Advertisement putting method and system based on big data analysis, medium and program product

Abstract

The invention discloses an advertisement putting method and system based on big data analysis, a medium and a program product, and the method comprises the steps: calculating the acceptance degree data of a target user for various advertisement contents; determining consumption conversion links of the target user in different time periods based on the purchase behavior data; determining a key advertisement content combination and an optimal display time sequence according to the consumption conversion link; screening out the advertisement content from a preset advertisement library, wherein the combination similarity of the advertisement content and the key advertisement content is greater than a preset first similarity threshold value, and the click frequency of the advertisement content is smaller than a preset frequency, and obtaining an advertisement content exploration pool; the advertisement content in the advertisement content exploration pool is inserted into the optimal display time sequence in a progressive mode, and a dynamic advertisement putting sequence is formed; and putting the final advertisement content corresponding to the dynamic advertisement putting sequence to the target user according to the dynamic advertisement putting sequence. According to the method and the device, advertisement contents conforming to potential interests are brought to users, and continuous enhancement of the delivery effect is realized on the basis of keeping the existing delivery effect.

Description

Technical Field

[0001] The present application belongs to the field of electronic digital data processing, and in particular, relates to an advertising delivery method, system, medium and program product based on big data analysis. Background Art

[0002] With the continuous development of Internet technology and the rapid growth of user data, the accuracy and effectiveness of advertising are particularly important. Traditional advertising methods mainly rely on manual experience and simple user portraits to make advertising decisions. This method is not only inefficient, but also difficult to accurately grasp the actual needs of users, resulting in a waste of advertising resources and unsatisfactory advertising results.

[0003] In the existing technology, it is possible to build a user interest model by collecting data such as user browsing history, search history, and shopping behavior, and then place advertisements based on the model prediction results. This method can improve the accuracy of advertisement placement to a certain extent and make the advertisement content closer to the actual needs of users.

[0004] However, excessive pursuit of superficial data indicators such as user click-through rates when making advertising decisions may cause the system to tend to push advertising content that users are familiar with or easy to accept, limiting users' opportunities to be exposed to new things, and causing advertising to fall into low-level repetition, making it difficult to stimulate users' potential consumption interests, thereby affecting the long-term effectiveness of advertising. Summary of the invention

[0005] The present application provides an advertising delivery method, system, medium and program product based on big data analysis, which is used to bring advertising content that meets the potential interests of users, and achieves continuous enhancement of delivery effects while maintaining existing delivery effects.

[0006] In the first aspect, the present application provides an advertising delivery method based on big data analysis to obtain historical click data and purchase behavior data of target users, where the historical click data includes the content of the clicked advertisement, the click time, and the click frequency, and the purchase behavior data includes the purchased product information, purchase time, and purchase amount; Calculate the repeated click rate and average click interval of each type of advertising content in the historical click data to obtain the target users' acceptance of each type of advertising content; Determine the target user's consumption conversion link in different time periods based on purchase behavior data. The consumption conversion link includes the sequence of advertising content from the first click to the final purchase. Determine the key advertising content combination and optimal display timing based on the consumer conversion link; Screening out advertising content whose similarity with the key advertising content combination is greater than a preset first similarity threshold and whose click frequency is less than a preset frequency in a preset advertising library, to obtain an advertising content exploration pool; Insert the advertising content in the advertising content discovery pool into the optimal display sequence in a progressive manner to form a dynamic advertising delivery sequence; The final advertising content corresponding to the dynamic advertising delivery sequence is delivered to the target users according to the dynamic advertising delivery sequence.

[0007] By adopting the above technical solution, the historical click data and purchase behavior data of the target user are obtained, and the user's acceptance of the advertising content is calculated by combining the repeated click rate and the average click interval, so as to grasp the user's advertising preference characteristics. Based on the user's consumption conversion link, the key advertising content combination and the optimal display sequence are determined, and the advertising content sequence that most effectively triggers the user's purchasing behavior can be identified. Advertising content with high similarity but low display frequency is selected from the preset advertising library to form an exploration pool, and it is inserted into the optimal display sequence in a progressive manner. This not only maintains the effectiveness of advertising delivery, but also expands the user's advertising experience range, so that advertising delivery conforms to the user's acceptance habits, maintains the freshness of the content, and achieves continuous enhancement of delivery effects on the basis of maintaining the existing delivery effects.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, determining a key advertising content combination and an optimal display timing according to a consumption conversion link specifically includes: Count the time distribution of target users’ click behaviors triggered by various advertising contents in the consumption conversion chain to obtain time sensitivity data; Based on time sensitivity data, the conversion probability of different advertising content combinations within a preset time window is calculated. The conversion probability represents the probability that the advertising content combination triggers the target user's purchase behavior. Selecting an advertising content combination whose conversion probability is greater than a preset probability threshold as a key advertising content combination; Analyze the optimal display time interval of each advertising content in the key advertising content combination, establish display rules, and generate the optimal display timing based on the display rules.

