designated user of the target application software includes: accumulating the basic audience degree, effective conversion degree and behavioral bias degree of each advertising category relative to the designated user at the current login time point and the product of their corresponding preset weights, so as to obtain the suitability coefficient of each advertising category relative to the designated user at the current login time point, and screening the advertising categories whose suitability coefficient is greater than or equal to the preset suitability coefficient benchmark threshold of the software advertising category suitability standard specification stored in the cloud database as the currently suitable advertising categories for the designated user of the target application software.

[0040] The embodiment of the present invention screens the appropriate advertising categories for the target application software relative to the designated user by evaluating the basic audience level, effective conversion level and behavioral bias level of each advertising category relative to the designated user at the current login time point, and uses this as a starting point to ensure the reliability of subsequent appropriate advertising content analysis, thereby greatly improving the accuracy and effectiveness of advertising delivery.

[0041] Reference Figure 3 As shown, the preferred advertisement screening module is used to extract the advertisements contained in the various suitable advertisement categories currently relative to the designated user in the target application software cooperative advertisement library, organize and mark them as various suitable advertisements, predict the attractiveness of the content of each suitable advertisement to the designated user, and screen the various preferred advertisements currently relative to the designated user of the target application software.

[0042] Specifically, predicting the attractiveness of each suitable advertising content to a designated user includes: identifying each image element and each keyword in each suitable advertising content, comparing them with each image element and each keyword identified and extracted from a high-frequency access content set and a deep-engagement content set within a historical preset period, obtaining image element feature similarity and keyword semantic similarity of each suitable advertising content relative to the designated user's recent behavioral tendency, accumulating the image element feature similarity and keyword semantic similarity to predict the first-order attractiveness of each suitable advertising content to the designated user.

[0043] It should be noted that the specific process of obtaining the feature similarity of the image elements of the above-mentioned suitable advertising contents relative to the recent behavioral tendencies of the specified user is as follows: the image elements generated by the identification and extraction of the high-frequency access content set and the deep participation content set within the historical preset period are imported into the image analysis software in advance, and then the image elements in a suitable advertising content are imported, and the basic feature similarities of each image element in the suitable advertising content and each image element in the comparison image element sequence are output one by one according to the image feature similarity tool in the image analysis software, and the maximum value is selected as the feature similarity of each image element in the suitable advertising content and the comparison image element sequence, and the image element feature similarity of the suitable advertising content relative to the recent behavioral tendencies of the specified user is obtained after mean calculation, thereby obtaining the image element feature similarity of each suitable advertising content relative to the recent behavioral tendencies of the specified user.

[0044] The keyword semantic similarity of each suitable advertising content relative to the recent behavioral tendency of a specified user is obtained by importing the comparison keyword sequence generated by identifying and extracting each keyword from the high-frequency access content set and the deep participation content set within a preset historical period into the language processing software in advance, and then importing each keyword in a suitable advertising content. According to the semantic similarity tool in the language processing software, the basic semantic similarity of each keyword in the suitable advertising content and each keyword in the comparison keyword sequence is output one by one, and the maximum value is selected as the semantic similarity of each keyword in the suitable advertising content and the comparison keyword sequence. After mean calculation, the keyword semantic similarity of the suitable advertising content relative to the recent behavioral tendency of the specified user is obtained, thereby obtaining the keyword semantic similarity of each suitable advertising content relative to the recent behavioral tendency of the specified user.

[0045] Exemplarily, the image analysis software may be specifically MATLAB or ImageJ, and the language processing software may be specifically NLTK.

[0046] Retrieve the authorized access behavior content for other integrated software on the platform within a preset period before the current login time point of the specified user, extract each image element and each keyword therein, compare them with each image element and each keyword in each suitable advertising content, obtain the image element feature similarity and keyword semantic similarity of each suitable advertising content relative to the current behavioral motivation of the specified user, and accumulate them to predict the second-order attractiveness of each suitable advertising content to the specified user.

[0047] The first-order and second-order attraction levels of each suitable advertising content for a designated user are accumulated with the product of their corresponding preset weights to obtain the attraction level of each suitable advertising content for the designated user.

[0048] The embodiments of the present invention combine the recent behavioral tendencies and current behavioral motivations of designated users to predict the attractiveness of each suitable advertising content to designated users. This not only helps advertisers to understand in advance the degree of fit between advertising content and user interests, thereby more accurately selecting and optimizing suitable advertising plans and avoiding the waste of resources caused by blind advertising, but also can significantly improve the pertinence and effectiveness of advertising, and maximize the attention and response willingness of designated users to advertising.

