CN118096266A

CN118096266A - Intelligent advertisement marketing system and method based on Internet

Info

Publication number: CN118096266A
Application number: CN202410464942.9A
Authority: CN
Inventors: 夏圣尧; 丁其磊
Original assignee: Shanghai Ocean University
Current assignee: Shanghai Ocean University
Priority date: 2024-04-18
Filing date: 2024-04-18
Publication date: 2024-05-28
Anticipated expiration: 2044-04-18
Also published as: CN118096266B

Abstract

The invention relates to the technical field of advertisement marketing, in particular to an intelligent advertisement marketing system and method based on the Internet, comprising a user analysis management unit, a data collection module, a user portrait construction module and a user portrait construction module, wherein the user analysis management unit is used for collecting user behavior data, the data analysis module is used for analyzing the user behavior data, and the user portrait construction module is used for classifying user groups according to user labels based on the result output by the data analysis module; the advertisement resource management unit is used for managing and distributing advertisement space resources for delivery; the targeted delivery unit matches advertisements with target audiences based on the user portraits, and adopts a targeted delivery algorithm to formulate an optimal advertisement delivery strategy; the putting optimization unit adjusts the bidding strategy based on the maximum expected benefits according to the net benefits of the advertisement position on each media platform resource, accurately reflects the liveness, loyalty and short-term buying habits of the user based on the buying frequency of the user, provides a basis for formulating a periodic marketing strategy, and further achieves accurate positioning of target audiences.

Description

Intelligent advertisement marketing system and method based on Internet

Technical Field

The invention relates to the technical field of advertisement marketing, in particular to an intelligent advertisement marketing system and method based on the Internet.

Background

The intelligent advertisement marketing system of the Internet is an advanced online advertisement service form, and utilizes advanced technical means such as big data, artificial Intelligence (AI), machine learning and the like to realize a complex system with functions such as accurate advertisement orientation, intelligent creative generation, automatic delivery and optimization and the like.

Intelligent advertising marketing is highly dependent on machine learning and artificial intelligence algorithms, which may suffer from black-box effects, i.e., difficult to interpret decision-making processes, and may sometimes erroneously push advertisements to inappropriate target groups, thus providing internet-based intelligent advertising marketing systems and methods.

Disclosure of Invention

The present invention is directed to an internet-based intelligent advertisement marketing system and method, which solve the problem that the intelligent advertisement marketing proposed in the above-mentioned background art is highly dependent on machine learning and artificial intelligence algorithms, and these algorithms may have black box effects, i.e. the decision process is difficult to explain, and may sometimes erroneously push advertisements to unsuitable target groups.

In order to achieve the above purpose, the present invention provides an internet-based intelligent advertisement marketing system, which comprises a user analysis management unit, wherein the user analysis management unit collects user behavior data through a data collection module, the data analysis module analyzes the user behavior data to identify the consumption frequency, content preference and product relevance characteristics of users, and a user portrait construction module generates user labels based on the output result of the data analysis module, classifies user groups according to the user labels to form different user group categories;

the advertisement resource management unit is used for managing and distributing advertisement position resources for delivery, and comprises an advertisement resource library, a media resource integration module and an advertisement resource management module;

The advertisement resource management module (23) distributes advertisements in the advertisement resource library (21) according to the flow characteristics and the values of different media channels by adopting an advertisement resource super-optimization algorithm according to the user portraits generated by the user analysis management unit (1);

the targeted delivery unit is used for matching advertisements with target audiences based on user portraits, combining advertisement targets and budgets and adopting a targeted delivery algorithm to formulate an optimal advertisement delivery strategy;

considering the time sensitivity of user behavior, the purchasing needs and interest preferences of users often have a certain timeliness, and recent purchasing behaviors may reflect the current interest hotspots or demand trends of users, so that pushing related advertisement content is easier to draw attention and response of users. Meanwhile, the possibility that advertisements can be seen and interacted by the advertisements can be improved according to the advertisements placed in the user active time intervals, and the optimized directed placement algorithm is introduced by the factors of the latest active time intervals, and is the following:

；

In the method, in the process of the invention, Representing a user representation; /(I)A feature description representing an advertisement; /(I)Representing an optimized representation of a userAnd characterization of advertisements/>A match score between; /(I)Representing personalized attributes in the user representation; /(I)Representing target crowd attributes in an advertisement positioning strategy; /(I)A set of interest tags representing a user; /(I)Representing a set of advertisement-related interest keywords; Historical consumption behavior habit data representing a user; /(I) Representing a product category with which the advertisement is associated; /(I)Representing a similarity function between two attributes; /(I)、/>And/>Respectively representing weight factors of different matching dimensions; /(I)Time sensitivity to user behavior; /(I)Time information representing delivery of the advertisement; /(I)The time sensitivity of the user behavior is represented by the influence weight of the user behavior on the matching result, and the importance of relevant factors such as the latest purchasing behavior, the activity period and the like in the targeted delivery is represented.

When (when)When the value of (2) is larger, the influence of time sensitivity on the matching result is larger, and the model more pays attention to the latest behavior or activity period of the user; when/>When the value of (2) is smaller, the influence of the time factor on the matching result is smaller, and the model is more dependent on the matching results of other dimensions; the advertisement delivery strategy can be adjusted by an advertiser according to the user flow and the liveness of different time periods through analysis of the user liveness time periods, so that the optimal utilization of advertisement budget is realized;

And the delivery optimizing unit is used for adjusting bidding strategies of advertisements on each media platform resource based on a delivery optimizing algorithm according to the net benefits of the advertisement positions on each media platform resource, tracking the behavior paths of users on different platforms by adopting a cross-platform joint optimizing algorithm, and optimizing the advertisement delivery strategies to realize the overall optimization of the cross-platform advertisement effect.

