Advertisement interaction degree evaluation method and system based on user behaviors
By constructing user interaction animation and behavioral pattern analysis, combined with factor analysis and regression analysis, the systematic and personalized shortcomings of existing advertising evaluation methods are solved, accurate evaluation and dynamic optimization of advertising interaction are achieved, and advertising delivery effect and return on investment are improved.
Patent Information
- Application Number
- CN202510588684.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-08
- Publication Date
- 2025-08-08
AI Technical Summary
The existing advertising evaluation methods lack systematicity and personalization, and cannot fully reflect the depth of user interaction, insufficient dynamic feedback and optimization capabilities, which affect the utilization rate of advertising resources and return on investment.
User interaction animation is constructed through machine learning algorithms, cluster analysis is used to identify behavioral patterns, combine factor analysis and regression analysis, generate interaction scores, and adjust advertising strategies in real time.
It has achieved comprehensive and accurate advertising interaction assessment, improved the matching degree between advertising content and user interests, and improved the adaptability and return on investment of the advertising delivery system.
Smart Images

Figure CN120450781A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of big data processing, and primarily to a method and system for evaluating advertising interactivity based on user behavior. Background Art
[0002] With the rapid development of internet technology and mobile smart devices, digital advertising has become a crucial channel for brand communication and product promotion. Traditional advertising effectiveness evaluation methods rely primarily on a single or limited set of metrics, such as impressions and click-through rates, to provide a simple quantitative analysis of the actual effectiveness of advertising. However, with the increasing complexity of user behavior, evaluating advertising effectiveness solely through a single metric is no longer able to fully reflect users' true interest in advertising content and the depth of their engagement.
[0003] In existing technologies, some advertising platforms have begun to introduce multi-dimensional indicators such as user dwell time, ad viewing completion, and social interaction (such as likes, comments, and shares) in the hope of more comprehensively evaluating the effectiveness of ad interaction. However, these methods often have the following problems: 1. Analysis of indicators in isolation There is a lack of systematic correlation analysis between various interaction indicators, and it is impossible to identify the potential behavioral motivations behind different user behavior patterns, resulting in one-sided evaluation results and limited optimization strategy effects.
[0004] 2. Lack of user portraits and personalized assessments Most current evaluation methods use overall average statistics and fail to conduct personalized interaction evaluation based on user characteristics (interest preferences, activity, response habits, etc.), resulting in poor application effects in segmented user groups.
[0005] 3. Insufficient dynamic feedback and real-time optimization capabilities Some system evaluation results lag behind the ad display process, and are unable to dynamically adjust ad content, bidding, or display strategies based on real-time interactive data, reducing ad resource utilization and return on investment.
[0006] 4. Insufficient identification of factors affecting advertising effectiveness Existing solutions lack quantitative modeling and in-depth exploration of the influence of various factors on advertising interactive behavior, making it difficult to effectively guide advertising creative design, delivery strategy formulation, and audience positioning optimization.
[0007] Therefore, there is an urgent need for a technical solution that can systematically evaluate advertising interactivity based on users' multi-dimensional interactive behavior data, explore potential behavioral patterns, and support dynamic feedback and personalized optimization to improve advertising effectiveness, enhance user experience and advertiser revenue. Summary of the Invention
[0008] To address the above shortcomings, the present invention proposes a method for evaluating advertising interactivity based on user behavior, the specific steps of which are as follows: The system collects key behavioral indicators of users in advertising interactions in real time according to preset conditions as user interaction behavior data, and uses machine learning algorithms to build user interaction profiles based on the user interaction behavior data; According to the engagement scoring mechanism, different weights are assigned to different interactive behaviors, including high-weighted behaviors, medium-weighted behaviors, and low-weighted behaviors, to generate the corresponding ad engagement score; Using cluster analysis algorithms to further mine the user interaction behavior data and identify potential user behavior patterns; The interaction score of the advertisement is correlated with the actual data of the advertisement to generate a predicted rate ranking of user interaction behavior; based on the real-time feedback of the advertisement interaction, the advertisement delivery platform adaptively adjusts the advertisement strategy.
[0009] Furthermore, the key behavioral indicators include: click-through rate, dwell time, scroll depth, ad viewing completion, communication interaction and bounce rate.