[0049] Specifically, the current screening condition of the target application software for each preferential advertisement for a designated user is that the attractiveness of each preferential advertisement content to the designated user is greater than or equal to a preset content attractiveness benchmark threshold of the software advertisement preferential standard specification stored in the cloud database.

[0050] The advertisement placement determination module is used to analyze the advertisement visual elements corresponding to the effective conversion of each advertisement position of the target application software, match and analyze each optimal advertisement with each advertisement position, determine the best advertisement for each advertisement position of the current target application software corresponding to the designated user, and thereby carry out advertisement placement work for the target application software page of the designated user.

[0051] Specifically, the analysis of the visual elements of advertisements corresponding to the effective conversions of each advertisement slot of the target application software includes: retrieving the cover display image of each effective conversion advertisement of each advertisement slot within the historical preset period of the target application software, converting the cover display image into the HSV color space to obtain the brightness value and saturation value of each pixel in the cover display image, and accordingly analyzing the light and dark contrast and purity contrast of the cover display image of each effective conversion advertisement of each advertisement slot within the historical preset period of the target application software, retrieving the light and dark contrast interval and purity contrast interval of the cover display image of the advertisement corresponding to the effective conversion of each advertisement slot of the target application software, and the light and dark contrast interval and purity contrast interval of the cover display image of the advertisement are used together as the visual elements of the advertisement.

[0052] It should be noted that the specific analysis process of the light-dark contrast ratio of the cover display image of each effective conversion advertisement in each ad position within the historical preset period of the above-mentioned target application software is as follows: extract the brightness value of each pixel in the cover display image of a certain effective conversion advertisement in a certain ad position within the historical preset period of the target application software, and calculate the average brightness value of the pixels in the cover display image. and brightness standard deviation ,by , Plan the upper pixel brightness threshold corresponding to the dark area of ​​the cover display image and the lower pixel brightness threshold corresponding to the bright area respectively. is a preset constant, obtain the brightness value of each pixel in the dark area and the brightness value of each pixel in the bright area of ​​the cover display image, calculate the average brightness value of the bright area and the average brightness value of the dark area in the cover display image, perform ratio analysis on the difference between the two and the sum of the two, and obtain the brightness and darkness contrast of the cover display image of the effective conversion advertisement of the ad position within the historical preset period of the target application software, thereby obtaining the brightness and darkness contrast of the cover display image of each effective conversion advertisement of each ad position within the historical preset period of the target application software.

[0053] It should also be explained that the specific process of retrieving the light and dark contrast intervals of the cover display images of advertisements corresponding to effective conversions of each ad slot of the target application software is as follows: taking the light and dark contrast of the cover display images of each effective conversion advertisement of each ad slot within the historical preset period of the target application software as the median of the interval, extending and shifting before and after the preset light and dark contrast, constructing a reference interval of the light and dark contrast of the cover display images of each effective conversion advertisement of each ad slot within the historical preset period of the target application software, comparing the reference intervals of the light and dark contrast of the cover display images of each effective conversion advertisement of the same ad slot, and selecting the reference interval segment with the largest number of overlaps as the light and dark contrast interval of the cover display images of advertisements corresponding to effective conversions of the ad slot of the target application software, thereby obtaining the light and dark contrast interval of the cover display images of advertisements corresponding to effective conversions of each ad slot of the target application software.

[0054] It should be noted in particular that the purity contrast of the cover display images of each effective conversion advertisement in each ad slot within the historical preset period of the target application software is consistent with the analysis principle of light and dark contrast, and the purity contrast interval of the cover display images of the advertisements corresponding to the effective conversions of each ad slot of the target application software is consistent with the analysis principle of light and dark contrast interval, which will not be elaborated here.

[0055] Specifically, the method of determining the best advertisement for each advertisement slot of the current target application software corresponding to the designated user includes: extracting the click price budget of the advertiser's cooperation specification to which each preferred advertisement belongs, comparing it with the click price of each advertisement slot corresponding to the current login time point of the designated user of the target application software, and screening each preferred advertisement of the current target application software that meets the click budget for each advertisement slot.