As a further improvement of the technical scheme, the data analysis module is based on a user behavior analysis algorithm, and the specific expression involved in the analysis of the user behavior data is as follows:

frequency of user consumption:

If the user set is U, the user Wherein, time window/>The specific expression of the consumption frequency of the user is: /(I)；

In the method, in the process of the invention,Representing user/>Frequency of purchases over the past 7 days; /(I)Representing user/>At the time point/>The number of purchases made; /(I)Represents a time window representing the time from the/>Beginning on day, a period of 7 days follows in succession; /(I)Representing a time point/>Belonging to time window/>；

User content preferences:

If there is a group of contents to be interacted with Wherein the content of the user interaction/>The specific expression of the user content preference is: /(I)；

In the method, in the process of the invention,Content representing user interactions; /(I)To indicate the function, the representation is at the time point/>User/>Whether or not to match content/>There is interaction, if there is better, then/>Otherwise/>；/>Representing user/>For content/>Is a preference degree of (a);

Product relevance:

if there are two products E and F, their support And confidence/>The relevance of products E and F is calculated as: /(I)；

In the method, in the process of the invention,Indicating a multiple of increase in probability of purchasing the product F compared to the base probability of purchasing the product F given the purchase of the product E.

Degree of elevation ofFor helping an analyst to understand the degree of association between two products:

if the degree of lift is greater than 1, indicating that there is a positive correlation between product E and product F, a customer purchasing product E is more likely to purchase product F;

if the degree of lift is equal to 1, it indicates that there is no association between product E and product F;

If the degree of lift is less than 1, it indicates that there is a negative correlation between product E and product F, and that the customer purchasing product E is unlikely to purchase product F instead.

As a further improvement of the technical scheme, the specific steps involved in the generation of the user label by the user portrait construction module are as follows;

s3.1, acquiring output user behavior data from a data analysis module, and analyzing to obtain consumption frequency A, content preference B and product relevance characteristics C;

S3.2, defining a group of user labels according to service requirements and analysis results, and mapping the consumption frequency A, the content preference B and the product relevance feature C to the defined user labels respectively;

the specific mapping rule is as follows:

Consumption frequency a mapping rule:

If the user purchases frequency in the past 7 days Higher, it can be mapped to a "high frequency consumer" tag/>；

If the purchase frequency is moderate, the label can be mapped into a conventional consumer label；

If the purchase frequency is low, it may be mapped to a "low frequency consumer" tag；

Content preference B mapping rules:

According to the content of the user Preference degree/>If a user frequently interacts with a certain class of content, such content preferences are mapped to corresponding interest tags/>In which,/>Representing an index variable representing different interest types, e.g./>Representing a scientific interest/>Representing fashion trends;

Product relevance feature C mapping rules:

if the user purchases product E and the degree of promotion between products E and F Higher, such users can be mapped to "associated product purchaser" tags/>；

If a user tends to purchase a package or a combination of complementary products, often immediately after purchase of product A, product B, the user of this behavior pattern may be mapped to a "package purchase preference" tag；

By combining the mapping rules, the user portrait construction module can accurately label the users based on the performances of the users in different dimensions, and is beneficial to the fine operation of the advertisement system aiming at different user groups.

S3.3, classifying the user groups by adopting user label classification based on defined user labels, and dividing the users into groups with similar characteristics; dividing into groups with similar characteristics, realizing accurate positioning of target audience, being beneficial to more effectively distributing advertisement budget, reducing ineffective exposure and waste, and by identifying common interests, behavior patterns or consumption habit characteristics of users, the system can pertinently push products, contents or services which are possibly interested in the users to various user groups, increase click rate and conversion rate, and simultaneously provide a powerful tool for enterprises to design popularization strategies of cross-selling, binding selling or other related product combinations;

Among the above steps, the specific steps involved in the classification of the user tag are:

s3.31, randomly selecting K users as initial clustering centers, and recording as ；

S3.32, for each userThe distance between the clustering center and K clustering centers is calculated, and the distance expression is as follows:

；

In the method, in the process of the invention, The label is a feature dimension; /(I)Representing user/>In/>Values on the individual features; /(I)Representing cluster centersIn/>Values on the individual features;

will user Assigned to the cluster corresponding to the nearest cluster center:

Order of the game ；

In the method, in the process of the invention,Representing user/>Index of clusters to which it should be assigned, i.e. user/>Finally belonging to the cluster category number; /(I)Representation of/>, for a given userSearching for a cluster K such that the distance of the user from the center of the cluster is minimal; k represents the total number of clusters; /(I)Represents the/>Individual users or data points; /(I)A cluster center representing the kth cluster; for user/>Assigning a cluster tag/>To make it correspond to the nearest cluster center/>, to the userDuring each iteration, each user is reassigned to the nearest cluster according to the principle until the cluster center is no longer significantly changed;

S3.33, calculating the average value of all users in each cluster, and taking the average value as a new cluster center;

The update formula of the new cluster center is as follows:

；

In the method, in the process of the invention, Representing all user sets in the currently allocated cluster K; /(I)Representing a list of all users in the current cluster KA new cluster center obtained by averaging the feature vectors of the images; /(I)Representing the number of users in cluster K;

the clustering center is made to be as close to the members in the cluster to which the clustering center belongs as possible, so that the clustering effect is optimized, and in each iteration process, the algorithm can converge to a local optimal solution by continuously updating the clustering center and redistributing sample points;

S3.34, performing convergence judgment:

stopping iteration if the clustering center reaches a preset threshold value or the clustering result is unchanged after a plurality of continuous iterations;

Otherwise, returning to the step S3.32 to continuously reassign the users and update the clustering center;

s3.34, after iteration is finished, a final clustering result is obtained, namely, clusters to which each user belongs are clustered, and user portraits can be constructed based on the clusters;

The user behavior data within each cluster is made as close as possible by constantly adjusting the position of the cluster center, thereby enabling an efficient subdivision of the user population.