[0010] Furthermore, the use of machine learning algorithms to construct user interaction profiles specifically includes: generating fine-grained user interest profiles by mapping and fusing the multi-dimensional features of user interaction behavior data, and training them into machine learning algorithm models. The model formula is as follows: ; in, represents the multidimensional interest vector of user i, n represents the total number of behavior samples, represents the behavior intensity score, represents the response delay of user i to the jth action, represents the behavioral diversity score of user i in the jth behavior, represents the feature vector of the interest category to which behavior j belongs, The weight hyperparameter representing the strength of the behavior, The weight hyperparameter representing the timeliness of the response, A weight hyperparameter representing the diversity of behaviors.
[0011] Furthermore, the specific implementation process of the cluster analysis algorithm is as follows: Extracting training features from the user interaction behavior data, the training features including: click-through rate, average dwell time, ad viewing completion, number of social interactions, bounce rate, average response time, and ad category preference distribution; normalizing numerical features to a mean of 0 and a standard deviation of 1, and representing ad category preferences using normalized vectors; The optimal number of clusters k is selected by the elbow method, and the sum of the squared errors (SSE) under different k values is calculated to obtain the inflection point. Using the determined k, perform K-Means clustering training, input the standardized feature vector of each user, and output the cluster label of each user; continuously re-cluster and dynamically update the cluster label.
[0012] Furthermore, the interaction score of the advertisement is subjected to correlation analysis with the actual data of the advertisement, specifically including: using factor analysis to identify potential factors that affect the interaction of the advertisement; using the potential factors as input features, and using regression analysis to evaluate the quantitative relationship between the interaction of the advertisement and the actual conversion rate.
[0013] The implementation process of the factor analysis includes: Bartlett's sphericity test was used to test the input data for factor variables, and the number of factors retained was determined based on the criterion of eigenvalue > 1 in principal component analysis; Orthogonal rotation is used to obtain a rotated factor loading matrix, which represents the contribution of each original variable to each factor.
[0014] The implementation process of the regression analysis includes constructing a regression equation as follows: ;
[0015] Among them, y represents the advertising conversion rate, Represents each ad engagement metric. represents the intercept term, represents the regression coefficient of each indicator, represents the error term; use The explanatory power of the values was evaluated, and the regression coefficient significance test was used to determine the factors with the greatest impact on the conversion rate.
[0016] According to a second aspect of the present invention, a user behavior-based advertising interactivity evaluation system is proposed, which is characterized by including the following configurations: A user profile building unit configured to collect key behavioral indicators of users in advertising interactions in real time according to preset settings as user interaction behavior data, and to build a user interaction profile based on the user interaction behavior data using a machine learning algorithm; An engagement scoring and weighting unit configured to assign differentiated weights to different engagement behaviors based on an engagement scoring mechanism, including high-weighted behaviors, medium-weighted behaviors, and low-weighted behaviors, to generate an engagement score for the corresponding advertisement; a behavior pattern analysis and prediction unit configured to further mine the user interaction behavior data using a cluster analysis algorithm to identify potential user behavior patterns; an interaction correlation analysis unit configured to perform correlation analysis on the interaction score of the advertisement and the actual data of the advertisement to generate a predicted rate ranking of user interaction behavior; The dynamic optimization and strategy adjustment unit is configured to adaptively adjust the advertising strategy of the advertising delivery platform based on real-time feedback on advertising interactivity.
[0017] According to a third aspect of the present invention, a computer program product is provided, on which one or more computer programs are stored. When the one or more computer programs are executed by a computer processor, the above method is implemented.
[0018] The above one or more technical solutions in the embodiments of the present application have at least one of the following technical effects: (1) Comprehensive and accurate engagement evaluation By integrating multi-dimensional behavioral data and systematically modeling it, we can overcome the shortcomings of existing technologies such as single evaluation indicators and one-sided analysis, and achieve a comprehensive and detailed evaluation of the interactive effects of advertising.
[0019] (2) Personalized optimization based on behavioral patterns By combining user group characteristics and individual behavior portraits, we develop differentiated advertising push and content optimization strategies to significantly improve the matching degree between advertising content and target user interests.