[0009] By adopting the above technical solution, the time distribution characteristics of the user click behavior triggered by the advertising content in the consumption conversion link are analyzed, the user's time sensitivity data for different advertising content is obtained, and then the conversion probability of the advertising content combination in a specific time window is calculated. The advertising content combination with a high conversion probability is selected as the key combination, and the display rules are established by analyzing the optimal display time interval of each advertising content to generate the optimal display sequence. It can combine the user's time preference characteristics with the display rules of the advertising content to ensure that the advertising content is displayed to the user at the most suitable time point, improve the timeliness and pertinence of advertising delivery, and thus effectively improve the advertising conversion rate.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, inserting the advertisement contents in the advertisement content discovery pool into the optimal presentation timing in a progressive manner specifically includes: Calculate the content similarity between each advertisement content in the advertisement content exploration pool and each advertisement content in the key advertisement content combination; The advertising contents in the advertising content exploration pool are grouped according to content similarity to obtain similar advertising groups, each of which corresponds to an advertising content in the key advertising content combination; In each group of similar advertisements, the advertisement contents are sequentially inserted after the display timing point of the corresponding key advertisement contents in the order of content similarity from high to low.

[0011] By adopting the above technical solution, the content similarity between each advertising content in the advertising content exploration pool and the key advertising content combination is calculated, the advertising content in the exploration pool is grouped, and the advertising content in each group of similar advertisements is inserted after the display timing point of the corresponding key advertisement in the order of similarity. This progressive advertising content insertion method forms a continuous transition in content between the new advertising content and the original key advertising content, which not only maintains the overall relevance of the advertising sequence, but also can smoothly introduce the new advertising content into the user's attention range. Through the combination of similarity control and progressive insertion, the consistency of advertising delivery and user experience are improved, so that the advertising sequence can maintain a stable conversion effect and have rich content.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, after delivering the final advertisement content corresponding to the dynamic advertisement delivery sequence to the target user according to the dynamic advertisement delivery sequence, the method further includes: Collect feedback data from target users in real time, including click behavior, dwell time, and interactive operations; Calculate the user engagement score of the final advertising content based on the feedback data; The advertising content with a user engagement score lower than a preset engagement threshold is removed from the advertising content exploration pool.

[0013] By adopting the above technical solution, the user feedback data on the advertisements is collected in real time, including click behavior, dwell time and interactive operations, the user engagement score of the advertisement content is calculated, and the advertisement content below the preset engagement threshold is removed from the exploration pool, so that the advertisement delivery system can dynamically adjust the quality level of the advertisement content library and continuously optimize the advertisement content exploration pool. By timely cleaning up advertisement content with low user engagement, it is ensured that the exploration pool retains high-quality advertisement content that can effectively attract user attention and interaction, thereby improving the overall quality and effect of advertisement delivery.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, calculating the user engagement score of the final advertisement content based on the feedback data specifically includes: Assign preset weight coefficients to click behavior, dwell time and interactive operation respectively; Calculate the comprehensive engagement index based on the feedback data. The comprehensive engagement index is the weighted sum of click behavior, dwell time, interaction operation and corresponding weight coefficients; Calculate the time decay coefficient of each final advertising content based on the comprehensive engagement index; Multiply the comprehensive engagement index by the time decay coefficient to get the user engagement score.

[0015] By adopting the above technical solution and assigning different weight coefficients to click behavior, dwell time and interactive operations, the actual degree of user participation in advertising content can be more comprehensively evaluated, reducing the evaluation bias caused by a single indicator. When calculating the comprehensive engagement index, the weighted sum method can reflect the differentiated contributions of different interactive behaviors to user engagement. The introduction of the time decay coefficient takes into account the characteristics of user interests changing over time, making recent user participation behavior more valuable as a reference. The user engagement score obtained by multiplying the comprehensive engagement index by the time decay coefficient reflects both the multi-dimensional characteristics of user participation behavior and the influence of time factors, thereby being able to more accurately quantify the actual delivery effect of advertising content.

[0016] In combination with some embodiments of the first aspect, in some embodiments, after removing the advertising content with a user engagement score lower than a preset engagement threshold from the advertising content exploration pool, the method further includes: Perform cluster analysis on the remaining advertising contents in the advertising content exploration pool to obtain advertising content subsets of different subject categories; Calculate the quantity ratio of each advertising content subset and compare it with the preset balanced distribution ratio, where the preset balanced distribution ratio includes the target ratio corresponding to each advertising content subset; Supplement the advertising content whose quantity ratio is less than the corresponding target ratio from the preset advertising library.

[0017] By adopting the above technical solution, cluster analysis is performed on the remaining advertising content in the advertising content exploration pool, and the distribution of advertising content in different subject categories can be found. By calculating the proportion of each advertising content subset and comparing it with the preset balanced distribution ratio, the imbalanced distribution of advertising content can be identified. According to the comparison results of the actual proportion of advertising content in each category and the target proportion, the insufficient advertising content can be supplemented from the preset advertising library in a targeted manner, which can maintain the reasonable distribution of different subject categories in the advertising content exploration pool, maintain the diversity of advertising content, avoid the excessive concentration or serious shortage of advertising content in certain subject categories, so that advertising delivery can cover a wider range of user interests and improve the overall advertising delivery effect.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, supplementing the advertising content whose proportion is less than the corresponding target proportion from the preset advertising library specifically includes: Determine the advertising content whose quantity ratio is less than the corresponding target ratio as the advertising content to be supplemented; Calculate the difference between the proportion of the quantity of the advertising content to be supplemented and the corresponding target proportion; Matching candidate advertisement content corresponding to the advertisement content to be supplemented in a preset advertisement library; Calculating the similarity between the candidate advertisement content and the existing content in the corresponding advertisement content subset; The candidate advertising content with a similarity less than a preset second similarity threshold is selected as the supplementary content, and the supplementary content is added to the advertising content exploration pool, and the number of the supplementary content is the same as the difference.

