A marketing advertising method based on media big data

Abstract

The present invention discloses a marketing advertising delivery method based on media big data, which specifically relates to the technical field of advertising delivery, including constructing an advertising mapping database according to advertising labels, user categories and advertising effects, screening out advertisements that are highly relevant to target user groups through similarity calculation, determining promoted users based on advertising effects, collecting the push success rate, bounce rate and conversion rate of promoted users, determining the information of interest of promoted users, combining the purchase history records of promoted users by constructing a knowledge graph of labels, and collecting the time difference between adjacent purchase histories of promoted users, analyzing the consumption history behavior of promoted users, determining the consumption information of promoted users, comprehensively analyzing the information of interest and consumption information of promoted users, quantitatively evaluating the potential value of users, and determining high-quality promoted users. The present invention helps to accurately match advertisements with users, optimize advertising delivery effects, and improve advertising conversion rates and delivery efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of advertising delivery, and more specifically, to a marketing advertising delivery method based on media big data. Background Art

[0002] With the rapid development of digital advertising, the accuracy and effectiveness of advertising delivery are receiving more and more attention. Traditional advertising delivery methods are usually based on simple matching of user behavior data and advertising content. The existing advertising delivery system lacks effective user portrait construction and intelligent matching mechanism of advertising tags, resulting in low relevance of advertising push, poor user experience, and low advertising conversion rate. At the same time, users' consumption history and behavior are often analyzed in isolation, failing to effectively combine users' interests, preferences and consumption behavior for comprehensive analysis.

[0003] In order to solve the above-mentioned defects, a technical solution is now provided. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a marketing advertisement delivery method based on media big data to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A marketing advertising method based on media big data specifically includes the following steps:

[0007] S1: Build an ad mapping database based on ad tags, user categories, and ad effects. Filter out ads that are highly relevant to the target user group through similarity calculations. Then, determine the users to promote based on ad effects.

[0008] S2: Collect the push success rate, bounce rate, and conversion rate of promoted users to determine the information of interest to the promoted users;

[0009] S3: By building a knowledge graph of tags and combining it with the promoted user's purchase history, and collecting the time difference between the promoted user's adjacent purchase history, the promoted user's consumption history behavior is analyzed to determine the promoted user's consumption information;

[0010] S4: Comprehensively analyze the promotion users’ interest information and consumption information, quantitatively evaluate the users’ potential value, and identify high-quality promotion users.

[0011] In a preferred embodiment, screening out advertisements that are highly relevant to the target user group includes:

[0012] The labels of the advertisements that need to be placed and the labels of the advertisements in the advertisement mapping database are represented by vectors, and the labels of the advertisements that need to be placed are marked as: , where m=1, 2, 3, ..., M, where M is a positive integer and m is the label number of the advertisement to be placed. The labels of advertisements in the advertisement mapping database are marked as: , i=1, 2, 3, ..., I, I is a positive integer, i is the number of the mapping relationship in the advertisement mapping database, j=1, 2, 3, ..., J, J is a positive integer, j is the number of the advertisement label in the i-th mapping relationship;

[0013] The similarity coefficient measures the similarity between tags by Jaccard similarity. The calculation formula of the similarity coefficient is: ;in, is the similarity coefficient;

[0014] Set a similarity coefficient threshold to filter out advertisements that are highly relevant to the target user group, compare the similarity coefficient of the advertisement to be placed with the similarity coefficient threshold, obtain the mapping relationship in the advertisement mapping database that is greater than the similarity coefficient threshold, and regard the mapping relationship that is greater than the similarity coefficient threshold as the advertisement that is highly relevant to the target user group.

[0015] In a preferred embodiment, determining the promoted user includes:

[0016] Set a click volume threshold to further screen the filtered ads. By comparing the click volume threshold with the click rate of ads that are highly relevant to the target user group, obtain ads that are highly relevant to the target user group and have a click volume greater than the click volume threshold, determine the user categories that have a click volume greater than the click volume threshold, and mark the user categories that have a click volume greater than the click volume threshold as promoted users.