S3.4, integrating the user tag and the group information to generate a user image, wherein the user image comprises basic information, interest and hobbies and purchasing behavior characteristics of a user, so as to form a comprehensive user description; each user representation represents a comprehensive and personalized user description for application scenarios such as personalized recommendation, precision marketing, customer relationship management, etc.

As a further improvement of the technical scheme, the advertisement resource library is used for storing all advertisement resources for delivery;

The media resource integration module is in butt joint with various media platforms through an API interface and is used for managing media platform resources for advertising, so that cross-platform advertising resource unified management and scheduling are realized;

And the advertisement resource management module distributes advertisements in the advertisement resource library according to the flow characteristics and the values of different media channels by adopting an advertisement resource super-optimization algorithm according to the user portrait generated by the user analysis management unit.

As a further improvement of the technical scheme, the advertisement resource super-optimization algorithm specifically comprises the following steps:

；

In the method, in the process of the invention, Represents the/>The advertisement is at the/>Exposure on the individual media platform resources; /(I)Represents the/>The advertisement is at the/>Click rate on individual media platform resources; /(I)Represents the/>The advertisement is at the/>Conversion rate on individual media platform resources; /(I)Represents the/>The advertisement is at the/>Average revenue per user brought on the individual media platform resources; /(I)Represents the/>The advertisement is at the/>Display cost per thousand times on the individual media platform resources; /(I)Representing the number of ad spots; /(I)Representing the number of media platform resources; /(I)Representing the expected profit sum generated by all the advertisement spots on each media platform resource minus the corresponding putting cost sum, and obtaining the net profit;

wherein, the constraint condition is:

Total budget limit constraint: ；

In the method, in the process of the invention, Representing the total budget;

Non-negative constraint:

；

In the method, in the process of the invention, Representation of all/>And all/>The inequality must be satisfied;

Period launch quota constraints:

；

In the method, in the process of the invention, Represents the/>All advertisement sets contained within a time period; /(I)Representing period/>An upper budget limit within; expressed in/> The sum of the exposure of all advertisements on each media platform resource in each time period; /(I)Representation of all/>The inequality described above must be true.

As a further improvement of the technical scheme, the directional delivery unit comprises a matching delivery module, the matching delivery module is used for carrying out accurate matching according to the user portrait and the advertisement positioning strategy based on a directional delivery algorithm, and pushing the most suitable advertisement content to the corresponding user, and the specific expression of the directional delivery algorithm is as follows:

；

Determining the most suitable advertisement content according to the characteristics of the user and the positioning strategy of the advertisement; the method comprises the steps of comprehensively considering a plurality of factors such as personalized attributes, interest labels, historical consumption behaviors and the like of users, determining the most suitable advertisement content by calculating the similarity among the factors and multiplying the factors by corresponding weight factors, wherein the weight factors can be adjusted according to specific conditions so as to adjust the matching priority according to different conditions.

As a further improvement of the technical scheme, the delivery optimizing unit comprises a data monitoring module, a target setting module and a bidding strategy optimizing module;

The data monitoring module is used for collecting and processing data related to advertisement delivery, the target setting module determines an optimization target based on a cross-platform joint optimization algorithm, the bidding strategy optimization module adjusts bidding strategies of advertisements on each media platform resource based on the optimization target through a delivery optimization algorithm according to expected benefits and target setting, and the delivery optimization algorithm dynamically adjusts the bidding strategies based on expected benefits and real-time data so as to achieve optimal delivery effects.

As a further improvement of the technical scheme, the specific expression of the delivery optimization algorithm is as follows:

；

In the method, in the process of the invention, Representing global bidding coefficients for uniformly adjusting cost proportions of all resources; /(I)Expressed in/>The amount of delivery on the individual media resources, namely the showing times and clicking times of the advertisements on the resources; /(I)Expressed in/>Average revenue on individual media assets; /(I)Representing the number of media assets; /(I)For index variables, representing the media assets used to traverse each of 1 through M; /(I)Expressed in/>Average cost on individual media assets;

wherein the constraint meets the total cost not exceeding the budget ：

；

In the method, in the process of the invention,Representing the maximum total cost allowed;

by adjusting global bidding coefficients And the delivery volume per media asset/>To maximize the overall net benefit;

based on the optimization objective of maximizing the total revenue, then:

；

In the method, in the process of the invention, Representing the maximized total benefit as an optimization target;

introducing constraint conditions into the objective function to form the objective function for dynamically adjusting global bid coefficients And the delivery volume per media asset/>：

；

In the method, in the process of the invention,Representing a lagrangian function;

by respectively corresponding to And/>Taking the partial derivative and setting to 0 to find the extreme point:

；

Solving the above equation set, for the first Delivered-on amount on individual media resources/>：

；

For global bidding coefficients：

；

In the method, in the process of the invention,Representing lagrangian multipliers for handling constraints of the problem; /(I)Represents the Lagrangian function L versus the delivered-quantity/>Is a partial derivative of (2); /(I)Representing the Lagrangian function L versus the global bid coefficient/>Is a partial derivative of (c).