[0020] (3) Explore key influencing factors and guide strategy optimization The factor analysis method is used to extract the main potential factors affecting advertising interactivity, providing a scientific basis for advertising creative design, display location optimization, time period arrangement, etc.
[0021] (4) Support dynamic feedback and real-time adjustment Introduce a dynamic feedback mechanism to track changes in user interactive behavior in real time, quickly respond to changes in market environment and user interests, and improve the adaptive ability and closed-loop optimization capability of the advertising delivery system.
[0022] (5) Self-trained user profile generation model improves prediction accuracy and profile update frequency The self-developed portrait generation model proposed by the present invention can automatically learn and update the user's multi-dimensional interest characteristics, accurately predict the user's interactive intention towards different advertising content, and significantly improve the personalization and precise conversion capabilities of advertising push.
[0023] (6) Improve advertising return on investment (ROI) Based on accurate evaluation and optimized decision-making, it significantly improves the click-through rate, conversion rate and brand influence of advertisements, helping advertisers achieve higher returns on investment. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] The accompanying drawings are included to provide a further understanding of the embodiments and are incorporated into and constitute a part of this specification. The accompanying drawings illustrate the embodiments and, together with the description, serve to explain the principles of the present invention. Other embodiments and many of the expected advantages of the embodiments will be readily apparent as they become better understood by reference to the following detailed description. The elements of the drawings are not necessarily to scale with respect to each other. Like reference numerals designate corresponding similar parts.
[0025] Figure 1 A schematic diagram of the structure of a method for evaluating advertisement interactivity based on user behavior according to an embodiment of the present invention is shown.
[0026] Figure 2 It is a structural diagram of a computer system suitable for implementing the electronic device of the embodiment of the present application. DETAILED DESCRIPTION
[0027] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the relevant invention are shown in the accompanying drawings.
[0028] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0029] Figure 1 FIG. 1 shows a schematic diagram of a method for evaluating advertisement interactivity based on user behavior according to an embodiment of the present invention. Figure 1 As shown: S1. The system collects key behavioral indicators of users in advertising interactions in real time according to preset settings as user interaction behavior data, and uses machine learning algorithms to construct user interaction profiles based on the user interaction behavior data; The key behavioral indicators include: click-through rate, dwell time, scroll depth, ad viewing completion, communication interaction and bounce rate.
[0030] Click-through rate: How often users click on your ad.
[0031] Dwell time: The time a user stays on an ad, reflecting the ad's appeal.
[0032] Scroll depth: If the ad is a scrollable content (such as a video ad or banner ad), the depth of the user's scrolling can be used as a measure of engagement.
[0033] Ad viewing completion rate: For video ads, the rate at which users complete watching the video.
[0034] Social interactions such as likes, shares, and comments: Likes, shares, or comments on social media ads can provide in-depth indicators of user interaction.
[0035] Bounce rate: The proportion of users who quickly leave the page without any interaction after the ad is displayed. A low bounce rate usually means that the ad has a high degree of interaction.
[0036] The use of machine learning algorithms to construct user interaction profiles specifically includes: generating fine-grained user interest profiles by mapping and fusing the multi-dimensional features of user interaction behavior data, and training them into machine learning algorithm models. The model formula is as follows: ; in, represents the multidimensional interest vector of user i, n represents the total number of behavior samples, represents the behavior intensity score, represents the response delay of user i to the jth action, represents the behavioral diversity score of user i in the jth behavior, represents the feature vector of the interest category to which behavior j belongs, The weight hyperparameter representing the strength of the behavior, The weight hyperparameter representing the timeliness of the response, A weight hyperparameter representing the diversity of behaviors.
[0037] The model can automatically capture the user behavior characteristics of the three dimensions of "depth + immediacy + breadth", and the output result is a vector for further data mining. It also supports dynamic updates. Every time a user adds a new piece of behavior data, the interest vector can be incrementally updated.
[0038] S2. Based on the engagement scoring mechanism, different weights are assigned to different engagement behaviors, including high-weighted behaviors, medium-weighted behaviors, and low-weighted behaviors, to generate an engagement score for the corresponding ad. In this embodiment, the following weight distribution method is provided, but is not limited to the following rules: Click behavior is considered as direct interaction between users and ads, reflecting the attractiveness and relevance of the ads. High weight refers to users who click on ad content, with a score of 1.0 or higher, indicating a high degree of attention to the ad content; medium weight refers to users who do not interact further or leave the page after clicking on the ad, with a score of 0.7.