[0017] In a preferred embodiment, determining the information of interest to the promoted user includes:

[0018] The information of interest to the promoted user is represented by the interest deviation coefficient;

[0019] The logic for obtaining the interest deviation coefficient is as follows: obtaining the number of times the promoted user receives different promotional advertisements in the historical records, determining the promotion user's push success rate, and marking the promotion user's push success rate as: CG, where , The number of times the promotional ad is clicked by the promotional user. The number of times the promotion user received different promotion ads in the historical records;

[0020] Obtain the bounce rate of promoted users in the historical records, and mark the bounce rate of promoted users in the historical records as: TC; obtain the conversion rate of promoted users in the historical records, and mark the conversion rate of promoted users in the historical records as: ZH;

[0021] Set the push success rate standard value , bounce rate standard value And the conversion rate standard value , obtain the deviations of push success rate, bounce rate, and conversion rate, and mark the deviations of push success rate, bounce rate, and conversion rate as: 、 as well as ,in, , , ;

[0022] Calculate the interest deviation coefficient, the calculation formula is: ;in, is the interest bias coefficient.

[0023] In a preferred embodiment, determining consumption information of promoted users includes:

[0024] The consumption information of the promoted users is expressed through the consumption correlation coefficient;

[0025] The logic for obtaining the consumption correlation coefficient is as follows: construct a knowledge graph of labels, where entities in the knowledge graph represent product labels, and edges represent the correlation between product labels. Edges between adjacent entities have different weights, and the weights are determined by determining the number of co-occurrences of each pair of labels in all products. The co-occurrences are then normalized using the Min-Max normalization method, and the expression is: ;in, is the weight between adjacent entities a and b, is the number of co-occurrences of adjacent entities a and b, is the minimum value of the co-occurrence counts of all label pairs, is the maximum value of the co-occurrence counts of all tag pairs;

[0026] Based on the promoted user's purchase history, determine the length of time from when the promoted user purchased the product to the current time, and mark the length of time from when the promoted user purchased the product to the current time as: , where k=1, 2, 3, ..., K, K is a positive integer, and k is the type of purchased goods;

[0027] Based on the label of the promoted ad, determine the entity of the promoted ad and mark the entity of the promoted ad as: , r=1, 2, 3, ..., R, R is a positive integer, r is the number of the promotion advertisement entity, and according to the purchase history of the promotion user, the entity of the promotion user who purchased the product is determined and the entity of the promotion user who purchased the product is marked as: , p=1, 2, 3, ..., P, P is a positive integer, p is the ID of the entity that the promoted user purchased the product, and the consumption correlation coefficient is calculated using the following formula: ,in, is the consumption correlation coefficient.

[0028] In a preferred embodiment, the consumption information of the promoted user is represented by a consumption performance coefficient, including:

[0029] The logic for obtaining the consumption performance coefficient is as follows: based on the user's purchase history, determine the time difference between adjacent purchase histories of the promoted user, and mark the time difference between adjacent purchase histories of the promoted user as: , where w=1, 2, 3, ..., W, W is a positive integer, and w is the number of the adjacent purchase history;

[0030] Set the time difference threshold and mark the time difference threshold as: , compare the time difference between the adjacent purchase histories of promoted users with the time difference threshold, obtain the time difference between the adjacent purchase histories of promoted users that is greater than the time threshold, and re-mark the time difference between the adjacent purchase histories of promoted users that is greater than the time threshold as: , where y=1, 2, 3, ..., Y, Y is a positive integer, and y is the number of the time difference between the adjacent purchase histories of the promoted user that is greater than the time threshold;

[0031] Calculate the consumption performance coefficient, the calculation formula is: ;in, is the consumption performance coefficient.

[0032] In a preferred embodiment, determining high-quality promotion users includes:

[0033] Through comprehensive analysis of interest information and consumption information, the interest deviation coefficient, consumption correlation coefficient and consumption performance coefficient are weighted and calculated to construct a promotion evaluation model and generate a promotion evaluation coefficient. The calculation formula is: ;in, is the promotion evaluation coefficient, 、 、 are the proportional coefficients of interest deviation coefficient, consumption correlation coefficient, and consumption performance coefficient, respectively. 、 、 are both greater than 0;

[0034] Set a promotion evaluation coefficient threshold, and compare the promotion evaluation coefficient of each promotion user with the promotion evaluation coefficient threshold. If the promotion evaluation coefficient of the promotion user is greater than the promotion evaluation coefficient threshold, the promotion user will be marked as a high-quality promotion user. If the promotion evaluation coefficient of the promotion user is less than the promotion evaluation coefficient threshold, the promotion user will not be marked.

