Data pushing method and device, electronic equipment and program product

Abstract

The embodiment of the invention provides a data pushing method and device, electronic equipment and a program product. Relates to the technical field of computers. The method comprises the steps of obtaining activity data corresponding to a to-be-promoted activity, target pushing effect data and a pre-generated user set; determining a target user from the user set based on the to-be-promoted activity and the number of to-be-pushed users corresponding to the user set, pushing activity data to the target user, and counting actual pushing effect data of the activity data in a preset time period; determining a pushing user number adjustment coefficient based on the actual pushing effect data and the target pushing effect data by adopting a preset closed control algorithm; and updating the number of the to-be-pushed users according to the pushing user number adjustment coefficient, and returning to execute the operation of determining the target user from the user set based on the number of the to-be-pushed users. According to the invention, the accuracy of data pushing can be improved, and the flow utilization efficiency and the user experience are improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a data push method, apparatus, electronic device and program product. Background Technology

[0002] Current data platforms typically match user profiles with activity characteristics, and then push activity data that may be of interest to different user groups.

[0003] However, current push strategies mainly rely on manual selection and static rules to determine the target audience for data pushes, which suffers from insufficient push flexibility and accuracy. This results in users receiving too much activity data that does not match their actual interests, which not only reduces the user experience but also reduces the traffic utilization efficiency of the data platform. Summary of the Invention

[0004] This invention provides a data push method, apparatus, electronic device, and program product to improve the accuracy of data push, enhance traffic utilization efficiency, and improve user experience.

[0005] In a first aspect, the data push method provided in the embodiments of the present invention includes:

[0006] Obtain activity data, target push effect data, and pre-generated user sets corresponding to the promotional activities;

[0007] Based on the number of users to be pushed to the promotion activity and user set, target users are determined from the user set, activity data is pushed to the target users, and the actual push effect data of the activity data within the preset time period is statistically analyzed.

[0008] A preset closed-loop control algorithm is used to determine the adjustment coefficient for the number of push users based on actual push effect data and target push effect data;

[0009] The number of users to be pushed to is updated by adjusting the coefficient based on the number of users to be pushed to, and the operation of determining the target user from the user set based on the number of users to be pushed to is returned.

[0010] Secondly, the data push device provided in the embodiments of the present invention includes:

[0011] The acquisition module is used to acquire activity data, target push effect data, and pre-generated user sets corresponding to the promotion activity;

[0012] The push module is used to determine the target users from the user set based on the number of users to be pushed to the user set corresponding to the promotion activity and the user set, push the activity data to the target users, and count the actual push effect data of the activity data within a preset time period.

[0013] The determination module is used to determine the adjustment coefficient for the number of push users based on the actual push effect data and the target push effect data, using a preset closed-loop control algorithm.

[0014] The adjustment module is used to adjust the coefficient to update the number of users to be pushed based on the number of users to be pushed, and return the operation of determining the target user from the user set based on the number of users to be pushed.

[0015] Thirdly, the electronic device provided in the embodiments of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the data push method as described in any embodiment of the present invention.

[0016] Fourthly, the computer program product provided in the embodiments of the present invention includes a computer program that, when executed by a processor, implements the data push method as described in any embodiment of the present invention.

[0017] In this embodiment of the invention, by acquiring activity data, target push effect data, and a pre-generated user set corresponding to the activity to be promoted, a user group with high potential interest can be prepared for the activity promotion, clear and quantifiable push effect requirements can be set for the activity promotion, and a reference standard can be provided for subsequent traffic control. Based on the number of users to be pushed to, target users are determined from the user set, activity data is pushed to the target users, and the actual push effect data of the activity data within a preset time period is statistically analyzed. This allows for accurate evaluation of the actual user feedback of the activity to be promoted, providing a data foundation for subsequent optimization of traffic allocation. A preset closed-loop control algorithm is used to determine the adjustment coefficient for the number of users to be pushed to based on the actual push effect data and the target push effect data, and the number of users to be pushed to is updated according to the adjustment coefficient. This allows for dynamic adjustment of traffic allocation based on the real-time feedback from the user group regarding the activity to be promoted, improving push accuracy and traffic utilization while enhancing user experience. Attached Figure Description

[0018] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a data push method provided in an embodiment of the present invention;

[0020] Figure 2 This is another flowchart illustrating the data push method provided in this embodiment of the invention;

[0021] Figure 3 This is a schematic diagram of a data push device provided in an embodiment of the present invention;

[0022] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] Figure 1 This is a flowchart illustrating a data push method provided in an embodiment of the present invention. The data push method provided in this embodiment is applicable to scenarios involving pushing activity data to users. This data push method can be executed by a data push device provided in this embodiment, which can be implemented using software and / or hardware. In a specific embodiment, the device can be integrated into an electronic device, such as a computer or server. See also... Figure 1 The data push method in this embodiment may include the following steps:

[0026] Step 101: Obtain the activity data, target push effect data, and pre-generated user set corresponding to the promotion activity.

[0027] An activity refers to a project planned by a data platform to promote specific content or services to users. An activity to be promoted refers to an activity in the promotion process. Activity data refers to the collection of information displayed to users by the activity, such as activity covers, promotional text, images, audio, and video. Specifically, an activity can be initiated by a business line of the data platform and promoted to users by displaying activity data in pre-set resource positions on the platform. Activities also have attributes such as lifecycle, activity purpose, and target audience tags. Taking a music platform as an example, business lines could include music album promotion, offline concert ticket sales, and membership package marketing; resource positions could include homepage recommendations, playlist pop-ups, and artist detail page navigation bars. Activities could include "Promoting Artist A's Music Album" or "Selling Concert Tickets for Artist B." The lifecycle of the activity "Promoting Artist A's Music Album" could be 2 weeks, the activity purpose could be "listening to the album" or "purchasing the album," and the target audience tags could be "young users aged 25 to 35" or "preferring Chinese pop music."

[0028] Push performance data refers to quantifiable metrics that measure user response and business results after campaign data is pushed to users through a data platform. For example, push performance data may include one or more metrics such as Click-Through Rate (CTR), Conversion Rate (CVR), Expected Cost Per Mille (ECPM), and average user dwell time. CTR is the percentage of users who click on the campaign out of the total number of users exposed to the campaign data, quantifying user engagement. CVR is the percentage of users who achieve the campaign's objectives out of the total number of users exposed to the campaign, quantifying the campaign's business conversion rate. ECPM is the expected revenue per thousand impressions, quantifying the commercial value created by the campaign. Average user dwell time is the average length of time a user spends on the campaign page after entering it, quantifying the campaign's attractiveness and user experience.