As a further improvement of the technical scheme, the specific expression related to the cross-platform joint optimization algorithm is as follows:

media assets Exposure to media assets/>The click effect is as follows:

；

media assets Exposure to media assets/>The click effect is as follows:

；

In the method, in the process of the invention, Representing advertisements in media assets/>Exposure to media assets/>The degree of impact of the click; /(I)Expressed in media resource/>Click rate of (2); /(I)Expressed in media resource/>Click rate of (2); /(I)Expressed in media resource/>Post exposure at media asset/>The probability of clicking is improved; /(I)Expressed in media resource/>Post exposure at media asset/>The probability of clicking is improved; /(I)Expressed in media resource/>Exposure times of (2); /(I)Expressed in media resource/>Exposure times of (2); /(I)An intensity coefficient representing an influence; /(I)An intensity coefficient representing an influence;

Then the objective function K is jointly optimized:

；

In the method, in the process of the invention, Representing media assets/>At the dose/>And bidding coefficients/>The lower independent advertising effect benefits; Representing the respective delivery amount/>, of each media resource And bidding coefficients/>Direct advertising effect revenue sum generated below,/>And/>All are index variables; /(I)Representing media assets/>Media resource/>Weight coefficient between; /(I)Representing media assets/>Media resource/>Weight coefficient between; /(I)And the function value is used for maximizing the function value to realize the optimal advertising effect of the cross-media resource platform for the overall optimization target.

By considering the synergistic effect among the platforms, the click probability of the user on other platforms can be improved by the exposure among the platforms, the synergistic effect helps to promote the overall advertising effect, and the gains of the independent release of each platform and the mutual influence among the platforms are comprehensively considered so as to realize the overall optimization of the cross-platform advertising effect.

On the other hand, the invention provides an intelligent advertisement marketing method based on the Internet, which is used for the intelligent advertisement marketing system based on the Internet and comprises the following steps:

S10.1, collecting user behavior data by a data collection module, calculating consumption frequency, content preference and content interaction condition of a user by a data analysis module through a user behavior analysis algorithm, generating user labels comprising the consumption frequency, the content preference and the product relevance according to the behavior characteristics of the user by a user portrait construction module, classifying the user into groups with similar characteristics, and forming user portraits;

S10.2, butting a plurality of media platform resources through a media resource integration module, uniformly managing and scheduling cross-platform advertisement space resources, and dynamically adjusting allocation strategies of advertisements in different media channels by using an advertisement resource super-optimization algorithm through an advertisement resource management module under the consideration of budget constraint and time slot quota throwing factors to maximize expected benefits;

S10.3, calculating the matching degree between the user portrait and the advertisement feature by adopting a matching delivery module and adopting a directional delivery algorithm, combining personalized attributes, interest labels and historical consumption behaviors to carry out accurate matching, formulating an optimal advertisement delivery strategy, pushing the most suitable advertisement content to the corresponding user, and simultaneously considering time sensitivity factors such as the latest activity period and purchasing behavior so as to improve the correlation and click conversion rate of the advertisement;

And S10.4, finally, monitoring advertisement delivery effect in real time through a delivery optimization unit, collecting and processing advertisement clicking and conversion related data, setting the maximized total income as an optimization target through a target setting module, dynamically adjusting global bidding coefficients and delivery amount on each media resource by adopting a delivery optimization algorithm, and realizing the maximization of advertisement benefit on the premise of meeting budget limit.

Compared with the prior art, the invention has the beneficial effects that:

1. In the intelligent advertisement marketing system and method based on the Internet, a data analysis module constructs accurate user portraits based on a user behavior analysis algorithm, identifies consumption frequency, content preference and product relevance characteristics of users, and provides powerful support for personalized advertisement delivery;

while based on the frequency of purchases by the user over the past 7 days The method accurately reflects the liveness, loyalty and short-term buying habits of the user, and provides basis for formulating a periodic marketing strategy; and users are clustered by adopting a user tag classification algorithm, and a clustering center is continuously optimized, so that the users in each cluster have highly similar behavior patterns and consumption characteristics, and further, the accurate positioning of target audience is realized.

2. In the intelligent advertisement marketing system and method based on the Internet, a directional delivery unit adopts a directional delivery algorithm to formulate an optimal advertisement delivery strategy, comprehensively considers personalized attributes, interest labels, historical consumption behaviors and time sensitivity factors of users, ensures that the most suitable advertisement content is pushed to a target user group at proper time, and thereby remarkably improves the advertisement click rate and conversion rate;

And the advertisement putting effect is monitored in real time through the putting optimization unit, and the preset optimization target is combined, so that the bidding strategy and the putting quantity are automatically adjusted by using the putting optimization algorithm, and the maximization of the putting effect is ensured to be realized within the limited budget.

Drawings

Fig. 1 is an overall flow diagram of the present invention.