[0035] Technical effects and advantages of the present invention:

[0036] The present invention proposes an intelligent advertising promotion method based on advertising tags, user behavior analysis and tag knowledge graph. By constructing an advertising mapping database, advertisements that are highly relevant to the target user group are screened out, and the user's push success rate, bounce rate, conversion rate and other data are combined to analyze their interest information. At the same time, the knowledge graph is used to combine the user's purchase history and time difference to deeply explore the user's consumption behavior, comprehensively evaluate their potential value, and finally identify high-quality promotion users. The present invention helps to accurately match advertisements with users, optimize the advertising delivery effect, and improve advertising conversion rate and delivery efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0038] Figure 1 The figure is a flow chart of a marketing advertising delivery method based on media big data according to the present invention. DETAILED DESCRIPTION

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0040] Example 1

[0041] Figure 1 A flowchart of a marketing advertising method based on media big data is provided, which specifically includes the following steps:

[0042] S1: Build an ad mapping database based on ad tags, user categories, and ad effects. Filter out ads that are highly relevant to the target user group through similarity calculations. Then, determine the users to promote based on ad effects.

[0043] S2: Collect the push success rate, bounce rate, and conversion rate of promoted users to determine the information of interest to the promoted users;

[0044] S3: By building a knowledge graph of tags and combining it with the promoted user's purchase history, and collecting the time difference between the promoted user's adjacent purchase history, the promoted user's consumption history behavior is analyzed to determine the promoted user's consumption information;

[0045] S4: Comprehensively analyze the promotion users’ interest information and consumption information, quantitatively evaluate the users’ potential value, and identify high-quality promotion users.

[0046] Accurate ad delivery is achieved by determining ad delivery tags. Ad tags are automatically generated using natural language processing (NLP) and machine learning (ML) technologies, including:

[0047] Through a series of preprocessing steps, it is converted into a format that can be used for machine learning and natural language processing;

[0048] By calculating the frequency of words in the ad copy and their prevalence in all ad copies, it helps identify the most discriminative keywords and extract meaningful features from the ad copy;

[0049] By building a classification model, advertising copy is mapped to predefined advertising label categories. Classification models include machine learning models such as support vector machines (SVM), random forests, or neural networks, which can be used for advertising label classification. By learning the relationship between advertising copy and labels, the model can eventually automatically generate labels for new advertising copy.

[0050] Identifying different types of users is a key step in targeted advertising, personalized recommendations, and market segmentation. Based on various user behavior records on the platform, such as browsing history, search history, purchase history, etc., users are divided through clustering algorithms to obtain users in different clusters.

[0051] Based on the ad tags, the ads are initially delivered. Based on the number of clicks on the ads by different categories of users, the user categories with the best ad effects are determined, and an ad mapping database is constructed. The specific steps for constructing the ad mapping database include:

[0052] The ad mapping database contains multiple tables that record the delivery effects of different ad tags and user categories, including ad tag table, user category table and ad effect table;

[0053] After each ad run, collect the ad performance data and enter it into the ad performance table. Associate the ad tag, user category, and ad performance data. For example: Ad tag: Electronics + Limited-time offer → User category: Potential buyer → Click-through rate = 5%;

[0054] As advertising continues, new combinations of advertising tags and user categories will be continuously added to the database.

[0055] Perform similarity analysis on the labels of ads that need to be placed and the labels of ads in the ad mapping database, calculate the similarity coefficient, and set a similarity coefficient threshold to determine whether the labels of ads that need to be placed and the labels of ads in the ad mapping database are similar enough to ensure that the labels of ads that need to be placed match user behavior well.

[0056] Among them, the labels of advertisements that need to be placed and the labels of advertisements in the advertisement mapping database are represented by vectors, and the labels of advertisements that need to be placed are marked as: , where m=1, 2, 3, ..., M, where M is a positive integer and m is the label number of the advertisement to be placed. The labels of advertisements in the advertisement mapping database are marked as: , i=1, 2, 3, ..., I, I is a positive integer, i is the number of the mapping relationship in the advertisement mapping database, j=1, 2, 3, ..., J, J is a positive integer, j is the number of the advertisement label in the i-th mapping relationship;

[0057] It should be noted that the mapping relationship represents the user category and advertising effect corresponding to the advertising label.