[0019] By adopting the above technical solution, the specific difference between the proportion of the number of advertising content to be supplemented and the target proportion is calculated, and the number of advertising content that needs to be supplemented can be accurately determined. When matching candidate advertising content, by calculating the similarity between the candidate advertising content and the existing content, and selecting the candidate advertising content with a similarity less than the preset second similarity threshold as the supplementary content, it is possible to reduce duplicate content when supplementing the advertising content, ensure the difference between the supplementary content and the existing content, and by strictly controlling the number of supplements, the number of advertisements of various themes in the advertising content exploration pool can accurately achieve the expected balanced distribution, thereby improving the quality of the advertising content exploration pool.

[0020] In a second aspect, an embodiment of the present application provides an advertising delivery system based on big data analysis, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and one or more processors call the computer instructions to enable the system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions, which, when executed on a system, causes the system to execute the method described in the first aspect and any possible implementation of the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer program product, characterized in that when the computer program product runs on a system, the system executes the method described in any possible implementation manner in the first aspect.

[0023] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. This application provides an advertising delivery method based on big data analysis, which obtains the historical click data and purchase behavior data of target users, and calculates the user's acceptance of advertising content in combination with the repeated click rate and the average click interval, so as to grasp the user's advertising preference characteristics. Based on the user's consumption conversion link, the key advertising content combination and the optimal display sequence are determined, and the advertising content sequence that most effectively triggers the user's purchasing behavior can be identified. Advertising content with high similarity but low display frequency is selected from the preset advertising library to form an exploration pool, and it is inserted into the optimal display sequence in a progressive manner, which not only maintains the effectiveness of advertising delivery, but also expands the user's advertising experience range, so that advertising delivery conforms to the user's acceptance habits, maintains the freshness of the content, and achieves continuous enhancement of the delivery effect on the basis of maintaining the existing delivery effect.

[0024] 2. This application provides an advertising delivery method based on big data analysis, which collects user feedback data on advertising delivery in real time, including click behavior, dwell time and interactive operations, calculates the user engagement score of the advertising content, and removes advertising content below the preset engagement threshold from the exploration pool, so that the advertising delivery system can dynamically adjust the quality level of the advertising content library and continuously optimize the advertising content exploration pool. By timely cleaning up advertising content with low user engagement, it is ensured that the exploration pool retains high-quality advertising content that can effectively attract user attention and interaction, thereby improving the overall quality and effectiveness of advertising delivery.

[0025] 3. The present application provides an advertising delivery method based on big data analysis. By clustering the remaining advertising content in the advertising content exploration pool, the distribution of advertising content in different subject categories can be discovered. By calculating the quantity proportion of each advertising content subset and comparing it with the preset balanced distribution ratio, the imbalanced distribution of advertising content can be identified. According to the comparison results of the actual proportion of each category of advertising content with the target proportion, the insufficient advertising content can be supplemented from the preset advertising library in a targeted manner, which can maintain the reasonable distribution of different subject categories in the advertising content exploration pool, maintain the diversity of advertising content, avoid the excessive concentration or serious shortage of advertising content in certain subject categories, so that advertising delivery can cover a wider range of user interests and improve the overall advertising delivery effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a flow chart of an advertising delivery method based on big data analysis in an embodiment of the present application.

[0027] Figure 2 It is a flow chart of an advertising content optimization method based on feedback data in an embodiment of the present application.

[0028] Figure 3 It is a schematic diagram of the physical device structure of an advertising delivery system based on big data analysis provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more listed items.

[0030] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0031] The following uses an embodiment and combines Figure 1 , an advertisement delivery method based on big data analysis in an embodiment of the present application is described: See also Figure 1, which is a flow chart of an advertising delivery method based on big data analysis in an embodiment of the present application.

[0032] S101, obtaining historical click data and purchase behavior data of target users; The system obtains the target user's historical click data and purchase behavior data. The historical click data includes the clicked ad content, click time and click frequency. The purchase behavior data includes the purchased product information, purchase time and purchase amount.

[0033] The system needs to obtain the target user's historical click data and purchase behavior data as input for subsequent analysis. Among them, the historical click data includes information such as the content of the advertisements clicked by the user, the time and frequency of the clicks, etc. These data reflect the user's interest and acceptance of different types of advertisements. The purchase behavior data includes the information of the goods actually purchased by the user, the purchase time and the amount, etc. These data reflect the final effect of the advertising, that is, whether the user is triggered by the advertisement and generates a purchase behavior.

[0034] The system can obtain this data through a variety of channels, such as cooperating with advertising platforms, e-commerce platforms, etc., and connecting to their user behavior log systems; or embedding data collection modules in its own products to record users' clicks and purchase behaviors within the products. For data from different sources, the system also needs to clean and integrate them to ensure the accuracy and consistency of the data, so as to facilitate the comprehensive use of various types of data in subsequent analysis.