[0056] Calculate the light-dark contrast and purity contrast of the cover display images of each optimal advertisement that meets the click budget of a certain ad slot, compare them with the light-dark contrast range and purity contrast range of the cover display images of the advertisements corresponding to the effective conversion of the ad slot, analyze the visual element matching degree of each optimal advertisement that meets the click budget of the ad slot, select the optimal advertisement with the largest visual element matching degree and corresponding click budget that meets the standard as the best delivery advertisement for the designated user of the ad slot, thereby determining the best delivery advertisement for the designated user of each ad slot of the current target application software.

[0057] It should be noted that the specific analysis process of the visual element matching degree of each preferred advertisement that meets the click budget of the above-mentioned ad position is: extract the upper and lower limits of the light and dark contrast interval of the advertisement cover display image corresponding to the effective conversion of the ad position and calculate the average, so as to obtain the reasonable reference light and dark contrast of the advertisement cover display image corresponding to the effective conversion of the ad position. , and subtract it from the lower limit of the brightness contrast interval of the advertisement cover display image corresponding to the effective conversion of the advertisement position to obtain the reasonable deviation brightness contrast of the advertisement cover display image corresponding to the effective conversion of the advertisement position Similarly, according to the purity contrast interval of the advertisement cover display image corresponding to the effective conversion of the advertisement position, obtain the reasonable reference purity contrast of the advertisement cover display image corresponding to the effective conversion of the advertisement position and reasonable deviation purity contrast , combined with the cover display pictures of each preferred advertisement that has reached the target of click budget for this ad slot, the light and dark contrast is calculated. and purity contrast , The IDs of the preferred ads that have reached the target click budget for this ad slot. , according to the formula Analyze the matching degree of visual elements of each optimal advertisement that meets the click budget target of the ad slot.

[0058] The embodiment of the present invention deeply analyzes the advertising visual elements corresponding to each advertising position of the target application software when achieving effective conversion, explores the key influence of the advertising visual level on the conversion effect, combines the advertiser budget corresponding to each preferred advertisement, and determines the best advertisement for each advertising position of the current target application software corresponding to the designated user, which helps to greatly improve the expected revenue and actual conversion effect of advertising delivery, and provides a valuable, scientific and reliable basis and solution for advertising operation decision-making.

[0059] The delivery effectiveness evaluation module is used to collect the interaction data of the designated users on the day for the advertisements delivered in the target application software, and evaluate the delivery effectiveness of the advertisements delivered to the designated users of the target application software on the day.

[0060] Specifically, the evaluation of the effectiveness of advertising delivery to designated users of the target application software on that day includes: analyzing the response interactivity of designated users to the advertisements delivered in the target application software on that day based on the interaction data of designated users to the advertisements delivered in the target application software on that day, including the conversion volume, sharing volume, like volume and comment volume, performing a ratio analysis with the user benchmark response interaction threshold stored in the cloud database, and using the ratio analysis result as an evaluation indicator for the effectiveness of advertising delivery to designated users of the target application software on that day.

[0061] It should be noted that the above-mentioned response interaction analysis process of the designated user on that day to the advertisements placed in the target application software is to perform ratio analysis on the conversion volume, sharing volume, like volume and comment volume of the designated user on that day to the advertisements placed in the target application software with the reference conversion volume, reference sharing volume, reference like volume and reference comment volume of the response interaction evaluation standard specification stored in the cloud database, and accumulate the results of each ratio analysis.

[0062] The embodiment of the present invention helps to instantly understand the implementation of advertising strategies for specific user groups on the day by evaluating the advertising delivery effectiveness for designated users of target application software on the day, so as to quickly discover potential problems and deficiencies, and provide highly targeted and timely data support for the adjustment and optimization of subsequent advertising delivery plans.

[0063] The cloud database is used to store occupational categories corresponding to each preset age range and each basic consumption ability level, store image element sets and keyword sets corresponding to each advertising category, store preset suitability coefficient benchmark thresholds of software advertising category suitability standards, store preset content attractiveness benchmark thresholds of software advertising optimal standards, store user benchmark response interaction thresholds, store reference conversion volumes, reference sharing volumes, reference like volumes and reference comment volumes of response interaction evaluation standards, and store value coefficient ranges and click prices corresponding to each value level set for the target application software ad slot.

[0064] The data sources in the cloud database of this embodiment are shown in Table 1 below.