[0058] The similarity coefficient measures the similarity between tags by Jaccard similarity. The calculation formula of the similarity coefficient is: ;in, is the similarity coefficient;

[0059] Set a similarity coefficient threshold to filter out advertisements that are highly relevant to the target user group. Compare the similarity coefficient of the advertisement to be placed with the similarity coefficient threshold to obtain mapping relationships in the advertisement mapping database that are greater than the similarity coefficient threshold. Count the mapping relationships that are greater than the similarity coefficient threshold as advertisements that are highly relevant to the target user group.

[0060] Set a click volume threshold to further screen the filtered ads. By comparing the click volume threshold with the click rate of ads that are highly relevant to the target user group, obtain ads that are highly relevant to the target user group and have a click volume greater than the click volume threshold, determine the user categories that have a click volume greater than the click volume threshold, and mark the user categories that have a click volume greater than the click volume threshold as promoted users.

[0061] The advantages of obtaining promotional users through screening include:

[0062] By setting similarity coefficients and click-through thresholds, you can filter out irrelevant or poorly performing ads, leaving only those that are most effective for the target user group. This can significantly increase click-through rates and conversion rates, and reduce advertising budget waste.

[0063] As advertising data accumulates, the system can continuously optimize advertising strategies based on click volume and similarity analysis. For example, if an ad has an unusually high click-through rate among a certain user group, the system can further promote it to a wider user group.

[0064] Through this data-based ad screening and optimization method, advertising platforms can more accurately understand the relationship between different ads and user groups, thereby providing stronger data support for advertising placement decisions.

[0065] By collecting the interest information and consumption information of the promoted users, the interest information of the promoted users is represented by the interest deviation coefficient, and the consumption information of the promoted users is represented by the consumption correlation coefficient and the consumption performance coefficient.

[0066] The logic for obtaining the interest deviation coefficient is as follows: obtaining the number of times the promoted user receives different promotional advertisements in the historical records, determining the promotion user's push success rate, and marking the promotion user's push success rate as: CG, where , The number of times the promotional ad is clicked by the promotional user. The number of times the promotion user received different promotion ads in the historical records;

[0067] Obtain the bounce rate of promoted users in the historical records, and mark the bounce rate of promoted users in the historical records as: TC; obtain the conversion rate of promoted users in the historical records, and mark the conversion rate of promoted users in the historical records as: ZH;

[0068] It should be noted that the bounce rate refers to the proportion of promoted users who leave the website after viewing only one page after clicking on the advertisement, and the conversion rate refers to the proportion of promoted users who complete the expected behavior of the advertisement. The higher the bounce rate of promoted users, the more carefully the promoted users will browse the advertisement content and usually have greater interest and patience in the advertisement. The higher the conversion rate of promoted users, the more likely the promoted users will make purchases after browsing the advertisement.

[0069] Set the push success rate standard value , bounce rate standard value And the conversion rate standard value , obtain the deviations of push success rate, bounce rate, and conversion rate, and mark the deviations of push success rate, bounce rate, and conversion rate as: 、 as well as ,in, , , ;

[0070] Calculate the interest deviation coefficient, the calculation formula is: ;in, is the interest bias coefficient.

[0071] From the formula, we can see that the larger the interest deviation coefficient is, the smaller the push success rate, bounce rate and conversion rate of the promoted user are, which means that the promoted user may not be interested in the push of advertisements.

[0072] The advantages of the consumption correlation coefficient include:

[0073] By building relationships between tags, the knowledge graph can help identify the deep connections between user purchasing behavior and ad tags. This allows ads to be accurately matched to content most relevant to user needs and interests, thereby increasing ad click-through rates and conversion rates.

[0074] By establishing associations between user interest tags and product tags, the knowledge graph can effectively understand users' changing interests and purchasing potential. Using these tags, the advertising system can quickly adapt to dynamically changing user needs and behaviors, enabling personalized ad recommendations.