[0035] It should be noted that the collection of the target user's historical click data and purchase behavior data in this application is agreed upon by the user and is used in legal scenarios.

[0036] S102, calculating the repeated click rate and average click interval of each type of advertising content in the historical click data, and obtaining the target user's acceptance degree data for each type of advertising content; This step measures the user's acceptance of various types of advertising content by calculating the repeat click rate and average click interval based on the target user's historical click data. The repeat click rate reflects the proportion of users clicking on the same type of advertisement multiple times, and the average click interval reflects the time interval between users clicking on the same type of advertisement multiple times. Both indicators represent the user's persistence in this type of advertisement from different perspectives. The higher the repeat click rate and the shorter the average click interval, the more likely the user is to be interested in this type of advertising content.

[0037] In specific implementation, the system can use a big data processing engine to perform real-time or offline calculations on massive amounts of historical click data. First, the click data is aggregated according to the category of the advertising content, and then the repeat click rate and average click interval are calculated for each type of advertisement. The definition of repeat clicks can be flexibly set according to the advertising recommendation scenario, such as multiple clicks in one advertising delivery, multiple clicks in one day, or multiple clicks in a month. For the calculation of the average click interval, it is necessary to comprehensively consider the timestamps of multiple click behaviors, and divide the time difference between the earliest and latest clicks by the number of clicks minus one.

[0038] S103, determining the consumption conversion links of target users in different time periods based on the purchase behavior data; The system determines the consumption conversion link of target users in different time periods based on the purchase behavior data. The consumption conversion link includes the sequence of advertising content from the first click to the final purchase.

[0039] Purchase behavior data reflects the user's final consumption decision. Analyzing this data helps to gain insight into the user's shopping psychology and habits. This step uses the target user's purchase behavior data to determine their consumption conversion link in different time periods. The consumption conversion link refers to the entire process from the user's first exposure to the advertisement to the purchase behavior, and the exposure and influence of a series of advertising content that the user experiences. Studies have shown that users' purchase decisions are usually not made overnight, but are affected by a series of advertisements. Understanding the combination and timing of this series of advertising content is of great significance to optimizing advertising delivery strategies.

[0040] In specific implementation, the system first needs to associate the purchase behavior data with the ad click data to identify the advertising content that the user may have been exposed to before each purchase behavior. This requires considering the time difference between the purchase behavior and the click behavior, as well as the relevance of the advertising content to the purchased goods. After determining the series of advertising content corresponding to each purchase, the system can sort these advertising content by time to obtain a series of consumption conversion links. For different time periods, the system can compare the similarities and differences of these links, analyze whether there are differences in users' shopping decisions in different time periods, and explore the patterns therein.

[0041] One possible technical problem is that in actual applications, not every purchase behavior can be matched one-to-one with previous ad clicks. Purchase behavior is affected by many factors, and users may place an order a long time after seeing an ad. One solution is that the system can use machine learning methods such as probabilistic graph models to automatically learn the correlation strength and time decay rules between different advertising content and purchase behaviors from a large amount of historical data, and flexibly match behaviors that are difficult to directly associate. At the same time, business rules and expert experience can also be used to constrain and optimize the construction of conversion links.

[0042] S104, determining a key advertising content combination and an optimal display timing according to the consumption conversion link; The system determines the key advertising content combination and optimal display timing based on the consumption conversion link. Specifically, it counts the time distribution of the target user's click behavior triggered by various advertising contents in the consumption conversion link to obtain time sensitivity data. Based on time sensitivity data, the conversion probability of different advertising content combinations within a preset time window is calculated. The conversion probability represents the probability that the advertising content combination triggers the target user's purchase behavior. Selecting an advertising content combination whose conversion probability is greater than a preset probability threshold as a key advertising content combination; Analyze the optimal display time interval of each advertising content in the key advertising content combination, establish display rules, and generate the optimal display timing based on the display rules.

[0043] After determining the user's consumption conversion link, this step further analyzes these links to determine the key advertising content combination that has the greatest impact on the user's purchase decision, as well as the optimal display timing of these key contents. The so-called key advertising content combination refers to some advertising content that frequently co-appears in the consumption conversion link and has a high correlation with the final purchase behavior. The combination of these contents can effectively trigger the user's desire to buy, so these contents should be given priority in subsequent advertising. The optimal display timing refers to the existence of an optimal display order and time interval between the mined key advertising contents, so that the influence of these advertising contents can be maximized.

[0044] In order to mine key advertising content combinations, the system can use algorithms such as frequent pattern mining and association rule mining to automatically discover frequently co-occurring advertising content combinations from a large number of consumer conversion links and evaluate their association strength with purchasing behavior. For calculating the optimal display timing, it can be modeled as a sequence optimization problem, the goal is to maximize the probability that the advertising content sequence triggers purchasing behavior. This can be solved with the help of sequence models such as hidden Markov models and recursive neural networks. At the same time, it is also necessary to consider the attribute characteristics of different advertising content and the historical behavior of users to provide a personalized solution to the sequence optimization problem.