[0065] Table 1 Detailed description of data sources in cloud database

[0066]

[0067] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. A data-driven intelligent advertising delivery platform, characterized in that: include: The user login response module is used to respond to the login operation of the specified user of the target application software, and determine the click price of each advertising position, the reference conversion rate of each advertising category and the reference audience group type of the target application software corresponding to the current login time point of the specified user; Advertisement category screening module, used to build basic portraits and conduct authorization behavior analysis for designated users, evaluate the basic audience level, effective conversion level, and behavioral bias of each advertisement category relative to the designated user at the current login time point, and thereby screen the appropriate advertisement categories for the target application software relative to the designated user; The optimal advertisement screening module is used to extract the advertisements contained in the currently suitable advertisement categories for the designated user in the cooperative advertisement library of the target application software, sort and mark them as suitable advertisements, predict the attractiveness of the content of each suitable advertisement to the designated user, and screen the currently preferred advertisements for the designated user of the target application software; The advertisement placement determination module is used to analyze the visual elements of advertisements corresponding to the effective conversion of each advertisement position of the target application software, match and analyze each optimal advertisement with each advertisement position, determine the best advertisement for each advertisement position of the current target application software corresponding to the designated user, and carry out advertisement placement work for the target application software page of the designated user; The delivery effectiveness evaluation module is used to collect the interaction data of the designated users on the day for the advertisements delivered in the target application software, and evaluate the delivery effectiveness of the advertisements delivered to the designated users of the target application software on the day; The cloud database is used to store the occupational categories corresponding to each preset age range and each basic consumption ability level, store the corresponding image element set and the relevant keyword set for each advertising category, store the preset suitability coefficient benchmark threshold of the software advertising category suitability standard specification, store the preset content attractiveness benchmark threshold of the software advertising optimal investment standard specification, and store the user benchmark response interaction threshold.

2. The data-driven intelligent advertising delivery platform according to claim 1, characterized in that: The evaluating of the basic audience level of each advertising category relative to the designated user at the current login time point includes: extracting the registration information of the designated user in the target application software, including age, gender and occupation; According to the preset age ranges and basic consumption ability levels stored in the cloud database, the occupational categories are divided accordingly, and the preset age range and basic consumption ability level of the specified user are determined. Combined with the gender of the specified user, the basic portrait of the specified user is constructed and matched with its group type. The advertising category with the reference audience group type of the specified user matching group type at the current login time point is obtained, and the basic audience level of the advertising category relative to the specified user at the current login time point is set to , and set the basic audience level of other advertising categories relative to the specified user to , , thereby obtaining the basic audience level of each advertising category relative to the specified user at the current login time.

3. The data-driven intelligent advertising delivery platform according to claim 1, characterized in that: The evaluation of the effective conversion degree of each advertising category relative to the designated user at the current login time point includes: extracting the designated user's recent behavior information in the target application software, including the high-frequency access content set and the deep engagement content set in the historical preset period, the number of effective conversions for each advertising category in the historical preset period, and the time point of each effective conversion; Divide each preset time period, count the number of effective conversions for each advertising category in each preset time period within the preset historical period of the specified user, extract the number of effective conversions for each advertising category in the preset time period at the current login time point, and calculate the effective conversion rate of the specified user relative to each advertising category at the current login time point; The effective conversion rate of each advertising category for a specified user at the current login time. Reference conversion rate of each advertising category corresponding to the target application software at the current login time of the specified user Make a comparison, is the number of each advertising category, , according to the formula Analyze the effective conversion level of each advertising category relative to the specified user at the current login time.

4. The data-driven intelligent advertising delivery platform according to claim 3, characterized in that: The evaluation of the degree of behavioral inclination of each advertising category relative to the designated user at the current login time point includes: extracting a high-frequency access content set and a deep-engagement content set within a historical preset period from the recent behavioral information of the designated user in the target application software, identifying and extracting image elements and keywords in the high-frequency access content set and the deep-engagement content set, obtaining each image element and each keyword of the recent behavioral inclination of the designated user in the target application software, comparing with the image element set and the keyword set corresponding to each advertising category stored in the cloud database, and obtaining the cumulative number of image elements involved in each advertising category of the recent behavioral inclination of the designated user in the target application software and the number of keywords involved , according to the formula Analyze the degree of behavioral bias of each advertising category relative to a specified user at the current login time.