[0075] Knowledge graphs can support in-depth analysis of the relationships between different tags. For example, by analyzing the co-occurrence patterns of a tag (such as "sports shoes") and other related tags (such as "sports accessories"), advertisers can determine which tag combinations are most attractive to specific user groups.

[0076] Knowledge graphs can mine cross-category relationships between product labels, which makes cross-category advertising possible.

[0077] The logic for obtaining the consumption correlation coefficient is as follows: construct a knowledge graph of labels, where entities in the knowledge graph represent product labels, and edges represent the correlation between product labels. Edges between adjacent entities have different weights, and the weights are determined by determining the number of co-occurrences of each pair of labels in all products. The co-occurrences are then normalized using the Min-Max normalization method, and the expression is: ;in, is the weight between adjacent entities a and b, is the number of co-occurrences of adjacent entities a and b, is the minimum value of the co-occurrence counts of all label pairs, is the maximum value of the co-occurrence counts of all tag pairs;

[0078] It should be noted that the number of co-occurrences refers to the number of times the two tags appear in the same product at the same time.

[0079] Based on the promoted user's purchase history, determine the length of time from when the promoted user purchased the product to the current time, and mark the length of time from when the promoted user purchased the product to the current time as: , where k=1, 2, 3, ..., K, K is a positive integer, and k is the type of purchased goods;

[0080] Based on the label of the promoted ad, determine the entity of the promoted ad and mark the entity of the promoted ad as: , r=1, 2, 3, ..., R, R is a positive integer, r is the number of the promotion advertisement entity, and according to the purchase history of the promotion user, the entity of the promotion user who purchased the product is determined and the entity of the promotion user who purchased the product is marked as: , p=1, 2, 3, ..., P, P is a positive integer, p is the ID of the entity that the promoted user purchased the product, and the consumption correlation coefficient is calculated using the following formula: ,in, is the consumption correlation coefficient.

[0081] As can be seen from the formula, the larger the consumption correlation coefficient, the more likely the promoted user is to pay attention to the advertisement in the future, and the higher the degree of correlation between the promoted advertisement and the user, the more likely the user is to purchase goods based on the promoted advertisement.

[0082] The logic for obtaining the consumption performance coefficient is as follows: based on the user's purchase history, determine the time difference between adjacent purchase histories of the promoted user, and mark the time difference between adjacent purchase histories of the promoted user as: , where w=1, 2, 3, ..., W, W is a positive integer, and w is the number of the adjacent purchase history;

[0083] Set the time difference threshold and mark the time difference threshold as: , compare the time difference between the adjacent purchase histories of promoted users with the time difference threshold, obtain the time difference between the adjacent purchase histories of promoted users that is greater than the time threshold, and re-mark the time difference between the adjacent purchase histories of promoted users that is greater than the time threshold as: , where y=1, 2, 3, ..., Y, Y is a positive integer, and y is the number of the time difference between the adjacent purchase history of the promoted user that is greater than the time threshold;

[0084] It should be noted that the time difference threshold is set by professional staff. The time difference threshold is a specific length of time used to measure the frequency of user purchases.

[0085] Calculate the consumption performance coefficient, the calculation formula is: ;in, is the consumption performance coefficient.

[0086] From the formula, we can see that the larger the consumption performance coefficient is, the longer the time interval between each consumption of the user may be, which means that the user's purchase frequency is low, which means that they may not be regular consumers of this type of product, or their consumption habits are not frequent. Therefore, the more users need to be promoted through advertising to activate their consumption needs.

[0087] Through comprehensive analysis of interest information and consumption information, the interest deviation coefficient, consumption correlation coefficient and consumption performance coefficient are weighted and calculated to construct a promotion evaluation model and generate a promotion evaluation coefficient. The calculation formula is: ;in, is the promotion evaluation coefficient, 、 、 are the proportional coefficients of interest deviation coefficient, consumption correlation coefficient, and consumption performance coefficient, respectively. 、 、 Both are greater than 0.

[0088] It can be seen from the formula that the larger the interest deviation coefficient and the smaller the consumption correlation coefficient and the consumption performance coefficient, the smaller the promotion evaluation coefficient, which means that the promotion user does not need promotional advertising for promotion. Conversely, the smaller the interest deviation coefficient and the larger the consumption correlation coefficient and the consumption performance coefficient, the larger the promotion evaluation coefficient, which means that the promotion user needs promotional advertising for promotion.