[0039] Video ads typically take a long time to connect with users, so viewing behavior has a high engagement value. A complete view is considered high-weighted: if the user watches more than 80% of the video, it will receive a score of 1.2, indicating that the ad content is highly engaging. A mid-view is considered medium-weighted: if the user watches between 50% and 80% of the video, it will receive a score of 0.8, indicating that the user has some interest. A short view is considered low-weighted: if the user watches less than 50% of the video, it will receive a score of 0.5, indicating that the ad is less appealing to the user.
[0040] The longer the dwell time, the higher the user's interest in the ad. This is especially true for content-rich ads, where longer dwell times generally indicate higher engagement. A dwell time exceeding 75% of the ad display duration receives a score of 1.0, indicating high engagement. A dwell time between 50% and 75% of the ad display duration receives a score of 0.7, indicating some interest. A dwell time less than 50% of the ad display duration receives a score of 0.3, indicating lower engagement.
[0041] The deeper the scrolling depth, the higher the user's interest and engagement with the ad content. Deep scrolling is weighted high: users scrolling to 75%-100% of the ad content receive a score of 1.0, indicating strong interest. Moderate scrolling is weighted medium: users scrolling to 40%-75% of the ad content receive a score of 0.7, indicating some interest. Shallow scrolling is weighted low: users scrolling less than 40% receive a score of 0.4, indicating low engagement.
[0042] The strength of social interaction is typically higher than a simple click because it reflects deep user engagement and a positive attitude toward the ad. Sharing ad content receives a score of 1.5, reflecting strong recommendation intent and typically associated with higher conversion potential. Commenting on ad content receives a score of 1.2, indicating a deep interest in the ad and a willingness to express opinions. Liking an ad receives a score of 0.8, indicating approval.
[0043] A low bounce rate generally indicates high ad engagement, while a high bounce rate indicates the ad content isn't engaging users. A bounce rate below 30% receives a score of 1.0, indicating the ad is effectively engaging users. A bounce rate between 30% and 50% receives a score of 0.6, indicating the ad is engaging. A bounce rate above 50% receives a score of 0.3, indicating poor ad performance.
[0044] For example, suppose a user behaves as follows: Clicking an ad: 2 points; watching a video in full: 1.5 points; dwell time: 1 point; scroll depth: 0.8 points; social interaction: 1 point; bounce rate: 0.5 points.
[0045] Comprehensive score = 2 (clicks) + 1.5 (video views) + 1 (stay time) + 0.8 (scroll depth) + 1 (social interaction) + 0.5 (bounce rate) Overall interaction score = 6.8 points.
[0046] After weighting, truly meaningful behaviors can be highlighted. For example, users who simply click on an ad but immediately exit will receive a low score, while users who stay longer and comment and share will receive a high score. This, combined with user profile modeling results, can more accurately reflect users' true interests and preferences, unaffected by noisy data.
[0047] During subsequent cluster analysis, the model primarily relies on weighted interaction feature vectors. Without appropriate weighting, clustering results can be affected by noise (such as occasional meaningless clicks), leading to inaccurate classification. Weighted allocation ensures that truly important behaviors play a leading role in clustering, resulting in more reasonable and commercially valuable user group segmentation.
[0048] Subsequent systems for real-time optimization and personalized recommendations need to adjust their strategies based on real-time interaction feedback. If each piece of interaction feedback is appropriately weighted, the system can more quickly identify which ad content, delivery timeframes, or demographic combinations are most effective and optimize accordingly. This weighted distribution ensures that feedback signals are both sensitive and accurate, avoiding misjudgments of results due to minor or low-quality interactions.
[0049] In summary, weight distribution is the "quality cornerstone" of the entire system modeling, and reasonable weight distribution can promote better technical effects.