[0029] Target push performance data refers to the expected push performance data of an activity in future stages. Target push performance data can be set differently based on the characteristics of the current promotion stage of the activity, and gradually increases as the promotion stage progresses. The promotion stage of an activity can be understood as several consecutive time intervals in the promotion process. Specifically, it can be divided according to the order of cumulative promotion time from short to long, the amount of data accumulated during the promotion process (such as cumulative clicks, cumulative revenue, etc.) from low to high, and the order of user reach from high to low.

[0030] For example, the promotion phases can be divided based on the cumulative promotion time of the promotion activity to determine the target push effect data: within 3 hours of the start of the promotion activity, the promotion activity is in the first phase, and the target push effect data can be a target CTR of 0.3%; within 3 to 6 hours of the start of the promotion activity, the promotion activity is in the second phase, and the target push effect data can be a target CTR of 0.5%; within 6 to 12 hours of the start of the promotion activity, the promotion activity is in the third phase, and the target push effect data can be a target CTR of 0.8%.

[0031] A user set is a group of users who have potential interest in a promotional campaign; it represents the candidate audience for the campaign. A pre-generated user set refers to the set of users generated before the campaign data is first pushed to users. This pre-generated user set can be manually selected by operations personnel or determined through a pre-set algorithm. For example, before a promotional campaign, active users from the data platform within the current week can be selected. Then, user tags pre-stored in the database can be matched with the target audience tags of the promotional campaign to obtain the matching degree between each user and the campaign. Finally, users with high matching degrees are selected based on a pre-set matching degree threshold to generate the user set.

[0032] Specifically, a user set corresponding to a promotional activity can be one or more, and the data push method of this embodiment of the invention is executed for each user set. For example, the number of user sets can be expanded based on the principle that "similar groups of people have similar interests." Machine learning algorithms can be used to extract user features from existing user sets, and then the similarity between the users to be expanded and the users in the existing user sets can be calculated. Users with higher similarity are selected to form a new user set, thereby ensuring that the expanded users have a high interest in the promotional activity.

[0033] Optionally, if there are multiple user sets, it is also possible to check in real time or periodically whether each user set meets the preset deletion conditions. If the actual push effect data of any user set among the multiple user sets meets the preset deletion conditions, then that user set is deleted. The preset deletion conditions include being lower than the average of the actual push effect data of each user set, being lower than half of the overall push effect data of the promotion activity, or being lower than the preset push effect data threshold.

[0034] Preset deletion conditions can be understood as rules used to determine if the push performance of any user set is too poor, requiring the deletion of that user set. Overall push performance data can be understood as the actual push performance data of the promotional activity across all user sets. Specifically, overall push performance data can be the average, median, or the best actual push performance data among all user sets corresponding to the promotional activity. Preset push performance data thresholds are the threshold values ​​for determining whether to delete a user set based on push performance data. Specifically, data platform operators can set a uniform preset push performance data threshold for each promotional activity based on historical experience, or they can set different preset push performance data thresholds for different promotional activities based on their business objectives.

[0035] Specifically, when the actual push performance data of a certain user set meets the preset deletion conditions, it means that the current actual push performance data of that user set is worse than the overall level of the various user sets corresponding to the promotion activity. Therefore, if the promotion activity is further promoted in that user set, the actual push performance data may deteriorate further, resulting in wasted traffic and user dissatisfaction.

[0036] In this embodiment, deleting the user set whose actual push effect data meets the preset deletion conditions can promptly terminate the continuous promotion of the activity to users who are not interested in the activity, concentrate traffic on users with high interest, thereby improving traffic utilization and ensuring user experience.

[0037] In this embodiment, by acquiring the activity data, target push effect data, and pre-generated user set corresponding to the activity to be promoted, it is possible to prepare a user group with high potential interest for the activity promotion, set clear and quantifiable push effect requirements for the activity promotion, and provide a reference standard for subsequent traffic control.

[0038] Step 102: Based on the number of users to be pushed to the user set corresponding to the promotion activity and user set, determine the target users from the user set, push the activity data to the target users, and count the actual push effect data of the activity data within the preset time period.

[0039] The number of users to be pushed to is the number of users in the current user set who will receive the activity data. The target users are the intended recipients of the activity data. For example, when pushing activity data to the user set for the first time, the initial number of users to be pushed to can be pre-set, or it can be determined by calculating the match rate between each user in the current user set and the activity to be promoted, obtaining the median match rate, identifying users with a match rate greater than the median match rate as target users, and determining the number of users with a match rate greater than the median match rate as the initial number of users to be pushed to. In subsequent pushes, the number of users to be pushed to is determined based on the actual push performance data and the target push performance data from the previous push. Specifically, the number of users to be pushed to can be increased if the actual push performance data is better than the target push performance data, and decreased if the actual push performance data is worse than the target push performance data.

[0040] The preset time period is the time period for acquiring and statistically analyzing the actual push effect data of the promotional activity. Actual push effect data refers to the real push effect data obtained by the data platform within the preset time period. Specifically, after the data platform adjusts the number of users to be pushed to and pushes activity data to target users, the target users will successively visit the activity homepage to achieve effective exposure of the activity data and generate interaction data. Based on the interaction data, the actual push effect data, such as the actual CTR, is further calculated. The preset time period can be extended based on the default preset time period, depending on whether the activity has reached the minimum exposure requirement. Specifically, if the activity has not reached the minimum exposure requirement, the preset time period can be extended to allow for sufficient exposure and ensure the representativeness of the collected actual push effect data. In this embodiment, target users are determined from the user set based on the number of users to be pushed to, activity data is pushed to target users, and the actual push effect data of the activity data within the preset time period is statistically analyzed. This allows for an accurate assessment of the actual user feedback of the promotional activity, providing a data foundation for subsequent optimization of traffic allocation.

[0041] Step 103: Using a preset closed-loop control algorithm, determine the adjustment coefficient for the number of push users based on the actual push effect data and the target push effect data.

[0042] The preset closed-loop control algorithm can be understood as a closed-loop control algorithm pre-designed by the data platform. It calculates feedback based on actual push effect data and target push effect data, and adjusts the number of users to be pushed to in the future according to the feedback. The push user number adjustment coefficient is a parameter output by the preset closed-loop control algorithm and is used to redetermine the number of users to be pushed to in the future.