[0089] Set a promotion evaluation coefficient threshold, and compare the promotion evaluation coefficient of each promotion user with the promotion evaluation coefficient threshold. If the promotion evaluation coefficient of the promotion user is greater than the promotion evaluation coefficient threshold, the promotion user will be marked as a high-quality promotion user, indicating that the more the promotion user needs promotion advertising for promotion, the more likely the promotion user is to make a purchase. If the promotion evaluation coefficient of the promotion user is less than the promotion evaluation coefficient threshold, the promotion user will not be marked.

[0090] It should be noted that the promotion evaluation coefficient is used to quantitatively evaluate the potential value of users, so as to make more accurate advertising decisions. By screening out high-quality promotion users, advertisements can be personalized for high-quality promotion users, which helps to improve the effectiveness of advertising delivery, ensure the efficient use of advertising resources, and improve the overall marketing return rate.

[0091] The present invention proposes an intelligent advertising promotion method based on advertising tags, user behavior analysis and tag knowledge graph. By constructing an advertising mapping database, advertisements that are highly relevant to the target user group are screened out, and the user's push success rate, bounce rate, conversion rate and other data are combined to analyze their interest information. At the same time, the knowledge graph is used to combine the user's purchase history and time difference to deeply explore the user's consumption behavior, comprehensively evaluate their potential value, and finally identify high-quality promotion users. The present invention helps to accurately match advertisements with users, optimize the advertising delivery effect, and improve advertising conversion rate and delivery efficiency.