5. The data-driven intelligent advertising delivery platform according to claim 1, characterized in that: The method of screening the currently suitable advertising categories of the target application software relative to the designated user includes: accumulating the basic audience degree, effective conversion degree and behavioral bias degree of each advertising category relative to the designated user at the current login time point and the product of their corresponding preset weights to obtain the suitability coefficient of each advertising category relative to the designated user at the current login time point, and screening the advertising categories whose suitability coefficient is greater than or equal to the preset suitability coefficient benchmark threshold of the software advertising category suitability standard specification stored in the cloud database as the currently suitable advertising categories of the target application software relative to the designated user.

6. The data-driven intelligent advertising delivery platform according to claim 4, characterized in that: The predicting the attractiveness of each suitable advertisement content to the designated user includes: identifying each image element and each keyword in each suitable advertisement content, comparing each image element and each keyword identified and extracted with the high-frequency access content set and the deep engagement content set within the historical preset period, obtaining the image element feature similarity and keyword semantic similarity of each suitable advertisement content relative to the recent behavior preference of the designated user, accumulating the image element feature similarity and keyword semantic similarity to predict the first-order attractiveness of each suitable advertisement content to the designated user; Retrieve the authorized access behavior content for other integrated software on the platform within a preset period before the current login time point of the specified user, extract each image element and each keyword therein, and compare them with each image element and each keyword in each suitable advertisement content, respectively, obtain the image element feature similarity and keyword semantic similarity of each suitable advertisement content relative to the current behavior motivation of the specified user, and accumulate them to predict the second-order attraction degree of each suitable advertisement content to the specified user; The first-order and second-order attraction levels of each suitable advertising content for a designated user are accumulated with the product of their corresponding preset weights to obtain the attraction level of each suitable advertising content for the designated user.

7. The data-driven intelligent advertising delivery platform according to claim 1, characterized in that: The screening condition of the target application software's current preferential advertisements for the designated user is that the attractiveness of the preferential advertisement content to the designated user is greater than or equal to the preset content attractiveness benchmark threshold of the software advertisement preferential standard specification stored in the cloud database.

8. The data-driven intelligent advertising delivery platform according to claim 1, characterized in that: The method of analyzing the visual elements of advertisements corresponding to effective conversions of each advertisement slot of the target application software includes: retrieving the cover display images of each effective conversion advertisement of each advertisement slot within a historical preset period of the target application software, converting the cover display images into an HSV color space to obtain the brightness value and saturation value of each pixel in the cover display images, analyzing the light-dark contrast and purity contrast of the cover display images of each effective conversion advertisement of each advertisement slot within a historical preset period of the target application software, retrieving the light-dark contrast interval and purity contrast interval of the cover display images of advertisements corresponding to effective conversions of each advertisement slot of the target application software, and using the light-dark contrast interval and purity contrast interval of the cover display images of advertisements as the visual elements of the advertisements.

9. The data-driven intelligent advertising delivery platform according to claim 8, characterized in that: Determining the optimal advertisement for each advertisement slot of the current target application software corresponding to the designated user includes: extracting the click price budget of the advertiser's cooperation specification of each optimal advertisement, comparing it with the click price of each advertisement slot corresponding to the target application software at the current login time of the designated user, and screening the optimal advertisements of the current target application software that meet the click budget of each advertisement slot; Calculate the light-dark contrast and purity contrast of the cover display images of each optimal advertisement that meets the click budget of a certain ad slot, compare them with the light-dark contrast range and purity contrast range of the cover display images of the advertisements corresponding to the effective conversion of the ad slot, analyze the visual element matching degree of each optimal advertisement that meets the click budget of the ad slot, select the optimal advertisement with the largest visual element matching degree and corresponding click budget that meets the standard as the best delivery advertisement for the designated user of the ad slot, thereby determining the best delivery advertisement for the designated user of each ad slot of the current target application software.

10. The data-driven intelligent advertising delivery platform according to claim 1, characterized in that: The evaluation of the effectiveness of advertising delivery to designated users of the target application software on that day includes: analyzing the response interactivity of designated users to the advertisements delivered in the target application software on that day based on the interaction data of designated users to the advertisements delivered in the target application software on that day, including the conversion volume, sharing volume, like volume and comment volume, performing a ratio analysis with the user benchmark response interactivity threshold stored in the cloud database, and using the ratio analysis result as an evaluation indicator for the effectiveness of advertising delivery to designated users of the target application software on that day.

Images