Hybrid advertisement carousel system based on dynamic optimization

Abstract

The invention discloses a hybrid advertisement carousel system based on dynamic optimization, and the system comprises the following modules: a data collection module which obtains user behavior data, an advertisement effect index and an environment variable in real time, the user behavior data comprises a click rate, a stay time length and an interaction frequency, and the environment variable comprises a time period, an equipment type and a network state; and the dynamic analysis module is used for analyzing the data acquired by the data acquisition module based on a machine learning model. When the method is implemented, the method has the following effects that the advertisement revenue is improved, the exposure ratio of the high-value advertisement is remarkably improved through dynamic optimization, and the rate of return on investment of the advertisement can also be improved; the resource utilization rate is high, traffic changes are responded in real time, the advertisement filling rate is remarkably improved, more advertisements can be played, and idle resources are reduced; the method is flexible and extensible, supports the quick adaptation of multiple advertisement forms and multiple terminal scenes, and puts advertisements according to the requirements of advertisers.

Description

Technical Field

[0001] The present application relates to the technical field of digital advertising delivery, and in particular to a hybrid advertising carousel system based on dynamic optimization. Background Art

[0002] With the continuous development of the self-media era, the form of advertising has also changed from traditional TV advertising to today's mobile phone advertising. That is, when users browse web pages, short videos and other information, advertisements are placed to increase the exposure of advertisements.

[0003] However, traditional advertising carousel systems usually use fixed sequences or simple carousel strategies, and cannot be dynamically adjusted based on user interests, real-time interactive data, or advertising effects. For example: static carousels lead to insufficient exposure of high-value ads, low user interest matching, and failure to change advertising delivery strategies in a timely manner based on user interests and the value of ads; lack of collaborative management of multiple advertising formats (such as video, graphics, and interactive ads), and often repeated delivery of single-format ads, resulting in poor advertising combination effects; and inability to respond to sudden traffic or changes in user behavior in real time, resulting in a waste of advertising resources. Summary of the invention

[0004] To this end, the present application provides a hybrid advertising carousel system based on dynamic optimization to solve the problem in the prior art that it is impossible to make dynamic adjustments based on user interests, real-time interactive data or advertising effects.

[0005] In order to achieve the above objectives, this application provides the following technical solutions:

[0006] A hybrid advertising carousel system based on dynamic optimization includes the following modules:

[0007] Data collection module: real-time acquisition of user behavior data, advertising effect indicators and environmental variables. The user behavior data includes click-through rate, dwell time and interaction frequency. The environmental variables include time period, device type and network status.

[0008] Dynamic analysis module: Analyzes the data collected by the data collection module based on the machine learning model, and outputs user interest tags, advertising value scores and priority weights;

[0009] Mixed ad pool: The mixed ad pool stores various forms of ad materials, including video ads, graphic ads, interactive ads, and dynamic creative ads, and extracts metadata tags associated with each type of ad.

[0010] Optimization and scheduling module: Based on the output results of the dynamic analysis module and combined with the advertising delivery goals, it generates the advertising play sequence and display strategy in real time;

[0011] Playback control module: rotates ads according to the optimized ad play sequence, and dynamically adjusts the playback rhythm and content combination of different forms of advertising materials based on real-time user feedback.

[0012] Preferably, the dynamic analysis module adopts the following method:

[0013] Build a user interest prediction model to generate interest vectors based on historical behavior data and real-time interaction data;

[0014] Calculate the advertising value score, combining the advertiser's bid, estimated click-through rate and user matching to score the advertising value;

[0015] Introduce a time decay factor to dynamically reduce the priority of repeatedly exposed ads.

[0016] Preferably, the management of the mixed advertisement pool includes:

[0017] Automatically categorize and store advertisements based on their format and metadata tags;

[0018] Set minimum exposure frequency constraints for each type of advertisement;

[0019] Supports dynamic insertion of temporary advertisements (such as promotional ads) and prioritizes matching with highly relevant users.

[0020] Preferably, the optimization scheduling module generates a playback sequence by the following steps:

[0021] (1) Filter the candidate ad set based on the similarity between the ad value score and the user interest vector;

[0022] (2) Based on the selected advertising set and combined with advertising format diversity constraints, avoid playing the same type of advertisements continuously;

[0023] (3) Use reinforcement learning algorithms to balance immediate benefits and long-term user experience and generate the optimal playback sequence.

[0024] Preferably, the dynamic adjustment of the playback control module includes:

[0025] When it is detected that the user skips the current ad, the display duration of subsequent ads of the same type will be shortened;

[0026] When users actively interact, extend the exposure time of relevant ads and insert supplementary content;

[0027] Dynamically switch ad resolution based on network bandwidth to ensure smooth playback.

[0028] Preferably, it also includes:

[0029] Configure differentiated advertising formats for different terminals;

[0030] Optimize ad layout and display logic based on screen size and interaction method.

[0031] Preferably, the system achieves closed-loop optimization in the following manner:

[0032] Collect user behavior data after each ad is played and update the input of the dynamic analysis module;

[0033] Regularly retrain machine learning models to adapt to user interest drift and advertising market changes.