[0043] Specifically, the preset closed-loop control algorithm uses the deviation between the target push effect data and the actual push effect data as feedback data. Using the algorithm's preset control logic, it calculates an adjustment coefficient for the number of users to be pushed to based on the feedback data, thereby adjusting the number of users to be pushed to. Then, it again statistically analyzes the actual push effect data of the adjusted promotional activity, calculates feedback data, and readjusts the number of users to be pushed to based on the feedback data, thus achieving closed-loop control of the number of users to be pushed to. The preset closed-loop control algorithm can employ fuzzy control algorithms, model predictive control algorithms, and proportional-integral-differential (PID) control algorithms, among others.

[0044] For example, the preset closed-loop control algorithm can be designed based on the PID control algorithm, and the push effect data can be used as the control coefficient of the PID control algorithm. Specifically, the difference between the actual push effect data and the target push effect data can be multiplied by the proportional gain to obtain the proportional term; the difference between the historical actual push effect data and the historical target push effect data corresponding to several previous activity data pushes can be accumulated and then multiplied by the integral gain to obtain the integral term; the difference between the actual push effect data and the target push effect data can be subtracted from the difference between the actual push effect data and the target push effect data corresponding to the previous activity data push and then multiplied by the derivative gain to obtain the derivative term; finally, the proportional term, integral term and derivative term are added together to obtain the adjustment coefficient for the number of push users.

[0045] Step 104: Adjust the coefficients based on the number of users to be pushed to update the number of users to be pushed to, and return to execute step 102.

[0046] Specifically, the number of users to be pushed to can be updated by combining the adjustment coefficient for the number of push users and the number of users in the user set. For example, the number of users to be pushed to, P, can be updated according to the following formula: P = N + K × N, where N is the current number of users to be pushed to, and K is the adjustment coefficient for the number of push users. Assuming that the current number of users to be pushed to is 2 million, and the adjustment coefficient for the number of push users is -2%, then the updated number of users to be pushed to = 2 million + (-2%) × 2 million = 1.96 million.

[0047] Specifically, if the actual push performance data is better than the target push performance data, it indicates that the target users have a high interest in the push campaign, so the number of users to be pushed to can be increased. If the actual push performance data is worse than the target push performance data, it indicates that there may be insufficient user interest in the campaign, and the number of users to be pushed to should be reduced to avoid wasting traffic resources and degrading user experience. For example, if there are three campaigns A, B, and C to be promoted, and the initial number of users to be pushed to is 1 million for each, after campaigns A, B, and C are pushed to the target users, if the target CTR of campaign A equals the actual CTR, the number of users to be pushed to remains unchanged; if the target CTR of campaign B is greater than the actual CTR, the number of users to be pushed to is increased to 120; if the target CTR of campaign C is less than the actual CTR, the number of users to be pushed to is reduced to 80.

[0048] In this embodiment, a preset closed-loop control algorithm is adopted to determine the adjustment coefficient of the number of push users based on the actual push effect data and the target push effect data, and to update the number of users to be pushed according to the adjustment coefficient of the number of push users. It can dynamically adjust the traffic allocation according to the real-time feedback of the user group on the promotion activity, thereby improving the user experience while improving the accuracy of push and the traffic utilization rate.

[0049] In this embodiment, by acquiring activity data, target push effect data, and a pre-generated user set corresponding to the promotion activity, a user group with high potential interest can be prepared for the promotion activity, and clear and quantifiable push effect requirements can be set for the promotion activity, providing a reference standard for subsequent traffic control. Based on the number of users to be pushed to, target users are determined from the user set, and activity data is pushed to the target users. The actual push effect data of the activity data within a preset time period is statistically analyzed, enabling accurate evaluation of the actual user feedback for the promotion activity and providing a data foundation for subsequent traffic allocation optimization. A preset closed-loop control algorithm is used to determine the adjustment coefficient for the number of push users based on the actual push effect data and the target push effect data. The number of users to be pushed is updated according to the adjustment coefficient, allowing for dynamic adjustment of traffic allocation based on the real-time feedback from the user group regarding the promotion activity. This improves push accuracy and traffic utilization while enhancing user experience.

[0050] Figure 2 This is another flowchart illustrating the data push method provided in this embodiment of the invention, such as... Figure 2 As shown, the specific steps include the following:

[0051] Step 201: Obtain the activity data and pre-generated user set corresponding to the activity to be promoted.

[0052] Optionally, at least one of the following checks can be performed on the promotional activity: campaign type check, online status check, skip list check, and page placement check. The campaign type check is used to check whether the promotional type of the activity to be promoted belongs to the specified promotional type; the online status check is used to check the online status of the activity to be promoted; the skip list check is used to check whether the activity to be promoted belongs to the preset skip list; and the page placement check is used to check whether the display position of the activity data of the activity to be promoted on the page belongs to the specified position.

[0053] The designated promotion type refers to the preset activity type that is applicable or not applicable to the data push method provided by this invention. Specifically, there are several targeted push activities in the data platform, such as activities that only allow members to participate. Whether to push such activities to users depends on the user's identity rather than the user's interests, and is not applicable to the data push method provided by this invention. Therefore, it is necessary to check the delivery type to avoid pushing irrelevant activities to users. Online status refers to whether the activity to be promoted is within its effective lifecycle. Specifically, online status usually includes "online" and "offline", representing whether the activity allows users to participate. Checking the online status of activities can avoid pushing activities that have ended to users. The preset skipped activity list refers to activities that are not applicable to the data push method provided by this invention. Specifically, the preset skipped activity list can be dynamically set by the data platform's operators to accurately respond to various sudden or special business situations, such as the display materials of a certain activity to be promoted having defects and needing to be temporarily taken offline for repair, or the activity data of a prize-winning activity only being pushed to users who participated in the lottery and won the prize. The display position of activity data on the page refers to the resource position used to display activity data, such as the homepage recommendation position, the pop-up window on the playlist page, the navigation bar on the artist details page, etc. The display position of activity data on the page can be specified by the data platform's operators. Specifically, the page placement check can be used to check whether the display position of the activity data of the activity to be promoted on the page is suitable or not suitable for the data push method provided by this invention.

[0054] Optionally, it can also be checked whether the user set is a defined user set. If so, the user set is not applicable to the data push method provided in this embodiment of the invention.

[0055] The selected user set refers to the set of users manually selected by the data platform's operations personnel when configuring a promotional campaign. Specifically, the users included in the selected user set are determined based on specific business rules and usually do not require dynamic adjustment.

[0056] In this embodiment, by checking the type of the promotion activity, online status, skip list, and page placement, and by checking whether the user set is the defined user set, it can ensure that the promotion activity is promoted within the scope permitted by the business, block invalid pushes that may harm the user experience or violate business rules, avoid the consumption of computing resources, and ensure the accuracy of data push and the controllability of push effect.