[0050] S3. Using a cluster analysis algorithm to further mine the user interaction behavior data and identify potential user behavior patterns; The specific implementation process of the cluster analysis algorithm is as follows: Data preparation and standardization: Extract training features from the user interaction behavior data. The training features include: click-through rate, average dwell time, ad viewing completion rate, number of social interactions, bounce rate, average response time, and ad category preference distribution. Numerical features are normalized to a mean of 0 and a standard deviation of 1. Ad category preferences are represented by normalized vectors. Cluster number selection: The optimal cluster number k is selected by the elbow method, and the sum of squared errors (SSE) under different k values is calculated to obtain the inflection point. Training model: Use a certain k to perform K-Means clustering training, input the standardized feature vector of each user, and output the cluster label of each user (for example: belonging to category 1 or category 2 user group); continuously re-cluster and dynamically update the cluster label.
[0051] The K-Means clustering model is suitable for discovering natural groups in user behavior data; it is fast to train and easy to interpret; and it has good scalability for large-scale user data.
[0052] Before selecting the number of clusters, the input data can be further normalized, and the category features can be One-hot encoded and incorporated into the feature vector.
[0053] By analyzing the feature mean of each cluster center, different user behavior patterns can be explained, as shown in Table 1, which outputs the data structure of an embodiment: Table 1
[0054] S4. Correlation analysis is performed on the interaction score of the advertisement and actual advertisement data to generate a predicted rate ranking of user interaction behavior; The association analysis specifically includes: using factor analysis to identify potential factors that affect advertising interactivity; using the potential factors as input features, and using regression analysis to evaluate the quantitative relationship between advertising interactivity and actual conversion rate.
[0055] Factor analysis extracts hidden latent factors from multiple ad interaction indicators and simplifies the data structure. When receiving the data processed in the above steps, to ensure sufficient sample size, the sample size is generally required to be at least 5-10 times the number of variables. Further steps are as follows: Bartlett's sphericity test was used to test the input data for factor variables, and the number of factors retained was determined based on the criterion of eigenvalue > 1 in principal component analysis; Orthogonal rotation is used to obtain a rotated factor loading matrix, which represents the contribution of each original variable to each factor.
[0056] Regression analysis quantitatively studies the impact of user ad engagement on ad conversion rate by establishing a regression model. The constructed regression equation is as follows: ; Among them, y represents the advertising conversion rate, Represents each ad engagement metric. represents the intercept term, represents the regression coefficient of each indicator, represents the error term; use The explanatory power of the values was evaluated, and the regression coefficient significance test was used to determine the factors with the greatest impact on the conversion rate.
[0057] To sum up, the correlation analysis is as follows: first perform factor analysis to extract a few core potential factors, then use these factor scores as input features to perform regression analysis to predict ad conversion rates or other target values. This can reduce feature redundancy, improve model stability, and make factors easier to interpret, which is much clearer than directly using a bunch of raw indicators.
[0058] S5. Based on real-time feedback on advertising engagement, the advertising platform adaptively adjusts its advertising strategy.
[0059] Based on the user interaction data obtained above, the system dynamically optimizes advertising strategies, including: Real-time bidding optimization: Dynamically adjust ad bids based on engagement data to improve exposure efficiency.
[0060] Ad content optimization: Make quick adjustments to ad content with low engagement, such as replacing creative materials and optimizing ad copy.
[0061] Personalized recommendations: Based on user interaction profiles, we push advertising content that best suits users’ interests in real time to increase interaction rates.
[0062] In addition, by integrating advertising interaction data from different platforms, a unified cross-channel user interaction profile is established, and a comprehensive evaluation of the overall advertising effect is achieved through a unified indicator system.
[0063] Design adaptive feedback mechanisms, including A / B testing, engagement feedback, and dynamic content personalization. A / B testing and engagement feedback involve running A / B tests between different ad versions, comparing engagement data in real time, and selecting the optimal version. Dynamic content personalization dynamically adjusts ad content based on real-time user engagement, achieving highly personalized delivery.
[0064] Through continuous feedback and optimization cycles, the system continuously improves the overall ROI of advertising and user experience.