[0034] Preferably, it also includes:

[0035] The visual dashboard is used to display advertising exposure, click-through rate, conversion rate and revenue indicators in real time;

[0036] Supports advertisers to customize delivery rules and effect warning thresholds.

[0037] Compared with the prior art, this application has at least the following beneficial effects:

[0038] Improved advertising revenue. Dynamic optimization significantly increases the exposure of high-value ads and improves the return on investment of ads.

[0039] High resource utilization, real-time response to traffic changes, and significantly improved ad fill rate, which means more ads can be played and less idle resources can be used;

[0040] It is flexible and scalable, supports rapid adaptation of multiple advertising formats and multiple terminal scenarios, and delivers advertisements according to advertisers' needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more intuitively illustrate the prior art and the present application, exemplary drawings are given below. It should be understood that the specific shapes and structures shown in the drawings should not generally be regarded as limiting conditions for implementing the present application; for example, those skilled in the art are capable of easily making conventional adjustments or further optimizations to the addition / reduction / attribution division, specific shapes, positional relationships, connection methods, dimensional ratios, etc. of certain units (components) based on the technical concepts and exemplary drawings disclosed in the present application.

[0042] Figure 1 A structural diagram of a hybrid advertising carousel system based on dynamic optimization provided in Example 1 of the present application. DETAILED DESCRIPTION

[0043] The present application is further described below in detail through specific embodiments in conjunction with the accompanying drawings.

[0044] like Figure 1 As shown, a hybrid advertising carousel system based on dynamic optimization includes the following modules:

[0045] Data collection module: real-time acquisition of user behavior data, advertising effect indicators and environmental variables. The user behavior data includes click-through rate, dwell time and interaction frequency. The environmental variables include time period, device type and network status.

[0046] Dynamic analysis module: Analyzes the data collected by the data collection module based on the machine learning model, and outputs user interest tags, advertising value scores and priority weights;

[0047] Mixed ad pool: The mixed ad pool stores various forms of ad materials, including video ads, graphic ads, interactive ads, and dynamic creative ads, and extracts metadata tags associated with each type of ad.

[0048] Optimization and scheduling module: Based on the output results of the dynamic analysis module and combined with the advertising delivery goals, it generates the advertising play sequence and display strategy in real time;

[0049] Playback control module: rotates ads according to the optimized ad play sequence, and dynamically adjusts the playback rhythm and content combination of different forms of advertising materials based on real-time user feedback.

[0050] When the technical solution of the present invention is implemented, the user's behavior data can be collected through the data collection module to determine the type of advertisements that the user may be interested in, and then the advertisements that the user may be interested in are selected from the mixed advertisement pool for delivery. In order to achieve the purpose of advertising delivery and increase advertising revenue, the display strategy is optimized through the optimization scheduling module.

[0051] The dynamic analysis module adopts the following method:

[0052] Build a user interest prediction model to generate interest vectors based on historical behavior data and real-time interaction data, such as the user's preferences for video types, which type of video the user stays on the longest, and which type of video the user interacts with the most.

[0053] Calculate the advertising value score, combine the advertiser's bid, estimated click-through rate and user matching to score the advertising value, estimate the advertising exposure rate, expected click-through rate and advertising return rate based on the price given by the advertiser, and predict the matching degree between the advertisement and the user based on the user's behavior habits;

[0054] By introducing a time decay factor, the priority of repeatedly exposed ads can be dynamically reduced. For repeatedly exposed ads, the priority of their exposure can be gradually reduced, which can not only avoid users from feeling disgusted with them, but also stimulate their interest.

[0055] The management of the mixed advertising pool includes:

[0056] Automatically categorize and store ads based on ad formats and metadata tags, so that ads can be matched with users and pushed to them;

[0057] Set a minimum exposure frequency constraint for each type of advertisement to ensure the number of times the advertisement is exposed, thereby increasing the number of times the advertisement is delivered;

[0058] It supports dynamic insertion of temporary ads (such as promotional ads) and gives priority to matching highly relevant users. It delivers temporary ads based on the tags of the temporarily inserted ads, and the target group is users with a high degree of match with the ad tags.

[0059] For the optimization scheduling module, the following implementation methods are provided:

[0060] The optimization scheduling module generates a playback sequence through the following steps: (1) screening candidate advertisement sets based on the similarity between advertisement value scores and user interest vectors; (2) avoiding continuous playback of similar advertisements based on the screened advertisement sets and advertisement format diversity constraints; and (3) using a reinforcement learning algorithm to balance immediate benefits and long-term user experience to generate an optimal playback sequence.

[0061] That is, during implementation, each advertisement value (the commercial value of the advertisement or the amount of money spent on the advertisement) is scored, and then an advertisement set that meets the user's interests is screened out based on the similarity of the user's interest vectors. The advertisement set contains several advertisements.