[0057] Optionally, the users in the user set are sorted in descending order of their predicted interest in the promotional activity; the predicted interest is obtained by inputting the feature information of the corresponding user and the feature information of the promotional activity into a preset neural network model and then using the output information of the preset neural network model.

[0058] Interest prediction values ​​quantify a specific user's interest in a particular promotional activity. Interest prediction values ​​can be percentages between 0 and 1. For example, user characteristics may include gender, age, region, preferred song genres, etc., while the promotional activity's characteristics may include target audience tags, activity creative types, and the activity data itself, preset by the data platform's operators when configuring the promotion. The preset neural network model is a deep learning model that predicts the user's interest in the promotional activity based on the input user and promotional activity characteristics. For example, the training data for the preset neural network model can be extracted from the data platform logs, including historical activity information, user characteristic information corresponding to users pushed to historical activity data, and whether users clicked on historical activities. Inputting user and promotional activity characteristics into the preset neural network model, the output information can be the probability of the user clicking on the activity; this output information can be directly used as the interest prediction value.

[0059] In this embodiment, by sorting users according to their predicted interest in the promotional activities, it can be ensured that the promotional activities are prioritized for potential high-interest user groups, thereby guaranteeing the effectiveness of the push.

[0060] Step 202: Determine the current promotion stage based on the current total clicks of the promotion activity.

[0061] The current total clicks are the sum of clicks from all user groups corresponding to the promotion activity. The current promotion stage is the specific time interval in which the current activity is located within its entire promotion process. Specifically, promotion stages are divided in ascending order of the total clicks of the activity. For example, if the total clicks of the activity are less than or equal to 10,000, the current promotion stage can be determined as the first promotion stage; if the total clicks are greater than 10,000 but less than or equal to 20,000, the current promotion stage can be determined as the second promotion stage; and if the total clicks are greater than 20,000 but less than or equal to 30,000, the current promotion stage can be determined as the third promotion stage.

[0062] Step 203: Based on the pre-set correspondence between promotion stages and target push effect data, determine the target push effect data corresponding to the current promotion stage; wherein, the target push effect data corresponding to the later promotion stage in time is greater than the target push effect data corresponding to the earlier promotion stage in time.

[0063] Specifically, the correspondence between the promotion phase and the target push effect data can be pre-set by the operations personnel. In the cold start phase of the campaign, before any user data is displayed, the target push effect data can be set very low, and the data is shown to users with the highest potential interest in the campaign. This results in a significantly better actual push effect than the target, allowing the pre-set closed-loop control algorithm to quickly increase the number of users to be pushed to, expanding the campaign's user reach and validating the target users' interest in the campaign and its business effectiveness. In subsequent promotion phases, a larger number of users to be pushed to leads to faster growth in user reach and total clicks, allowing the campaign to enter the next promotion phase more quickly, thus increasing the target push effect data. In later promotion phases, when user reach has reached a certain level and the user group with the highest potential interest has been reached, the remaining users have relatively lower interest. In this case, the target push effect data can be set to a higher value and kept constant, ensuring a smooth push to the remaining user group while maintaining a good user experience. Furthermore, the number of users to be pushed to can be reduced when the push effect declines to avoid user backlash. For example, if the promotion phases of the promotion activity are divided into the first promotion phase, the second promotion phase, and the third promotion phase in ascending order of total clicks, then the target push effect data for the first promotion phase can be set to CTR=0.3%, the target push effect data for the second promotion phase can be set to CTR=0.6%, and the target push effect data for the third promotion phase can be set to CTR=0.8%.

[0064] Optional: The target push performance data is the target CTR, and the actual push performance data is the actual CTR; or, the target push performance data is the target ECPM, and the actual push performance data is the actual ECPM.

[0065] Specifically, CTR is the percentage of users who click on an activity out of the total number of users who are exposed to the activity, quantifying user engagement. The formula for calculating CTR is: CTR = (Total Clicks / Total Impressions) × 100%. ECPM is the expected revenue per thousand impressions, quantifying the commercial value created by the activity. The formula for calculating ECPM is: ECPM = (Activity Revenue / Total Impressions) × 1000.

[0066] In this embodiment, the target push effect data corresponding to the current promotion stage is determined based on the correspondence between the promotion stage and the target push effect data. This enables the adjustment strategy for the number of users to be pushed to in the future to be adapted to the differentiated target of the current promotion stage, avoiding the insufficient adaptability of a single strategy throughout the entire promotion stage, thereby improving the accuracy of traffic control.

[0067] Step 204: Determine the first user in the user set whose position number corresponds to the number of users to be pushed to, and determine the first user and all users before the first user as target users, and push the activity data to the target users.

[0068] The first user refers to a specific user within the user set corresponding to the number of users to be pushed to. Specifically, users in the user set are sorted in descending order based on their predicted interest in the promotional activity. The smaller the position number in the user set, the higher the probability that the user is interested in the promotional activity. The users whose position numbers are the first user and those preceding the first user represent the group with the highest potential interest in the promotional activity within the current user set, and therefore can be identified as target users. For example, assuming a user set initially contains 5 million users, and their position numbers are sorted in descending order of their predicted interest in the promotional activity, if the number of users to be pushed to is 1 million, then users with position numbers greater than or equal to 1 and less than or equal to 1 million are identified as target users. When a user requests access to the platform homepage, it can be determined whether the user's predicted interest in the promotional activity is greater than the predicted interest value corresponding to the first user. If it is greater, the user belongs to the target user group, and the promotional activity can be exposed to that user.

[0069] Step 205: Determine the number of unique visitors corresponding to the user set, and determine whether the preset data volume insufficient condition is met based on the number of unique visitors.

[0070] Unique visitors refer to the total number of users who visit the page where the event is held. Specifically, if the same user visits the event page multiple times, the unique visitor count will still be 1. Unique visitors represent the number of users who have been effectively exposed to the event promotion.

[0071] Optionally, insufficient data conditions include at least one of the following:

[0072] The number of unique visitors is less than a preset threshold, which is determined based on the current size of the user set and the minimum exposure percentage threshold.

[0073] The number of unique visitors is less than the fixed minimum exposure threshold;

[0074] The current size of the user set is greater than the preset size threshold.