The meaning of each reference sign in the figure is:

1. A user analysis management unit; 11. a data collection module; 12. a data analysis module; 13. a user portrait construction module;

2. An advertisement resource management unit; 21. an advertisement resource library; 22. a media resource integration module; 23. an advertising resource management module;

3. a directional delivery unit; 4. and a throwing optimizing unit.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Example 1

Referring to fig. 1, an internet-based intelligent advertising marketing system is provided, which includes a user analysis management unit 1, wherein the user analysis management unit 1 collects user behavior data through a data collection module 11;

The data collection module 11 obtains the cookies information and the equipment identifier of the user; the user behavior data comprise browsing history, search records, purchasing behavior and social media interaction data of a user on the APP;

The data analysis module 12 analyzes the user behavior data, identifies the consumption frequency, content preference and product relevance characteristics of the user, and the user portrait construction module 13 generates user labels based on the results output by the data analysis module 12, classifies the user groups according to the user labels to form different user group categories, so that personalized advertisement strategies and service schemes can be customized for each group;

Collecting user information in a mode of cookies, equipment identifiers and user behavior data, analyzing user attributes, interest preferences and consumption habits based on the user information to form a user tag system, and completing user portrait construction, wherein the user analysis management unit 1 comprises a data collection module 11, a data analysis module 12 and a user portrait construction module 13;

The data analysis module 12 analyzes the user behavior data based on a user behavior analysis algorithm, and the specific expression involved in the analysis is:

frequency of user consumption:

If the user set is U, the user Wherein, time window/>The specific expression of the consumption frequency of the user is:

；

In the method, in the process of the invention, Representing user/>Frequency of purchases over the past 7 days; /(I)Representing user/>At the time point/>The number of purchases made; /(I)Represents a time window representing the time from the/>Beginning on day, a period of 7 days follows in succession; /(I)Representing a time point/>Belonging to time window/>；

By passing throughThe parameter indexes can know the liveness, loyalty and buying habits of the user in a specific time window, 7 days are adopted as a time window, in the period, the buying habits, the liveness period, the interest preference and other characteristics of the user are relatively stable and easy to observe, the peripheral consumption rule of the user can be identified by analyzing the consumption frequency in 7 days, a basis is provided for formulating an accurate marketing strategy, the time length of 7 days is short enough for online advertisement putting, the short-term effect of advertisements can be rapidly evaluated, and the advertisement content, the putting time and the frequency can be timely adjusted according to the data change of 7 days, so that the conversion rate and the return on investment are improved;

User content preferences:

If there is a group of contents to be interacted with Wherein the content of the user interaction/>The specific expression of the user content preference is:

；

In the method, in the process of the invention, Content representing user interactions; /(I)To indicate the function, the representation is at the time point/>User/>Whether or not to match content/>There is interaction, if there is better, then/>Otherwise/>；/>Representing user/>For content/>Is a preference degree of (a);

the expression can calculate the preference degree of the user on each content, and can be used for constructing the content preference portrait of the user so as to provide basis for personalized advertisement recommendation;

Product relevance:

If the degree of lifting is less than 1, it means that there is a negative correlation between product E and product F, and the customer who purchased product E is unlikely to purchase product F instead;

By analyzing the purchase history data of the user, the degree of promotion between different products can be calculated. Based on the product relevance with high lifting degree, the advertisement system can more accurately target advertisements of related products to potential customers, and the advertisement delivery effect and conversion rate are improved; the advertisement can be accurately put to the interested users by guiding the advertisement putting through the lifting degree, so that the click rate and the conversion rate of the advertisement are improved, the advertisement cost is reduced, and the return on investment of the advertisement is further improved.

The specific steps involved in the generation of the user labels by the user portrait construction module 13 are as follows;

s3.1, acquiring output user behavior data from the data analysis module 12, and analyzing to obtain consumption frequency A, content preference B and product relevance characteristics C;

the specific mapping rule is as follows:

Consumption frequency a mapping rule:

Content preference B mapping rules:

Product relevance feature C mapping rules:

By combining the mapping rules, the user portrait construction module 13 can accurately label the users according to the performances of the users in different dimensions, thereby being beneficial to the fine operation of the advertisement system for different user groups.

S3.3, classifying the user groups by adopting a user tag classification algorithm based on defined user tags, and dividing the users into groups with similar characteristics; dividing into groups with similar characteristics, realizing accurate positioning of target audience, being beneficial to more effectively distributing advertisement budget, reducing ineffective exposure and waste, and by identifying common interests, behavior patterns or consumption habit characteristics of users, the system can pertinently push products, contents or services which are possibly interested in the users to various user groups, increase click rate and conversion rate, and simultaneously provide a powerful tool for enterprises to design popularization strategies of cross-selling, binding selling or other related product combinations;

；/>

will user Assigned to the cluster corresponding to the nearest cluster center:

Order of the game ；

The update formula of the new cluster center is as follows:

；

S3.34, performing convergence judgment:

The intelligent advertisement marketing system based on the Internet further comprises an advertisement resource management unit 2, wherein the advertisement resource management unit 2 is used for managing and distributing advertisement space resources for delivery, and the advertisement resource management unit 2 comprises an advertisement resource library 21, a media resource integration module 22 and an advertisement resource management module 23;

wherein the advertisement resource library 21 is used for storing all advertisement resources for delivery, including but not limited to web site banner advertisements, social media screen inserting advertisements and APP screen opening advertisements;

The media resource integration module 22 is in butt joint with various media platforms through an API interface and is used for managing media platform resources for advertising, so that cross-platform advertising resource unified management and scheduling are realized;

The advertisement resource management module 23 adopts an advertisement resource super-optimization algorithm to distribute advertisements in the advertisement resource library 21 according to the flow characteristics and the value of different media channels according to the user portrait generated by the user analysis management unit 1.