[0045] S105, selecting advertisement contents whose similarity with the key advertisement content combination is greater than a preset first similarity threshold and whose click frequency is less than a preset frequency from the preset advertisement library, and obtaining an advertisement content exploration pool; The system screens out new advertising content with exploration value in the advertising library, providing more candidates for subsequent advertising sequence optimization. Specifically, the system calculates the similarity between each advertising content in the advertising library and the known key advertising content combination, and selects content with a similarity higher than a certain threshold; at the same time, in order to avoid selecting advertisements that users have already developed aesthetic fatigue, the historical click frequency of candidate content is also restricted. Advertising content that meets these two conditions is added to the advertising content exploration pool and becomes a candidate for subsequent dynamic insertion into the advertising sequence. This method can explore some new content that is similar to known key advertisements and has potential on the one hand, and on the other hand, it can control the breadth and depth of exploration and balance development and utilization.

[0046] To calculate the similarity between advertising content, some technologies of natural language processing and computer vision can be used. For example, for text ads, algorithms such as topic models and Word2Vec can be used to calculate the semantic similarity of text; for image ads, convolutional neural networks can be used to extract high-level semantic features of images, and then calculate the similarity between features. For complex types of rich media ads, it is necessary to combine the similarity measurements of different modalities such as text, images, and videos. In addition, the system can also optimize the calculation of similarity by combining metadata information of advertising content (such as advertisers, delivery channels, audience targeting conditions, etc.).

[0047] S106, inserting the advertising content in the advertising content exploration pool into the optimal display timing in a progressive manner to form a dynamic advertising delivery sequence; The system inserts the advertising content in the advertising content exploration pool into the optimal display timing in a progressive manner to form a dynamic advertising delivery sequence. Specifically: the content similarity between each advertising content in the advertising content exploration pool and each advertising content in the key advertising content combination is calculated; The advertising contents in the advertising content exploration pool are grouped according to content similarity to obtain similar advertising groups, each of which corresponds to an advertising content in the key advertising content combination; In each group of similar advertisements, the advertisement contents are sequentially inserted after the display timing point of the corresponding key advertisement contents in the order of content similarity from high to low.

[0048] In the previous steps, the system has obtained a relatively stable sequence of key advertising content and a pool of new advertising content to be explored. The task of this step is to dynamically insert the new advertising content in the exploration pool into the existing optimal display sequence in a progressive manner, thereby forming a dynamic and personalized advertising delivery sequence. The so-called progressive insertion means that the addition of new advertising content is gradual and controllable, which not only makes use of the exploration pool to enrich the delivery content, but also ensures that the impact on the original optimal sequence structure is not too large. In this way, the overall effect of advertising delivery can be guaranteed to the greatest extent while exploring.

[0049] In specific implementation, the system must first map and associate the new advertising content in the advertising content exploration pool with the original content in the key advertising content sequence. One idea is to calculate the similarity matrix of the new and old advertising content to find the most similar insertion position of each new advertising content in the original sequence. You can also consider the contextual relevance and complementarity of the new advertising content and the original sequence, as well as the structural impact of the new advertisement on the original sequence after inserting the new advertisement. After determining the insertion position of each new advertisement, the system can control the actual insertion probability and frequency of the new advertisement to achieve progressive exploration. For example, you can set an initial insertion probability, and dynamically adjust its insertion probability as the effect of the new advertising content is observed.

[0050] A key issue in this step is how to balance the exploration of new advertising content and the protection of the original sequence structure, so as to promote the discovery of new content while preventing the destruction of the original sequence structure due to excessive insertion. One idea is that the system can establish a sequence structure evaluation model to measure the difference between the sequence structure after the insertion of the new advertisement and the original sequence. When inserting a new advertisement, the difference is strictly controlled within a certain threshold. The system can also track the performance indicators of new and old advertising content in real time. Once it is found that the new advertising content performs poorly or better than the original content, its insertion strategy is dynamically adjusted to shrink or expand the display of the new content. This is actually an online parameter adjustment and optimization process, which requires balancing the diversity of content and the stability of delivery.

[0051] S107: delivering the final advertisement content corresponding to the dynamic advertisement delivery sequence to the target user according to the dynamic advertisement delivery sequence.

[0052] After calculation and optimization in the previous steps, the system finally obtains a dynamically generated, personalized advertising sequence. The task of this step is to deliver the actual series of advertising content corresponding to the sequence to the target users according to the optimized time sequence and display strategy. This is the last link in the optimization of advertising delivery, which directly affects the actual viewing experience of users and the monetization effect of the platform. The delivery channels, timing, frequency, material style, etc., all need to be set strictly according to the dynamic delivery sequence obtained by the previous optimization.

[0053] In specific implementation, the system needs to connect with the advertising delivery platform (such as advertising alliances, DSPs, etc.), and send the specific advertising content corresponding to each ad position in the dynamic sequence to the delivery platform through the API interface. This requires that the material mapping and optimization at the advertising creative level have been completed when the dynamic sequence is generated in the previous step. The system must also reasonably control the delivery rhythm, ensuring the exposure rate of key advertising content while avoiding overly intensive advertising bombardment that may cause user disgust. The delivery time interval can be dynamically adjusted based on the time stamp and content weight of the sequence. At the same time, it is also necessary to monitor the quality of the delivery channel, eliminate some false traffic and invalid exposure, and improve the cost-effectiveness of advertising delivery.