[0075] Specifically, in step 204, after pushing the activity data to the target users, it is necessary to statistically analyze the actual push effect data within a preset time period to evaluate the push effect. The insufficient data condition can be understood as the activity not being sufficiently exposed to the target users within the default preset time period, potentially leading to insufficient statistical data on the actual push effect. Since unique visitors represent the number of users effectively exposed to the activity data, we can first statistically analyze the number of unique visitors within the default preset time period to preliminarily determine whether the activity data was exposed to a sufficient number of users within that period, and thus determine whether the actual promotion effect data within the default preset time period is representative.

[0076] The current size of the user set refers to the total number of users currently included in the user set. Exposure percentage refers to the proportion of users in the user set exposed to the activity data out of the total number of users currently included in the user set. The minimum exposure percentage threshold is the lowest percentage of users in the user set that should be exposed to the activity data. The preset threshold can be understood as the minimum number of unique visitors determined based on the current size of the user set and the minimum exposure percentage threshold, which can meet the minimum exposure rate requirement of the promotional activity in the user set.

[0077] For example, the preset threshold can be calculated using the following formula: Preset threshold = Current size of the user set × Minimum exposure percentage threshold. If the current size of the user set is 1 million and the minimum exposure percentage threshold is 1%, then the preset threshold is 1 million × 1% = 10,000. The fixed minimum exposure threshold refers to the minimum number of unique visitors required for the user set, such as 200,000. The fixed minimum exposure threshold is a preset, fixed threshold and is independent of the size of the user set. The preset size threshold can be understood as a critical value for the number of users used to determine whether the user set is too large, such as 2.5 million.

[0078] Specifically, if the number of unique visitors is less than a preset threshold, it indicates that the proportion of users exposed to the current push activity in the user set is insufficient; if the number of unique visitors is less than a fixed minimum exposure threshold, it indicates that the absolute number of users exposed to the current push activity in the user set is insufficient. Therefore, the statistical time for the actual push effect can be delayed to collect more user feedback data and improve the representativeness of the actual push effect data; if the current size of the user set is greater than a preset size threshold, it indicates that the time required for the activity data to be exposed to all users is long. Therefore, the statistical time for the actual push effect data can also be delayed to wait for the activity data to be fully exposed to the target users.

[0079] In this embodiment, the determination of whether the preset data shortage condition is met based on the number of unique visitors can accurately determine whether the pushed activity data has been sufficiently exposed to the target users, thereby delaying the statistical time of the actual push effect data in the event of insufficient data.

[0080] Step 206: If it is determined that the data volume is insufficient, then query the database for the historical window offset corresponding to the promotion activity and user set.

[0081] Specifically, in order to obtain sufficiently representative actual push effect data, if the user set meets the condition of insufficient data volume, it is necessary to extend the statistical time of actual push effect data, that is, extend the preset time period.

[0082] The statistical time window is a sliding time window used to define a preset time period. The start point of the statistical time window is the time when the actual push effect data begins to be statistically analyzed, and the end point is the time when the actual push effect data ends to be statistically analyzed. The default start point of the statistical time window can be the time when activity data is pushed to the target user plus a preset time length, such as 30 minutes. The end point of the statistical time window can be delayed by a fixed time length relative to the start point, such as 10 minutes, to allow for the reporting of target user interaction data and the inclusion of new interaction data generated during the calculation of the actual push effect data. Extending the preset time period can be achieved by sliding the default statistical time window's end point backward by a certain step.

[0083] For example, assuming that activity data is pushed to target users in the user set starting at 8:00 AM, the default start point of the statistical time window can be 8:00 plus 30 minutes, i.e., 8:30, and the end point of the statistical time window can be 8:30 plus 10 minutes, i.e., 8:40. Therefore, the current statistical time window is a 10-minute time window between 8:30 and 8:40, with the preset time period being 8:30 to 8:40.

[0084] The window offset can be understood as the duration by which the default statistical time window slides backward from its end point. The historical window offset is the window offset used when the preset time period was last determined. Specifically, if the user set of the promotion activity meets the insufficient data condition, the historical window offset corresponding to the promotion activity and user set is queried in the database; if the insufficient data condition is not met, the time period corresponding to the default statistical time window is used as the preset time period.

[0085] Step 207: If the historical window offset is found, determine the current window offset based on the historical window offset and the preset time step, and update the historical window offset based on the current window offset.

[0086] The preset time step refers to the unit of time used to extend the statistical time window each time the insufficient data condition is met. Specifically, if a historical window offset is found, it indicates that the promotional activity and user set have not met the insufficient data condition for the first time. Therefore, the preset time step should be added to the historical window offset to obtain the current window offset. After determining the current window offset, it can be written to the database as the historical window offset for the next statistical analysis of actual push performance data.

[0087] For example, at 8:00 AM, the number of users to be pushed was adjusted. The default statistical time window is from 8:30 AM to 8:40 AM. However, within the default statistical time window, the number of unique visitors meets the condition of insufficient data volume. At this time, it is necessary to query the historical window offset. Assuming that the preset time step is 5 minutes, the historical window offset found is 10 minutes. Then, add 5 minutes to 10 minutes to determine the current window offset as 15 minutes, and update the historical data offset in the database to 15 minutes.

[0088] Step 208: If no historical window offset is found, determine the current window offset based on the preset time step and write the current window offset to the database.

[0089] Specifically, if no historical window offset is found, it means that the promotion activity and user set meet the condition of insufficient data volume for the first time. At this time, the preset time step can be determined as the current window offset.

[0090] Step 209: Perform a time extension operation on the default statistical time window based on the current window offset, and use the time period corresponding to the extended statistical time window as the preset time period.

[0091] Specifically, the time extension operation refers to sliding the end point of the default statistical time window backward by the current window offset by the corresponding time length. For example, if the current window offset is 15 minutes and the default statistical time window is from 8:30 to 8:40, then the statistical time window after performing the time extension operation will be from 8:30 to 8:55.

[0092] In this embodiment, the default statistical time window is extended based on the current window offset. The time period corresponding to the extended statistical time window is used as the preset time period. This ensures that the actual push effect data is obtained based on a sufficient amount of data, thus improving the reliability and accuracy of the actual push effect data.

[0093] Step 210: Calculate the actual push effect data of the activity data within the preset time period.

[0094] Step 211: Using a preset closed-loop control algorithm, determine the adjustment coefficient for the number of push users based on the actual push effect data and the target push effect data.

[0095] Optionally, the preset closed-loop control algorithm is a PID control algorithm. Using this algorithm, the number of users to be pushed is adjusted based on actual push performance data and target push performance data, including operations for Sa1-Sa4.

[0096] Sa1. Determine the current deviation value based on the difference between the actual push effect data and the target push effect data, and determine the proportion data based on the current deviation value and the preset proportion coefficient.