In this embodiment, the advertisement resource super-optimization algorithm specifically includes:

；

wherein, the constraint condition is:

Total budget limit constraint: ；

In the method, in the process of the invention, Representing the total budget;

Non-negative constraint:

；

Period launch quota constraints:

；

The advertisement resource super-optimization algorithm is used, and the distribution of the advertisement resources in different media channels is dynamically adjusted according to the flow characteristics of the user portraits and the media platform so as to realize the maximum advertisement benefit;

Considering the flow characteristics and the value of different media platform resources, the advertisement resources are accurately distributed to the media platform with the most potential value, so that the resource waste can be avoided, the advertisement delivery efficiency is improved, the advertisement delivery is more accurate and effective, the advertisement resources are delivered to the media platform with higher flow and higher value by optimizing the advertisement position distribution, the exposure and the conversion rate of the advertisement can be increased, and the advertising income is increased. This is beneficial to both advertisers and advertising platforms, which can promote the revenue level of both parties;

by calculating the expected benefits of each advertisement position and combining with the budget constraint condition to make a decision, when a preset threshold value is reached or a specific rule is triggered, the system can immediately react to adjust the delivery strategy, such as reducing the budget allocation of advertisements with poor effect, improving the bid of high-quality advertisement positions or delivering more advertisement contents conforming to the characteristics of a target user group.

The intelligent advertisement marketing system based on the Internet also comprises a directional throwing unit 3, wherein the directional throwing unit 3 adopts a directional throwing algorithm to formulate an optimal advertisement throwing strategy based on matching advertisements of user figures and target audiences and combining advertisement targets and budgets;

The directional delivery unit 3 comprises a matching delivery module, the matching delivery module is based on a directional delivery algorithm, and performs accurate matching according to a user portrait and an advertisement positioning strategy, and pushes the most suitable advertisement content to a corresponding user, and the specific expression of the directional delivery algorithm is as follows:

；

In the method, in the process of the invention, Representing a user representation; /(I)A feature description representing an advertisement; /(I)Representing user portraits/>And characterization of advertisements/>Matching scores between the two advertisements are used for measuring the adaptation degree of a certain advertisement to a certain user; /(I)The personalized attribute, age, sex and occupation in the user portrait are represented; /(I)Representing target crowd attributes in an advertisement positioning strategy; /(I)A set of interest tags representing a user; /(I)Representing a set of advertisement-related interest keywords; /(I)Historical consumption behavior habit data representing a user; representing a product category with which the advertisement is associated; /(I) Representing a similarity function between two attributes; /(I)、/>And/>Respectively representing weight factors of different matching dimensions;

In this embodiment, considering the time sensitivity of the user behavior, the purchase demand and interest preference of the user generally have a certain timeliness, and the recent purchase behavior may reflect the current interest hotspots or demand trends of the user, so that the related advertisement content is pushed to more easily cause the attention and response of the user. Meanwhile, the possibility that advertisements can be seen and interacted by the advertisements can be improved according to the advertisements placed in the user active time intervals, and the optimized directed placement algorithm is introduced by the factors of the latest active time intervals, and is the following:

；

Wherein, Time sensitivity to user behavior, representing a recent activity period factor; /(I)Time information representing delivery of the advertisement; /(I)The time sensitivity of the user behavior is represented by the influence weight of the user behavior on the matching result, and the importance of relevant factors such as the latest purchasing behavior, the activity period and the like in the targeted delivery is represented.

When (when)When the value of (2) is larger, the influence of time sensitivity on the matching result is larger, and the model more pays attention to the latest behavior or activity period of the user; when/>When the value of (2) is smaller, the influence of the time factor on the matching result is smaller, and the model is more dependent on the matching results of other dimensions; the advertisement delivery strategy can be adjusted by an advertiser according to the user flow and the liveness of different time periods through analysis of the user liveness time periods, so that the optimal utilization of advertisement budget is realized.

The intelligent advertisement marketing system based on the Internet further comprises a throwing optimizing unit 4, wherein the throwing optimizing unit 4 adjusts the bidding strategy of advertisements on each media platform resource based on a throwing optimizing algorithm according to the net income of the advertisement position on each media platform resource, and tracks the behavior paths of users on different platforms by adopting a cross-platform joint optimizing algorithm, so that the advertisement throwing strategy is optimized, and the overall optimization of the cross-platform advertisement effect is realized;

The delivery optimizing unit 4 comprises a data monitoring module, a target setting module and a bid strategy optimizing module;

The data monitoring module is used for collecting and processing data related to advertisement delivery, including advertisement clicking and conversion data, the data can be from real-time data flow or historical data of an advertisement platform and used for evaluating the performance of advertisements on different media platform resources and calculating net benefits, the target setting module is used for determining an optimization target based on a cross-platform joint optimization algorithm, the optimization target comprises the maximum total benefits, the maximum Return On Investment (ROI) and the minimum cost, the bidding strategy optimization module is used for adjusting bidding strategies of the advertisements on each media platform resource through a throwing optimization algorithm based on the optimization target according to expected benefits and target setting, and the throwing optimization algorithm is used for dynamically adjusting the bidding strategies based on the expected benefits and the real-time data so as to achieve optimal delivery effect.

In this embodiment, the specific expression of the delivery optimization algorithm is:

；

wherein the constraint meets the total cost not exceeding the budget ：

；

based on the optimization objective of maximizing the total revenue, then:

；

In the method, in the process of the invention,Representing a lagrangian function; /(I)

By respectively corresponding toAnd/>Taking the partial derivative and setting to 0 to find the extreme point:

；

For global bidding coefficients：

；

The advertising amount and bidding strategy on different media resources are dynamically adjusted to maximize total income and ensure that budget is not exceeded, thereby improving advertising effect and efficiency and enabling advertisers to obtain maximum returns within limited budget.