[0054] In the above embodiment, the historical click data and purchase behavior data of the target user are obtained, and the user's acceptance of the advertising content is calculated in combination with the repeated click rate and the average click interval, so that the user's advertising preference characteristics can be grasped. Based on the user's consumption conversion link, the key advertising content combination and the optimal display sequence are determined, and the advertising content sequence that most effectively triggers the user's purchasing behavior can be identified. Advertising content with high similarity but low display frequency is selected from the preset advertising library to form an exploration pool, and it is inserted into the optimal display sequence in a progressive manner. This not only maintains the effectiveness of advertising delivery, but also expands the user's advertising experience range, so that advertising delivery conforms to the user's acceptance habits, maintains the freshness of the content, and achieves continuous enhancement of delivery effects while maintaining the existing delivery effects.

[0055] In order to achieve that the advertising delivery method based on big data analysis provided in the embodiment of the present application can continue to maintain good delivery effects in actual applications, the system needs to monitor and dynamically adjust the delivery effects in real time after completing the execution of the advertising delivery sequence. Therefore, on the basis of the above embodiments, the present application also provides an advertising content optimization method based on feedback data, which realizes continuous optimization of advertising delivery effects by establishing a dynamic evaluation and adjustment mechanism for advertising content. This optimization method and the above advertising delivery method complement each other and together constitute a complete advertising delivery optimization system. Figure 2 , a method for optimizing advertising content based on feedback data in an embodiment of the present application is described: See also Figure 2 , which is a flow chart of an advertising content optimization method based on feedback data in an embodiment of the present application.

[0056] S201, collecting feedback data of target users in real time, and calculating the user engagement score of the final advertising content based on the feedback data; The system collects feedback data from target users in real time and calculates the user engagement score of the final advertising content based on the feedback data, where the feedback data includes click behavior, dwell time and interactive operations. Specifically, the click behavior, dwell time and interactive operations are assigned preset weight coefficients respectively; Calculate the comprehensive engagement index based on the feedback data. The comprehensive engagement index is the weighted sum of click behavior, dwell time, interaction operation and corresponding weight coefficients; Calculate the time decay coefficient of each final advertising content based on the comprehensive engagement index; Multiply the comprehensive engagement index by the time decay coefficient to get the user engagement score.

[0057] In this step, the system evaluates the actual effect of the advertising content by collecting feedback data generated by target users when browsing the advertising content in real time. The feedback data mainly includes indicators such as click behavior, dwell time and interactive operations, which can comprehensively reflect the user's interest and participation in the advertising content. The system uses these feedback data as the basis for calculating the user engagement score, and obtains a score that reflects the quality and attractiveness of the advertising content by quantitatively analyzing user behavior.

[0058] In the specific implementation, the system can embed a data collection module in the advertising display page to record the user's click, stay and interaction behaviors in real time. The collected raw data can be transmitted to the background server for analysis and calculation. For different types of feedback data, the system can set corresponding weight coefficients to reflect the importance of each indicator. For example, a higher weight can be given to click behavior, while a relatively lower weight can be given to stay time and interactive operations. The comprehensive engagement index is obtained by weighted calculation of feedback data and weight coefficients. On this basis, the system can also introduce a time decay factor in combination with the display time of the advertising content to reduce the engagement score of the advertising content released earlier, and pay more attention to the impact of recent feedback data. Finally, the comprehensive engagement index is multiplied by the time decay coefficient to obtain a dynamically updated user engagement score.

[0059] S202, removing advertising content with a user engagement score lower than a preset engagement threshold from the advertising content exploration pool; In this step, the system screens and optimizes the advertising content based on the calculated user engagement score. By setting a preset engagement threshold, the system can automatically identify advertising content with low engagement and insufficient appeal, and remove it from the advertising content exploration pool. This process can achieve dynamic updating and optimization of advertising content, ensuring that high-quality, high-engagement advertising content is always retained in the exploration pool, improving the overall delivery effect.

[0060] In specific implementations, the system can set a reasonable preset engagement threshold based on historical data and business needs. The threshold can be a fixed value or a dynamically adjusted range. When the user engagement score of a certain advertising content is lower than the threshold, the system can automatically mark it as low-quality content and remove it from the exploration pool. At the same time, the system can also record relevant information of the removed content, such as content theme, creative elements, etc., for subsequent analysis and optimization.

[0061] S203, performing cluster analysis on the remaining advertising contents in the advertising content exploration pool to obtain advertising content subsets of different subject categories; In this step, the system further analyzes and processes the optimized advertising content exploration pool. Through the clustering algorithm, the system can automatically classify advertising content with similar themes or features into the same subset, forming a combination of advertising content in different themes. This process can help the system better understand and grasp the distribution characteristics of advertising content, and provide support for subsequent content optimization and recommendation.

[0062] In the specific implementation, the system can select appropriate clustering algorithms, such as K-means clustering, hierarchical clustering, etc., to automatically group advertising content. Clustering algorithms can calculate the similarity between content based on multi-dimensional information such as text features, image features, and video features of advertising content, and classify the content into different clusters based on the similarity. Each cluster represents a subset of advertising content in a subject category. The system can visualize the clustering results to intuitively analyze the distribution of content in different subject categories.