[0065] Based on the above steps, the present invention proposes an advertising interactivity evaluation system based on user behavior, which includes the following configurations, configured to execute any content of the above advertising interactivity evaluation method: A user profile building unit configured to collect key behavioral indicators of users in advertising interactions in real time according to preset settings as user interaction behavior data, and to build a user interaction profile based on the user interaction behavior data using a machine learning algorithm; An engagement scoring and weighting unit configured to assign differentiated weights to different engagement behaviors based on an engagement scoring mechanism, including high-weighted behaviors, medium-weighted behaviors, and low-weighted behaviors, to generate an engagement score for the corresponding advertisement; a behavior pattern analysis and prediction unit configured to further mine the user interaction behavior data using a cluster analysis algorithm to identify potential user behavior patterns; an interaction correlation analysis unit configured to perform correlation analysis on the interaction score of the advertisement and the actual data of the advertisement to generate a predicted rate ranking of user interaction behavior; The dynamic optimization and strategy adjustment unit is configured to adaptively adjust the advertising strategy of the advertising delivery platform based on real-time feedback on advertising interactivity.
[0066] Reference below Figure 2 , which shows a structural diagram of a computer system 200 suitable for implementing an electronic device of an embodiment of the present application. Figure 2 The electronic device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.
[0067] like Figure 2 As shown, the computer system 200 includes a central processing unit (CPU) 201, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 202 or a program loaded from a storage unit 208 into a random access memory (RAM) 203. Various programs and data required for the operation of the system 200 are also stored in the RAM 203. The CPU 201, the ROM 202, and the RAM 203 are connected to each other via a bus 204. An input / output (I / O) interface 205 is also connected to the bus 204.
[0068] The following components are connected to the I / O interface 205: an input section 206 including a keyboard, a mouse, and the like; an output section 207 including a liquid crystal display (LCD) and speakers; a storage section 208 including a hard disk; and a communication section 209 including a network interface card such as a LAN card or a modem. The communication section 209 performs communication processing via a network such as the Internet. A drive 210 is also connected to the I / O interface 205 as needed. A removable medium 211, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 210 as needed, so that computer programs read therefrom can be installed into the storage section 208 as needed.
[0069] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 209, and / or installed from the removable medium 211. When the computer program is executed by the central processing unit (CPU) 201, the above-mentioned functions defined in the method of the present application are executed. It should be noted that the computer-readable storage medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including, but not limited to, an electromagnetic signal, an optical signal, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium that can transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
[0070] Computer program code for performing the operations of the present application can be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., through the Internet using an Internet service provider).
[0071] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.
[0072] The modules described in the embodiments of the present application may be implemented by software or hardware.
[0073] As another aspect, the present application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or it may exist independently and not be assembled into the electronic device. The above computer-readable storage medium carries one or more programs. When the above one or more programs are executed by the electronic device, the electronic device: the system collects the key behavioral indicators of users in advertising interaction in real time according to the preset as user interaction behavior data, and uses a machine learning algorithm to build a user interaction portrait based on the user interaction behavior data; according to the interaction scoring mechanism, different weights are assigned according to different interaction behaviors, including: high-weight behavior, medium-weight behavior and low-weight behavior, to generate an interaction score for the corresponding advertisement; a cluster analysis algorithm is used to further mine the user interaction behavior data to identify potential user behavior patterns; the interaction score of the advertisement is correlated with the actual data of the advertisement to generate a predicted rate ranking of the user interaction behavior; based on the real-time feedback of the advertisement interaction, the advertising delivery platform adaptively adjusts the advertising strategy.
[0074] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the above-mentioned inventive concept. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.
Claims
1. A method for evaluating advertising interactivity based on user behavior, characterized in that: include: The system collects key behavioral indicators of users in advertising interactions in real time according to preset conditions as user interaction behavior data, and uses machine learning algorithms to build user interaction profiles based on the user interaction behavior data; According to the engagement scoring mechanism, different weights are assigned to different interactive behaviors, including high-weighted behaviors, medium-weighted behaviors, and low-weighted behaviors, to generate the corresponding ad engagement score; Using cluster analysis algorithms to further mine the user interaction behavior data and identify potential user behavior patterns; The interaction score of the advertisement is correlated with the actual data of the advertisement to generate a predicted rate ranking of user interaction behavior; based on the real-time feedback of the advertisement interaction, the advertisement delivery platform adaptively adjusts the advertisement strategy.