The specific expression related to the cross-platform joint optimization algorithm is as follows:

media assets Exposure to media assets/>The click effect is as follows:

；

media assets Exposure to media assets/>The click effect is as follows:

；

Then the objective function K is jointly optimized:

；

Wherein the impression optimization algorithm adjusts the bid strategy (i.e., global bid coefficient) of the advertisement on each media platform resource based on the joint optimization objective function K) Delivery volume per platform/>; By considering the synergistic effect among the platforms, the click probability of the user on other platforms can be improved by the exposure among the platforms, the synergistic effect helps to promote the overall advertising effect, and the gains of the independent release of each platform and the mutual influence among the platforms are comprehensively considered so as to realize the overall optimization of the cross-platform advertising effect.

Example 2:

embodiment 2 of the present invention differs from embodiment 1 in that this embodiment describes an internet-based intelligent advertisement marketing method used by an internet-based intelligent advertisement marketing system.

The intelligent advertisement marketing method based on the Internet is used for the intelligent advertisement marketing system based on the Internet, and comprises the following steps:

s10.1, collecting user behavior data by a data collecting module 11, calculating the consumption frequency, content preference and content interaction condition of a user by a data analyzing module 12 through a user behavior analyzing algorithm, generating user labels comprising the consumption frequency, the content preference and the product relevance according to the behavior characteristics of the user by a user portrait construction module 13, classifying the user into groups with similar characteristics, and forming a user portrait;

S10.2, a plurality of media platform resources are butted through a media resource integration module 22, cross-platform advertisement space resources are uniformly managed and scheduled, an advertisement resource super-optimization algorithm is applied by an advertisement resource management module 23, expected benefits are maximized under the consideration of budget constraint and time slot quota throwing factors, and distribution strategies of advertisements in different media channels are dynamically adjusted;

And S10.4, finally, monitoring advertisement delivery effect in real time through a delivery optimizing unit 4, collecting and processing advertisement clicking and conversion related data, setting the maximized total income as an optimizing target through a target setting module, dynamically adjusting global bidding coefficients and delivery amount on each media resource by adopting a delivery optimizing algorithm, and realizing the maximization of advertisement benefit on the premise of meeting budget limit.

The foregoing has shown and described the basic principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the above-described embodiments, and that the above-described embodiments and descriptions are only preferred embodiments of the present invention, and are not intended to limit the invention, and that various changes and modifications may be made therein without departing from the spirit and scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.

Claims

1. An internet-based intelligent advertising marketing system, comprising:

The user analysis management unit (1), the user analysis management unit (1) collects user behavior data through the data collection module (11), the data analysis module (12) analyzes the user behavior data to identify the consumption frequency, content preference and product relevance characteristics of the user, the user portrait construction module (13) generates user labels based on the output result of the data analysis module (12), and classifies user groups according to the user labels to form different user group categories;

An advertising resource management unit (2), wherein the advertising resource management unit (2) is used for managing and distributing advertising space resources for delivery, and the advertising resource management unit (2) comprises an advertising resource library (21), a media resource integration module (22) and an advertising resource management module (23);

The targeted delivery unit (3), the targeted delivery unit (3) is based on matching advertisements with target audiences of user portraits, combines advertisement targets and budgets, adopts a targeted delivery algorithm to formulate an optimal advertisement delivery strategy, introduces a latest active period factor to optimize the targeted delivery algorithm, and the optimized targeted delivery algorithm is as follows:

；

In the method, in the process of the invention, Representing a user representation; /(I)A feature description representing an advertisement; /(I)Representing optimized user portraits/>And characterization of advertisements/>A match score between; /(I)Representing personalized attributes in the user representation; /(I)Representing target crowd attributes in an advertisement positioning strategy; /(I)A set of interest tags representing a user; /(I)Representing a set of advertisement-related interest keywords; /(I)Historical consumption behavior habit data representing a user; /(I)Representing a product category with which the advertisement is associated; /(I)Representing a similarity function between two attributes; /(I)、/>And/>Respectively representing weight factors of different matching dimensions; /(I)Time sensitivity to user behavior; /(I)Time information representing delivery of the advertisement; /(I)An influence weight of time sensitivity of the user behavior on the matching result is represented;

And the delivery optimizing unit (4) adjusts the bidding strategy of the advertisement on each media platform resource based on the delivery optimizing algorithm according to the net income of the advertisement position on each media platform resource, and tracks the behavior paths of the user on different platforms by adopting a cross-platform joint optimizing algorithm so as to optimize the advertisement delivery strategy.

2. The internet-based intelligent advertising marketing system of claim 1, wherein the data analysis module (12) is based on a user behavior analysis algorithm, and the specific expression involved in analyzing the user behavior data is:

frequency of user consumption:

In the method, in the process of the invention,Representing user/>Frequency of purchases over the past 7 days; /(I)Representing user/>At the time point/>The number of purchases made; /(I)Representing a time window; /(I)Representing a time point/>Belonging to time window/>；

User content preferences:

In the method, in the process of the invention,Content representing user interactions; /(I)Is an indication function; /(I)Representing user/>For content/>Is a preference degree of (a);

Product relevance:

3. The internet-based intelligent advertising marketing system of claim 1, wherein the specific steps involved in the user profile construction module (13) generating the user tag are;

s3.1, acquiring output user behavior data from a data analysis module (12) for analysis to obtain consumption frequency A, content preference B and product relevance characteristics C;

s3.3, classifying the user groups by adopting user label classification based on defined user labels, and dividing the users into groups with similar characteristics;

and S3.4, integrating the user tag and the group information to generate a user image.