[0062] Then classify the ads in the ad set, such as shopping ads, game ads, application ads, and product ads (of course, other classifications are also possible: graphic ads, video ads, etc.), filter out one or several ads of each type, and when delivering the ads, make sure that the same type of ads are played in different orders to avoid the situation where ads of the same type are delivered one after another.

[0063] Finally, the revenue generated by the advertisements is evaluated by the order and type of advertisements. Based on the evaluation results, the order and type of advertisements are optimized to generate the optimal advertisement sequence.

[0064] The dynamic adjustment of the playback control module includes:

[0065] When it is detected that the user skips the current ad, the display duration of subsequent ads of the same type will be shortened;

[0066] When users actively interact, extend the exposure time of relevant ads and insert supplementary content;

[0067] Dynamically switch ad resolution based on network bandwidth to ensure smooth playback.

[0068] Also includes:

[0069] Configure differentiated advertising formats for different terminals (mobile, PC, smart TV);

[0070] Optimize ad layout and display logic based on screen size and interaction method.

[0071] The system achieves closed-loop optimization by:

[0072] Collect user behavior data after each ad is played and update the input of the dynamic analysis module;

[0073] Regularly retrain machine learning models to adapt to user interest drift and advertising market changes.

[0074] Also includes:

[0075] The visual dashboard is used to display the advertising exposure, click-through rate, conversion rate and revenue indicators in real time, so as to grasp the advertising data in a timely and clear manner;

[0076] It supports advertisers to customize delivery rules and effect warning thresholds, that is, to deliver advertisements according to the needs of advertisers, meeting the needs of different advertisers.

[0077] The technical features of the above embodiments may be arbitrarily combined (as long as there is no contradiction in the combination of these technical features). To make the description concise, not all possible combinations of the technical features in the above embodiments are described; these embodiments that are not explicitly written should also be considered to be within the scope of this specification.

Claims

1. A hybrid advertising carousel system based on dynamic optimization, characterized in that: Includes the following modules: Data collection module: real-time acquisition of user behavior data, advertising effect indicators and environmental variables. The user behavior data includes click-through rate, stay time, and interaction frequency. The environmental variables include time period, device type and network status. Dynamic analysis module: Analyzes the data collected by the data collection module based on the machine learning model, and outputs user interest tags, advertising value scores and priority weights; Mixed Ad Pool: The mixed ad pool stores various forms of ad materials, including video ads, graphic ads, interactive ads, and dynamic creative ads, and extracts metadata tags associated with each type of ad. Optimization and scheduling module: Based on the output results of the dynamic analysis module and combined with the advertising delivery goals, it generates the advertising play sequence and display strategy in real time; Playback control module: rotates ads according to the optimized ad play sequence, and dynamically adjusts the playback rhythm and content combination of different forms of advertising materials based on real-time user feedback.

2. A hybrid advertisement carousel system based on dynamic optimization according to claim 1, characterized in that: The dynamic analysis module adopts the following method: Build a user interest prediction model to generate interest vectors based on historical behavior data and real-time interaction data; Calculate the advertising value score, combining the advertiser's bid, estimated click-through rate and user matching degree; Introduce a time decay factor to dynamically reduce the priority of repeatedly exposed ads.

3. The hybrid advertisement carousel system based on dynamic optimization according to claim 1, characterized in that: The management of the mixed advertising pool includes: Automatically categorize and store advertisements based on their format and metadata tags; Set minimum exposure frequency constraints for each type of advertisement; Supports dynamic insertion of temporary advertisements and prioritizes matching with highly relevant users.

4. The hybrid advertisement carousel system based on dynamic optimization according to claim 1, characterized in that: The optimization scheduling module generates a playback sequence by the following steps: (1) Filter the candidate ad set based on the similarity between the ad value score and the user interest vector; (2) Based on the selected advertising set and combined with advertising format diversity constraints, avoid playing the same type of advertisements continuously; (3) Use reinforcement learning algorithms to balance immediate benefits and long-term user experience and generate the optimal playback sequence.

5. The hybrid advertisement carousel system based on dynamic optimization according to claim 1, characterized in that: The dynamic adjustment of the playback control module includes: When it is detected that the user skips the current ad, the display duration of subsequent ads of the same type will be shortened; When users actively interact, extend the exposure time of relevant ads and insert supplementary content; Dynamically switch ad resolution based on network bandwidth to ensure smooth playback.

6. The hybrid advertisement carousel system based on dynamic optimization according to claim 1, characterized in that: Also includes: Configure differentiated advertising formats for different terminals; Optimize ad layout and display logic based on screen size and interaction method.

7. The hybrid advertisement carousel system based on dynamic optimization according to claim 1, characterized in that: The system achieves closed-loop optimization by: Collect user behavior data after each ad is played and update the input of the dynamic analysis module; Regularly retrain machine learning models to adapt to user interest drift and advertising market changes.

8. The hybrid advertisement carousel system based on dynamic optimization according to claim 1, characterized in that: Also includes: The visual dashboard is used to display advertising exposure, click-through rate, conversion rate and revenue indicators in real time; Supports advertisers to customize delivery rules and effect warning thresholds.

Images