[0097] Specifically, the current deviation value can be obtained by subtracting the target push performance data from the actual push performance data. This deviation value is then multiplied by a preset scaling factor to obtain the scaling factor. The larger the current deviation value, the larger the scaling factor, enabling rapid response to fluctuations in push performance.

[0098] Sa2. Determine the current cumulative deviation value based on the current deviation value and the previously obtained cumulative deviation value, and determine the integral term data based on the current cumulative deviation value and the preset integral coefficient.

[0099] The cumulative deviation value is obtained by summing historical deviation values ​​calculated after each promotion. Specifically, the current cumulative deviation value is obtained by adding the previous cumulative deviation value to the current deviation value, and then multiplying the current cumulative deviation value by a preset integration coefficient to obtain the integral term data. The larger the cumulative deviation value, the greater the long-term error between the target push effect data and the actual push effect data. Calculating the integral term data can ensure that the actual push effect is stable at the target push effect.

[0100] Sa3. Determine the current deviation rate of change based on the current deviation value and the previous deviation value, and determine the differential term data based on the current deviation rate of change and the preset differential coefficient.

[0101] Specifically, the current deviation value can be subtracted from the previous deviation value to obtain the current deviation change rate. Multiplying this rate by a preset differential coefficient yields the differential term data. The current deviation change rate measures the rate of change of the deviation between the target push effect data and the actual push effect data. Calculating the differential term data can prevent drastic fluctuations in the number of target users.

[0102] In addition, each combination of promotion activity, user set, and current promotion stage independently stores the cumulative deviation value and deviation change rate value corresponding to the last push. When adjusting the number of users to be pushed next time, the integral term data and differential term data are calculated based on the stored cumulative deviation value and deviation change rate value.

[0103] For example, suppose the push performance data is CTR. One promotional campaign has already been conducted, with a target CTR of 5% and an actual CTR of 3%. The target CTR for the current batch is 5%, and after the promotion ends, the actual CTR is 4.5%. Therefore, the current deviation value e = 5% - 4.5% = 0.5%. The formula for calculating the proportional data p is p = Kp × e; where Kp is the proportional coefficient, and Kp = 2. Therefore, p = 2 × 0.5% = 1%. The formula for calculating the integral data is i = Ki × (e' + e); where e' is the accumulated deviation value obtained from the previous campaign, i.e., e' = 5% - 3% = 2%, e' + e is the current accumulated deviation value, Ki is the integral data, and Ki = 0.5. Therefore, i = 0.5 × (2% + 0.5%) = 1.25%. The formula for calculating the differential term data d is d = Ki × (e - ep); where ep is the deviation value obtained in the previous time, so ep = 2%, e - ep is the current deviation change rate value, Kd is the differential term coefficient and Kd = 1, so d = 1 × (0.5% - (-2%)) = -1.5%.

[0104] Sa4. Determine the adjustment coefficient for the number of push users based on the proportional, integral, and differential data.

[0105] Specifically, the adjustment coefficient for the number of push users can be obtained by summing the proportional, integral, and differential data. Continuing the previous example, the current adjustment coefficient for the number of push users can be p + i + d = 1% + 1.25% - 1.5% = 0.75%.

[0106] In this embodiment, a PID control algorithm is used to determine the adjustment coefficient for the number of push users based on the actual push effect data and the target push effect data. This can reduce the deviation between the actual and target push effects and improve the accuracy and stability of traffic control.

[0107] Optionally, the methods for determining the preset proportional coefficient, preset integral coefficient, and preset differential coefficient include operations for Sb1-Sb2:

[0108] Sb1. Determine the current promotion stage based on the current total clicks corresponding to the promotion activity.

[0109] Sb2. Based on the pre-set correspondence between promotion stages and PID parameters, determine the PID parameters corresponding to the current promotion stage. The PID parameters include preset proportional coefficient, preset integral coefficient, and preset derivative coefficient. The PID parameters corresponding to later promotion stages are smaller than those corresponding to earlier promotion stages.

[0110] Specifically, the reason why the PID parameter for later promotion phases is smaller than that for earlier promotion phases is as follows: In earlier promotion phases, rapid exploration and trial and error are needed to find the optimal push strategy for the current user set. Therefore, a larger PID parameter is required so that the preset PID control algorithm can quickly adjust the user number adjustment coefficient based on the deviation between the target push effect data and the actual push effect data. In later promotion phases, the user coverage of the promotion activity has already reached a high level, and the promotion goal is to stabilize the actual push effect data near the target push effect data. On the other hand, most users with potential high interest in the promotion activity have already been reached. Continuing to promote the activity to users with relatively low potential interest will increase the risk of negative user feedback. Therefore, the PID parameter can be reduced to decrease the sensitivity of the user number adjustment coefficient to the deviation between the target and actual push effect data. This avoids blindly and significantly increasing the user number adjustment coefficient because the actual push effect data is lower than the target push effect data, which could lead to a decline in user experience.

[0111] In this embodiment, the PID parameters corresponding to the current promotion stage are determined according to the pre-set correspondence between the promotion stage and the PID parameters. This enables the preset closed-loop control algorithm to automatically adapt to the targets of different promotion stages and adaptively adjust the number of users to be pushed, thereby improving the flexibility of traffic control.

[0112] Step 212: Perform at least one of the following operations: boundary check, frequency check, and adjustment rationality check, and trigger step 213 if all checks pass;

[0113] Among them: boundary check is used to check whether the number of users to be pushed, determined by the adjustment coefficient of the number of push users, is within the preset boundary range; frequency check is used to check whether the time length between the current time and the last time the number of users to be pushed was updated meets the preset time interval requirement; adjustment rationality check is used to check whether the adjustment direction corresponding to the adjustment coefficient of the number of push users is consistent with the expected adjustment direction, and the expected adjustment direction is consistent with the change direction of the actual push effect data.

[0114] Specifically, if any of the boundary checks, frequency checks, or adjustment rationality checks fails, the operation of updating the number of users to be pushed based on the adjustment coefficients will not be performed. In this case, the parameter values ​​of the preset closed-loop control algorithm, such as the PID algorithm, need to be restored to their initial values.

[0115] The preset boundary range can include an upper and / or lower bound on the number of users to be pushed to. For example, if the upper bound of the preset boundary range is 2 million and the lower bound is 40 million, then when the number of users to be pushed to is less than 2 million or greater than 40 million, the number of users to be pushed to will not be updated.

[0116] The preset time interval requirement can be the minimum time interval between the current time and the time of the last update on the number of users to be pushed, such as 5 minutes.