[0092] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0093] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0094] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0095] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0096] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0097] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0098] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0099] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A marketing advertising method based on media big data, characterized in that: The specific steps include: S1: Based on the advertisement tags, user categories, and advertisement effects, an advertisement mapping database is constructed. Advertisements that are highly relevant to the target user group are screened out by calculating the similarity between the tags of the advertisements to be placed and the tags of the advertisements in the advertisement mapping database. Based on the advertisement effects, promotion users are determined. A similarity coefficient threshold is set to screen out advertisements that are highly relevant to the target user group. The similarity coefficient of the advertisements to be placed is compared with the similarity coefficient threshold to obtain mapping relationships in the advertisement mapping database that are greater than the similarity coefficient threshold. Mapping relationships that are greater than the similarity coefficient threshold are considered as advertisements that are highly relevant to the target user group. S2: Collect the push success rate, bounce rate, and conversion rate of promoted users to determine the information of interest to the promoted users; S3: By building a knowledge graph of tags and combining it with the promoted user's purchase history, and collecting the time difference between the promoted user's adjacent purchase history, the promoted user's consumption history behavior is analyzed to determine the promoted user's consumption information; S4: Comprehensively analyze the user's interest information and consumption information, quantitatively evaluate the user's potential value, and identify high-quality promotion users; Among them, determining the promotion users includes: Set a click volume threshold to further screen the filtered ads. By comparing the click volume threshold with the click-through rate of ads that are highly relevant to the target user group, obtain ads that are highly relevant to the target user group and have a click volume greater than the threshold. Determine the user category with a click volume greater than the threshold, and mark the user category with a click volume greater than the threshold as promoted users. Determine the information that users are interested in, including: The information of interest to the promoted user is represented by the interest deviation coefficient; The logic for obtaining the interest deviation coefficient is as follows: obtaining the number of times the promoted user receives different promotional advertisements in the historical records, determining the promotion user's push success rate, and marking the promotion user's push success rate as: CG, where , The number of times the promotional ad is clicked by the promotional user. The number of times the promotion user received different promotion ads in the historical records; Obtain the bounce rate of promoted users in the historical records, and mark the bounce rate of promoted users in the historical records as: TC; obtain the conversion rate of promoted users in the historical records, and mark the conversion rate of promoted users in the historical records as: ZH; Set the push success rate standard value , bounce rate standard value And the conversion rate standard value , obtain the deviations of push success rate, bounce rate, and conversion rate, and mark the deviations of push success rate, bounce rate, and conversion rate as: 、 as well as ,in, , , ; Calculate the interest bias coefficient, the calculation formula is: ;in, is the interest bias coefficient; Determine the consumption information of the promoted user, including: The consumption information of the promoted users is expressed through the consumption correlation coefficient; The logic for obtaining the consumption correlation coefficient is as follows: construct a knowledge graph of labels, where entities in the knowledge graph represent product labels, and edges represent the correlation between product labels. Edges between adjacent entities have different weights, and the weights are determined by determining the number of co-occurrences of each pair of labels in all products. The co-occurrences are then normalized using the Min-Max normalization method, and the expression is: ;in, is the weight between adjacent entities a and b, is the number of co-occurrences of adjacent entities a and b, is the minimum value of the co-occurrence counts of all label pairs, is the maximum value of the co-occurrence counts of all tag pairs; Based on the promoted user's purchase history, determine the length of time from when the promoted user purchased the product to the current time, and mark the length of time from when the promoted user purchased the product to the current time as: , where k=1, 2, 3, ..., K, K is a positive integer, and k is the type of purchased goods; Based on the label of the promoted ad, determine the entity of the promoted ad and mark the entity of the promoted ad as: , r=1, 2, 3, ..., R, R is a positive integer, r is the number of the promotion advertisement entity, and according to the purchase history of the promotion user, the entity of the promotion user who purchased the product is determined and the entity of the promotion user who purchased the product is marked as: , p=1, 2, 3, ..., P, P is a positive integer, p is the ID of the entity that the promoted user purchased the product, and the consumption correlation coefficient is calculated using the following formula: ,in, is the consumption correlation coefficient; The consumption information of promoted users is expressed through consumption performance coefficients, including: The logic for obtaining the consumption performance coefficient is as follows: based on the user's purchase history, determine the time difference between adjacent purchase histories of the promoted user, and mark the time difference between adjacent purchase histories of the promoted user as: , where w=1,2,3,…,W, W is a positive integer, and w is the number of the adjacent purchase history; Set the time difference threshold and mark the time difference threshold as: , compare the time difference between the adjacent purchase histories of promoted users with the time difference threshold, obtain the time difference between the adjacent purchase histories of promoted users that is greater than the time threshold, and re-mark the time difference between the adjacent purchase histories of promoted users that is greater than the time threshold as: , where y=1, 2, 3, ..., Y, Y is a positive integer, and y is the number of the time difference between the adjacent purchase histories of the promoted user that is greater than the time threshold; Calculate the consumption performance coefficient, the calculation formula is: ;in, is the consumption performance coefficient; Identify high-quality promotion users, including: Through comprehensive analysis of interest information and consumption information, the interest deviation coefficient, consumption correlation coefficient and consumption performance coefficient are weighted and calculated to construct a promotion evaluation model and generate a promotion evaluation coefficient. The calculation formula is: ;in, is the promotion evaluation coefficient, 、 、 are the proportional coefficients of interest deviation coefficient, consumption correlation coefficient, and consumption performance coefficient, respectively. 、 、 are both greater than 0; Set a promotion evaluation coefficient threshold, and compare the promotion evaluation coefficient of each promotion user with the promotion evaluation coefficient threshold. If the promotion evaluation coefficient of the promotion user is greater than the promotion evaluation coefficient threshold, the promotion user will be marked as a high-quality promotion user. If the promotion evaluation coefficient of the promotion user is less than the promotion evaluation coefficient threshold, the promotion user will not be marked.

2. A marketing advertising method based on media big data according to claim 1, characterized in that: Filter out ads that are highly relevant to the target user group, including: The labels of the advertisements that need to be placed and the labels of the advertisements in the advertisement mapping database are represented by vectors, and the labels of the advertisements that need to be placed are marked as: , where m=1, 2, 3, ..., M, where M is a positive integer and m is the label number of the advertisement to be placed. The labels of advertisements in the advertisement mapping database are marked as: , i=1, 2, 3, ..., I, I is a positive integer, i is the number of the mapping relationship in the advertisement mapping database, j=1, 2, 3, ..., J, J is a positive integer, j is the number of the advertisement label in the i-th mapping relationship; The similarity coefficient measures the similarity between tags through Jaccard similarity. The calculation formula of the similarity coefficient is: ;in, is the similarity coefficient.

Images