2. The advertising interactivity evaluation method according to claim 1, characterized in that: The key behavioral indicators include: click-through rate, dwell time, scroll depth, ad viewing completion, communication interaction and bounce rate.
3. The advertising interactivity evaluation method according to claim 1, characterized in that: The use of machine learning algorithms to construct user interaction profiles specifically includes: generating fine-grained user interest profiles by mapping and fusing the multi-dimensional features of user interaction behavior data, and training them into machine learning algorithm models. The model formula is as follows: ; in, represents the multidimensional interest vector of user i, n represents the total number of behavior samples, represents the behavior intensity score, represents the response delay of user i to the jth action, represents the behavioral diversity score of user i in the jth behavior, represents the feature vector of the interest category to which behavior j belongs, The weight hyperparameter representing the strength of the behavior, The weight hyperparameter representing the timeliness of the response, A weight hyperparameter representing the diversity of behaviors.
4. The advertising interactivity evaluation method according to claim 1, wherein: The specific implementation process of the cluster analysis algorithm is as follows: Extracting training features from the user interaction behavior data, the training features including: click-through rate, average dwell time, ad viewing completion, number of social interactions, bounce rate, average response time, and ad category preference distribution; normalizing numerical features to a mean of 0 and a standard deviation of 1, and representing ad category preferences using normalized vectors; The optimal number of clusters k is selected by the elbow method, and the sum of the squared errors (SSE) under different k values is calculated to obtain the inflection point. Using the determined k, perform K-Means clustering training, input the standardized feature vector of each user, and output the cluster label of each user; continuously re-cluster and dynamically update the cluster label.
5. The advertising interactivity evaluation method according to claim 1, wherein: The correlation analysis between the interaction score of the advertisement and the actual data of the advertisement specifically includes: using factor analysis to identify potential factors that affect the interaction of the advertisement; using the potential factors as input features, and using regression analysis to evaluate the quantitative relationship between the interaction of the advertisement and the actual conversion rate.
6. The advertising interactivity evaluation method according to claim 5, characterized in that: The implementation process of the factor analysis includes: Bartlett's sphericity test was used to test the input data for factor variables, and the number of factors retained was determined based on the criterion of eigenvalue > 1 in principal component analysis; Orthogonal rotation is used to obtain a rotated factor loading matrix, which represents the contribution of each original variable to each factor.
7. The advertising interactivity evaluation method according to claim 5, characterized in that: The implementation process of the regression analysis includes constructing a regression equation as follows: ; Among them, y represents the advertising conversion rate, Represents each ad engagement metric. represents the intercept term, represents the regression coefficient of each indicator, represents the error term; use The explanatory power of the values was evaluated, and the regression coefficient significance test was used to determine the factors with the greatest impact on the conversion rate.
8. An advertising interaction evaluation system based on user behavior, characterized in that: Includes the following configuration: A user profile building unit configured to collect key behavioral indicators of users in advertising interactions in real time according to preset settings as user interaction behavior data, and to build a user interaction profile based on the user interaction behavior data using a machine learning algorithm; An engagement scoring and weighting unit configured to assign differentiated weights to different engagement behaviors based on an engagement scoring mechanism, including high-weighted behaviors, medium-weighted behaviors, and low-weighted behaviors, to generate an engagement score for the corresponding advertisement; a behavior pattern analysis and prediction unit configured to further mine the user interaction behavior data using a cluster analysis algorithm to identify potential user behavior patterns; an interaction correlation analysis unit configured to perform correlation analysis on the interaction score of the advertisement and the actual data of the advertisement to generate a predicted rate ranking of user interaction behavior; The dynamic optimization and strategy adjustment unit is configured to adaptively adjust the advertising strategy of the advertising delivery platform based on real-time feedback on advertising interactivity.
9. A computer program product, characterized in that A computer program is stored thereon, which implements the method according to any one of claims 1 to 7 when executed by a processor.
10. A computing system, characterized in that The method comprises a processor and a memory, wherein the processor is configured to execute the method according to any one of claims 1 to 7.
Citation Information
Cited By
Right content delivery strategy generation method and device based on AI big data
CN120894081A
OTT advertisement optimization system based on pointing behavior of remote controller
CN121352884A
Advertisement multi-contact effect feedback method and system based on DDA
CN122155789A