[0117] The expected adjustment direction can be understood as the adjustment direction of the number of users to be pushed to, aiming to make the actual push effect data closer to the target push effect data. Specifically, when the actual push effect data is greater than the target push effect data, the number of users to be pushed to is increased; when the actual push effect data is less than the target push effect data, the number of users to be pushed to is decreased. The adjustment direction corresponding to the push user number adjustment coefficient can be understood as the adjustment direction of the push user number calculated by a preset closed-loop control algorithm. Specifically, if the push user number adjustment coefficient is positive, the adjustment direction corresponding to this coefficient is to increase the number of users to be pushed to; if the push user number adjustment coefficient is negative, the adjustment direction corresponding to this coefficient is to decrease the number of users to be pushed to.

[0118] In this embodiment, performing boundary checks can prevent the number of users to be pushed to from being too small or too large, ensuring the stability and controllability of the traffic of the promotion activity; performing frequency checks can prevent the number of users to be pushed to from being repeatedly adjusted in a short period of time, causing traffic fluctuations in the promotion activity; performing adjustment rationality checks can prevent unreasonable adjustments to the number of users to be pushed to from being performed, ensuring the accuracy of traffic control.

[0119] Step 213: Adjust the coefficients based on the number of users to be pushed to update the number of users to be pushed to, and return to execute step 202.

[0120] In this embodiment, by acquiring activity data, target push effect data, and a pre-generated user set corresponding to the promotion activity, a user group with high potential interest can be prepared for the promotion activity, and clear and quantifiable push effect requirements can be set for the promotion activity, providing a reference standard for subsequent traffic control. Sorting users according to their predicted interest in the promotion activity ensures that the promotion activity is prioritized for push to potential high-interest user groups, thereby guaranteeing the push effect. Determining the target push effect data corresponding to the current promotion stage based on the correspondence between the promotion stage and the target push effect data allows the subsequent adjustment strategy for the number of users to be pushed to to be adapted to the differentiated goals of the current promotion stage, avoiding a single strategy. The system addresses the lack of adaptability throughout the promotion phase, thereby improving the accuracy of traffic control. It identifies target users from the user set based on the number of users to be pushed to, pushes activity data to these users, and statistically analyzes the actual push effect data within a preset time period. This allows for accurate evaluation of actual user feedback on the promotional activity, providing a data foundation for subsequent traffic allocation optimization. It determines whether the preset data shortage condition is met based on the number of unique visitors, accurately judging whether the pushed activity data has received sufficient exposure to the target users, thus allowing for a delay in the statistical time of actual push effect data in cases of insufficient data. Finally, it extends the default statistical time window based on the current window offset, thus delaying the... The extended statistical time window, serving as a preset time period, ensures that the actual push performance data is based on sufficient statistical data, improving the reliability and accuracy of the actual push performance data. A preset closed-loop control algorithm is employed to determine the adjustment coefficient for the number of push users based on the actual and target push performance data, and updates the number of users to be pushed to according to this adjustment coefficient. This allows for dynamic adjustment of traffic allocation based on real-time feedback from user groups regarding the promotional activity, improving both traffic utilization and user experience. A PID control algorithm is also used to determine the adjustment coefficient for the number of push users based on the actual and target push performance data, reducing the impact of discrepancies between the actual and target push performance. To improve the accuracy and stability of traffic control, the algorithm addresses deviations in performance. By determining the PID parameters corresponding to the current promotion stage based on a pre-set relationship between the promotion stage and the PID parameters, the pre-defined closed-loop control algorithm automatically adapts to the targets of different promotion stages, adaptively adjusting the number of users to be pushed to, thus improving the flexibility of traffic control. Boundary checks prevent the number of users to be pushed from being too small or too large, ensuring the stability and controllability of the traffic for the promotional activity. Frequency checks prevent repeated adjustments to the number of users to be pushed within a short period, avoiding traffic fluctuations. Adjustment rationality checks prevent unreasonable adjustments to the number of users to be pushed, ensuring the accuracy of traffic control.

[0121] Figure 3This is a schematic diagram of a data push device provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the device includes:

[0122] The acquisition module 301 is used to acquire the activity data, target push effect data and pre-generated user set corresponding to the promotion activity;

[0123] The push module 302 is used to determine the target users from the user set based on the number of users to be pushed to the user set corresponding to the promotion activity and the user set, push the activity data to the target users, and count the actual push effect data of the activity data within a preset time period.

[0124] The determination module 303 is used to determine the adjustment coefficient for the number of push users based on the actual push effect data and the target push effect data using a preset closed control algorithm;

[0125] The adjustment module 304 is used to adjust the coefficient to update the number of users to be pushed based on the number of users to be pushed, and return to perform the operation of determining the target user from the user set based on the number of users to be pushed.

[0126] In one embodiment, the preset closed-loop control algorithm is a PID control algorithm. The determining module 303 uses the preset closed-loop control algorithm to determine the adjustment coefficient for the number of push users based on the actual push effect data and the target push effect data, including:

[0127] The current deviation value is determined based on the difference between the actual push effect data and the target push effect data, and the proportion data is determined based on the current deviation value and the preset proportion coefficient.

[0128] The current accumulated deviation value is determined based on the current deviation value and the previously obtained integral term data; the integral term data is determined based on the current accumulated deviation value and the preset integral coefficient.

[0129] The current deviation rate of change is determined based on the current deviation value and the previous deviation value, and the differential term data is determined based on the current deviation rate of change and the preset differential coefficient.

[0130] The adjustment coefficient for the number of users pushed to the platform is determined based on the proportional, integral, and differential data.

[0131] In one embodiment, the method for determining the preset proportional coefficient, preset integral coefficient, and preset differential coefficient includes:

[0132] The current promotion stage is determined based on the current total number of clicks for the activity to be promoted.

[0133] Based on the pre-set correspondence between promotion stages and PID parameters, determine the PID parameters corresponding to the current promotion stage. The PID parameters include the preset proportional coefficient, preset integral coefficient, and preset derivative coefficient.

[0134] The PID parameters for later promotion phases are smaller than those for earlier promotion phases.

[0135] In one embodiment, the method for determining target push effect data includes:

[0136] The current promotion stage is determined based on the current total clicks of the campaign to be promoted.

[0137] Based on the pre-set correspondence between promotion stages and target push performance data, determine the target push performance data corresponding to the current promotion stage;

[0138] Among them, the target push effect data corresponding to the later promotion stage is greater than the target push effect data corresponding to the earlier promotion stage.