[0063] In the process of implementing this step, you may encounter problems where the clustering results are not accurate or practical enough, which will affect the subsequent content optimization effect. To solve this problem, the system can introduce manual intervention and feedback mechanisms to allow business experts to review and adjust the clustering results. Experts can reclassify certain boundary samples or outliers based on their own experience and judgment to improve the accuracy of clustering. At the same time, the system can also use expert feedback to iteratively optimize the clustering algorithm and dynamically adjust clustering parameters and strategies to make it more suitable for actual business scenarios. In addition, the system can also regularly evaluate and update clustering results, and adjust the division of topic categories in a timely manner according to changes in advertising content to ensure the timeliness and accuracy of clustering results.

[0064] S204, calculating the quantity ratio of each advertising content subset, and comparing it with the preset balanced distribution ratio; The system calculates the quantity proportion of each advertising content subset and compares it with the preset balanced distribution ratio, which includes the target proportion corresponding to each advertising content subset.

[0065] In this step, after completing the clustering of the advertising content, the system further analyzes the quantity distribution of each theme category subset. By calculating the proportion of the number of advertising content in each subset to the total, the system can fully understand whether the content distribution of different theme categories is balanced. At the same time, the system also needs to compare the actual distribution with the preset balanced distribution ratio to find possible content deviations or deficiencies.

[0066] In a specific implementation, the system can count the number of advertising content in each theme category subset based on the clustering results, and calculate its percentage of the total content. The preset balanced distribution ratio can be set according to business needs and user preferences. For example, it is required that the proportion of the number of content in each theme category is close, or that the proportion of content in some popular theme categories is slightly higher than that of other categories. The system compares the actual distribution calculated with the preset ratio and calculates the degree of difference between the two.

[0067] S205: Supplement the advertising content whose proportion is less than the corresponding target proportion from the preset advertising library.

[0068] The system supplements the advertising content whose quantity ratio is less than the corresponding target ratio from the preset advertising library. Specifically: the advertising content whose quantity ratio is less than the corresponding target ratio is determined as the advertising content to be supplemented; Calculate the difference between the proportion of the quantity of the advertising content to be supplemented and the corresponding target proportion; Matching candidate advertisement content corresponding to the advertisement content to be supplemented in a preset advertisement library; Calculating the similarity between the candidate advertisement content and the existing content in the corresponding advertisement content subset; The candidate advertising content with a similarity less than a preset second similarity threshold is selected as the supplementary content, and the supplementary content is added to the advertising content exploration pool, and the number of the supplementary content is the same as the difference.

[0069] In this step, the system selects relevant content from the preset advertising library to supplement the theme categories with uneven quantity distribution. Through content supplementation, the system can dynamically expand and optimize the advertising content exploration pool to ensure that the quantity of content in each theme category meets the preset balanced distribution requirements. At the same time, content supplementation also helps to introduce new creative elements and forms of expression, and improve the diversity and attractiveness of advertising content.

[0070] In the specific implementation, the system first needs to identify the topic categories whose quantity ratio is less than the corresponding target ratio, and calculate the size of the content gap. Then, the system can retrieve related or similar candidate content in the preset advertising library based on the characteristics of the topic category. Candidate content can be matched and sorted based on elements such as titles, keywords, and text descriptions. In order to avoid content duplication or excessive similarity, the system also needs to calculate the similarity between the candidate content and the existing content in the exploration pool, and eliminate content whose similarity exceeds the preset threshold. Finally, the system selects content with the same quantity as the size of the gap from the candidate content with a higher ranking and lower similarity, and adds it to the corresponding topic category subset.

[0071] In the above embodiment, user feedback data on the delivered advertisements, including click behavior, dwell time, and interactive operations, is collected in real time, the user engagement score of the advertisement content is calculated, and the advertisement content below the preset engagement threshold is removed from the exploration pool, so that the advertisement delivery system can dynamically adjust the quality level of the advertisement content library and continuously optimize the advertisement content exploration pool. By timely cleaning up advertisement content with low user engagement, it is ensured that the exploration pool retains high-quality advertisement content that can effectively attract user attention and interaction, thereby improving the overall quality and effect of advertisement delivery. Cluster analysis is performed on the remaining advertisement content in the advertisement content exploration pool to find out the distribution of advertisement content in different subject categories. By calculating the proportion of each advertisement content subset and comparing it with the preset balanced distribution ratio, the situation of uneven distribution of advertisement content can be identified. According to the comparison results of the actual proportion of advertisement content in each category and the target proportion, the insufficient advertisement content is supplemented from the preset advertisement library in a targeted manner, so as to maintain the reasonable distribution of different subject categories in the advertisement content exploration pool, maintain the diversity of advertisement content, avoid the situation where the advertisement content of certain subject categories is over-concentrated or seriously insufficient, so that advertisement delivery can cover a wider range of user interests and improve the overall advertisement delivery effect.

[0072] The following describes the system in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , which is a schematic diagram of the physical device structure of an advertising delivery system based on big data analysis provided in an embodiment of the present application.

[0073] It should be noted that Figure 3 The structure of the system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0074] like Figure 3As shown, the system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through a bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0075] The following components are connected to the I / O interface 305: an input section 306 including a camera, an infrared sensor, etc.; an output section 307 including a liquid crystal display (LCD) and a speaker, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0076] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0077] It should be noted that the computer-readable medium shown in the embodiment of the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may 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 with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.