4. The internet-based intelligent advertising marketing system according to claim 1, wherein the advertising asset library (21) is configured to store all advertising assets for delivery;

The media resource integration module (22) is in butt joint with various media platforms through an API interface and is used for managing media platform resources for advertising, so that cross-platform advertising resource unified management and scheduling are realized;

And the advertisement resource management module (23) distributes advertisements in the advertisement resource library (21) according to the flow characteristics and the value of different media channels by adopting an advertisement resource super-optimization algorithm according to the user portrait generated by the user analysis management unit (1).

5. The internet-based intelligent advertising marketing system of claim 4, wherein the advertising resource super-optimization algorithm is specifically:

；

In the method, in the process of the invention, Represents the/>The advertisement is at the/>Exposure on the individual media platform resources; /(I)Represents the/>The advertisement is at the/>Click rate on individual media platform resources; /(I)Represents the/>The advertisement is at the/>Conversion rate on individual media platform resources; Represents the/> The advertisement is at the/>Average revenue per user brought on the individual media platform resources; /(I)Represents the/>The advertisement is at the/>Display cost per thousand times on the individual media platform resources; /(I)Representing the number of ad spots; /(I)Representing the number of media platform resources; /(I)Representing the expected profit sum generated by all the advertisement spots on each media platform resource minus the corresponding putting cost sum, and obtaining the net profit;

wherein, the constraint condition is:

Total budget limit constraint: ；

In the method, in the process of the invention, Representing the total budget;

Non-negative constraint:

；

Period launch quota constraints:

；

6. The internet-based intelligent advertisement marketing system according to claim 1, wherein the targeting unit (3) comprises a matching delivery module, the matching delivery module is based on a targeting delivery algorithm, and performs accurate matching according to the user portraits and advertisement positioning strategies, and pushes the most suitable advertisement content to the corresponding user, and the specific expression of the targeting delivery algorithm is:

。

7. the internet-based intelligent advertising marketing system according to claim 1, wherein the impression optimization unit (4) comprises a data monitoring module, a targeting module, and a bid policy optimization module;

The data monitoring module is used for collecting and processing data related to advertisement delivery, the target setting module is used for determining an optimization target based on a cross-platform joint optimization algorithm, and the bidding strategy optimization module is used for adjusting the bidding strategy of advertisements on each media platform resource through the delivery optimization algorithm based on the optimization target according to expected benefits and target setting.

8. The internet-based intelligent advertising marketing system of claim 7, wherein the delivery optimization algorithm is specifically expressed as:

；

In the method, in the process of the invention, Representing global bid coefficients; /(I)Expressed in/>The amount of delivery on the individual media resources; /(I)Expressed in/>Average revenue on individual media assets; /(I)Representing the number of media assets; /(I)Is an index variable; /(I)Expressed in/>Average cost on individual media assets;

wherein the constraint meets the total cost not exceeding the budget ：

；

based on the optimization objective of maximizing the total revenue, then:

；

For global bidding coefficients：

；

In the method, in the process of the invention,Representing lagrangian multipliers; /(I)Represents the Lagrangian function L versus the delivered-quantity/>Is a partial derivative of (2); /(I)Representing the Lagrangian function L versus the global bid coefficient/>Is a partial derivative of (c).

9. The internet-based intelligent advertising marketing system of claim 8, wherein the cross-platform joint optimization algorithm involves the following specific expressions:

media assets Exposure to media assets/>The click effect is as follows:

；

media assets Exposure to media assets/>The click effect is as follows:

；

Then the objective function K is jointly optimized:

；

10. An internet-based intelligent advertisement marketing method for an internet-based intelligent advertisement marketing system according to any one of claims 1 to 9, comprising the steps of:

S10.1, collecting user behavior data by a data collecting module (11), calculating the consumption frequency, content preference and content interaction condition of a user by a data analyzing module (12) through a user behavior analyzing algorithm, generating user labels comprising the consumption frequency, the content preference and the product relevance according to the behavior characteristics of the user by a user portrait construction module (13), and classifying the user into groups with similar characteristics to form user portraits;

S10.2, butting a plurality of media platform resources through a media resource integration module (22), uniformly managing and scheduling cross-platform advertisement space resources, and dynamically adjusting the distribution strategy of advertisements in different media channels by using an advertisement resource super-optimization algorithm through an advertisement resource management module (23) under the consideration of budget constraint and time slot quota factors, and maximizing expected benefits;

S10.3, calculating the matching degree between the user portrait and the advertisement features by adopting a directional delivery algorithm through a matching delivery module, carrying out accurate matching by combining personalized attributes, interest labels and historical consumption behaviors, formulating an optimal advertisement delivery strategy, and pushing the most suitable advertisement content to the corresponding user;

And S10.4, finally, monitoring advertisement delivery effect in real time through a delivery optimizing unit (4), collecting and processing advertisement clicking and conversion related data, setting the maximized total income as an optimizing target through a target setting module, dynamically adjusting the global bidding coefficient and the delivery amount on each media resource by adopting a delivery optimizing algorithm, and realizing the maximization of advertisement benefit on the premise of meeting budget limit.