[0078] The flowcharts 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 invention. Among them, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing 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 from the order 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 or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0079] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the system described in the above embodiment; or may exist independently without being assembled into the system. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of a system, the system implements the method provided in the above embodiment.

[0080] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0081] As used in the above embodiments, the term "when..." may be interpreted as "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted as "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0082] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk), etc.

[0083] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. An advertising delivery method based on big data analysis, characterized in that: include: Obtain the target user's historical click data and purchase behavior data, wherein the historical click data includes the clicked advertisement content, click time and click frequency, and the purchase behavior data includes the purchased commodity information, purchase time and purchase amount; Calculating the repeated click rate and average click interval of each type of advertisement content in the historical click data to obtain the target user's acceptance degree data for each type of advertisement content; Determine the consumption conversion link of the target user in different time periods based on the purchase behavior data, wherein the consumption conversion link includes an advertisement content sequence from the first click to the final purchase; Determine the key advertising content combination and optimal display timing according to the consumption conversion link; Screening out advertising content whose similarity with the key advertising content combination is greater than a preset first similarity threshold and whose click frequency is less than a preset frequency in a preset advertising library, to obtain an advertising content exploration pool; Inserting the advertising content in the advertising content exploration pool into the optimal display timing in a progressive manner to form a dynamic advertising delivery sequence; The final advertisement content corresponding to the dynamic advertisement delivery sequence is delivered to the target user according to the dynamic advertisement delivery sequence.

2. The method according to claim 1, characterized in that Determining the key advertising content combination and the optimal display timing according to the consumption conversion link specifically includes: Counting the time distribution of the target user's click behavior triggered by the various types of advertising content in the consumption conversion link to obtain time sensitivity data; Based on the time sensitivity data, calculating the conversion probability of different advertising content combinations within a preset time window, the conversion probability representing the probability of the advertising content combination triggering the target user's purchase behavior; Selecting the advertising content combination whose conversion probability is greater than a preset probability threshold as a key advertising content combination; Analyze the optimal display time interval of each advertisement content in the key advertisement content combination, establish a display rule, and generate an optimal display timing based on the display rule.

3. The method according to claim 1, characterized in that The inserting the advertisement contents in the advertisement content discovery pool into the optimal presentation timing in a progressive manner specifically includes: Calculating the content similarity between each advertisement content in the advertisement content exploration pool and each advertisement content in the key advertisement content combination; Grouping the advertisement contents in the advertisement content exploration pool according to the content similarity to obtain similar advertisement groups, each of the similar advertisement groups corresponding to one advertisement content in the key advertisement content combination; In each group of similar advertisements, the advertisement contents are sequentially inserted after the display timing point of the corresponding key advertisement contents in the order of the content similarity from high to low.

4. The method according to claim 1, characterized in that After delivering the final advertisement content corresponding to the dynamic advertisement delivery sequence to the target user according to the dynamic advertisement delivery sequence, the method further includes: Collecting feedback data of the target user in real time, the feedback data including click behavior, stay time and interactive operation; Calculating a user engagement score for the final advertising content based on the feedback data; The advertisement content whose user engagement score is lower than a preset engagement threshold is removed from the advertisement content exploration pool.

5. The method according to claim 4, characterized in that The calculating the user engagement score of the final advertisement content based on the feedback data specifically includes: Assigning preset weight coefficients to the click behavior, the stay duration and the interactive operation respectively; Calculating a comprehensive engagement index according to the feedback data, the comprehensive engagement index being a weighted sum of the click behavior, the dwell time, and the interactive operation and a corresponding weight coefficient; Calculating the time decay coefficient of each of the final advertising contents based on the comprehensive engagement index; The comprehensive engagement index is multiplied by the time decay coefficient to obtain a user engagement score.

6. The method according to claim 4, characterized in that After removing the advertisement content with the user engagement score lower than the preset engagement threshold from the advertisement content exploration pool, the method further includes: Performing cluster analysis on the remaining advertisement contents in the advertisement content exploration pool to obtain advertisement content subsets of different subject categories; Calculating the quantity ratio of each of the advertising content subsets, and comparing it with a preset balanced distribution ratio, wherein the preset balanced distribution ratio includes a target ratio corresponding to each of the advertising content subsets; Advertisement contents whose proportion is less than the corresponding target proportion are supplemented from the preset advertisement library.

7. The method according to claim 6, characterized in that The step of supplementing the advertising content whose proportion is less than the corresponding target proportion from the preset advertising library specifically includes: Determine the advertising content whose quantity ratio is less than the corresponding target ratio as the advertising content to be supplemented; Calculating the difference between the proportion of the quantity of the to-be-supplemented advertising content and the corresponding target proportion; Matching candidate advertisement content corresponding to the advertisement content to be supplemented in the preset advertisement library; Calculating the similarity between the candidate advertisement content and the existing content in the corresponding advertisement content subset; The candidate advertisement content whose similarity is less than a preset second similarity threshold is selected as supplementary content, and the supplementary content is added to an advertisement content exploration pool, and the number of the supplementary content is the same as the difference.

8. An advertising delivery system based on big data analysis, characterized in that: The system comprises: One or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the system to execute the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a system, the system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a system, the system is caused to execute the method according to any one of claims 1 to 7.

Images