[0139] In one embodiment, the users in the user set are sorted in descending order of their predicted interest in the promotional activity; the predicted interest value is obtained by inputting the feature information of the corresponding user and the feature information of the promotional activity into a preset neural network model, and based on the output information of the preset neural network model. The push module 302 determines the target users from the user set based on the number of users to be pushed to corresponding to the promotional activity and the user set, including:

[0140] Identify the first user in the user set whose position number corresponds to the number of users to be pushed to, and then identify the first user and all users before the first user as target users.

[0141] In one embodiment, the method for determining the preset time period includes:

[0142] Determine the number of unique visitors corresponding to the user set, and determine whether the preset data volume insufficient condition is met based on the number of unique visitors;

[0143] If the insufficient data condition is met, the current window offset is determined, and the default statistical time window is extended based on the current window offset. The time period corresponding to the extended statistical time window is then used as the preset time period.

[0144] In one embodiment, determining the current window offset includes:

[0145] Query the database for the historical window offsets corresponding to the promotional activity and user set;

[0146] If a historical window offset is found, the current window offset is determined based on the historical window offset and the preset time step, and the historical window offset is updated based on the current window offset.

[0147] If no historical window offset is found, the current window offset is determined based on a preset time step and written to the database.

[0148] In one embodiment, the insufficient data condition includes at least one of the following:

[0149] The number of unique visitors is less than a preset threshold, which is determined based on the current size of the user set and the minimum exposure percentage threshold.

[0150] The number of unique visitors is less than the fixed minimum exposure threshold;

[0151] The current size of the user set is greater than the preset size threshold.

[0152] In one embodiment, the device further includes a checking module, which is used to perform at least one of boundary checks, frequency checks and adjustment rationality checks before updating the number of users to be pushed based on the adjustment coefficient according to the number of push users, and triggers the operation of updating the number of users to be pushed based on the adjustment coefficient according to the number of push users if all checks pass.

[0153] Among them: boundary check is used to check whether the number of users to be pushed, determined by the adjustment coefficient of the number of push users, is within the preset boundary range; frequency check is used to check whether the time length between the current time and the last time the number of users to be pushed was updated meets the preset time interval requirement; adjustment rationality check is used to check whether the adjustment direction corresponding to the adjustment coefficient of the number of push users is consistent with the expected adjustment direction, and the expected adjustment direction is consistent with the change direction of the actual push effect data.

[0154] In one embodiment, the target push performance data is the target click-through rate (CTR), and the actual push performance data is the actual CTR; or,

[0155] The target push performance data is the target revenue per thousand impressions (ECPM), while the actual push performance data is the actual ECPM.

[0156] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules is merely an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the functional modules described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0157] The apparatus of this invention, by acquiring activity data, target push effect data, and a pre-generated user set corresponding to the promotion activity, can prepare a user group with high potential interest for the promotion activity, set clear and quantifiable push effect requirements for the promotion activity, and provide a reference standard for subsequent traffic control. Based on the number of users to be pushed to, it determines target users from the user set, pushes activity data to the target users, and statistically analyzes the actual push effect data of the activity data within a preset time period. This allows for accurate evaluation of the actual user feedback of the promotion activity, providing a data foundation for subsequent optimization of traffic allocation. Employing a preset closed-loop control algorithm, it determines a push user number adjustment coefficient based on the actual push effect data and the target push effect data, and updates the number of users to be pushed according to the push user number adjustment coefficient. This enables dynamic adjustment of traffic allocation based on the real-time feedback from the user group regarding the promotion activity, improving push accuracy and traffic utilization while enhancing user experience.

[0158] The following is for reference. Figure 4 It shows a schematic diagram of the structure of a computer system 400 suitable for implementing an electronic device according to embodiments of the present invention. Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0159] like Figure 4 As shown, the computer system 400 includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 402 or programs loaded from storage section 408 into random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of the computer system 400. The CPU 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0160] The following components are connected to I / O interface 405: input section 406 including keyboard, mouse, etc.; output section 407 including cathode ray tube, liquid crystal display, etc., and speakers, etc.; storage section 408 including hard disk, etc.; and communication section 409 including network interface card, such as modem, etc. Communication section 409 performs communication processing via a network such as the Internet. Drive 410 is also connected to I / O interface 405 as needed. Removable media 411, such as disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.

[0161] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit (CPU) 401, it performs the functions defined above in the system of this invention.

[0162] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical fiber, a portable compact disk read-only memory, an optical storage device, a magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, etc., or any suitable combination thereof.

[0163] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0164] The modules and / or units described in the embodiments of this invention can be implemented in software or hardware. The described modules and / or units can also be housed in a processor; for example, a processor may include an acquisition module, a push module, a determination module, and an adjustment module. The names of these modules do not necessarily limit the functionality of the module itself.

[0165] In another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist alone and not assembled into the device. The computer-readable medium carries one or more programs, which, when executed by the device, cause the device to include:

[0166] The system acquires activity data, target push performance data, and a pre-generated user set corresponding to the promotion activity; determines target users from the user set based on the number of users to be pushed to, corresponding to the promotion activity and the user set; pushes activity data to the target users and calculates the actual push performance data of the activity data within a preset time period; uses a preset closed-loop control algorithm to determine the adjustment coefficient for the number of push users based on the actual push performance data and the target push performance data; updates the number of users to be pushed according to the adjustment coefficient, and returns to execute the operation of determining target users from the user set based on the number of users to be pushed.

[0167] The technical solution of this invention, by acquiring activity data, target push effect data, and a pre-generated user set corresponding to the activity to be promoted, can prepare a user group with high potential interest for the activity promotion, set clear and quantifiable push effect requirements for the activity promotion, and provide a reference standard for subsequent traffic control. Based on the number of users to be pushed to, target users are determined from the user set, activity data is pushed to the target users, and the actual push effect data of the activity data within a preset time period is statistically analyzed. This allows for accurate evaluation of the actual user feedback of the activity to be promoted, providing a data foundation for subsequent optimization of traffic allocation. A preset closed-loop control algorithm is used to determine the adjustment coefficient for the number of users to be pushed to based on the actual push effect data and the target push effect data, and the number of users to be pushed to is updated according to the adjustment coefficient. This allows for dynamic adjustment of traffic allocation based on the real-time feedback from the user group regarding the activity to be promoted, improving push accuracy and traffic utilization while enhancing user experience.

[0168] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the data push method provided in any embodiment of this invention.

[0169] In the implementation of a computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​as well as conventional procedural programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including local area networks (LANs) or wide area networks (WANs), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0170] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0171] It should be noted that the collection, use, storage, sharing, and transfer of user personal information involved in the technical solution of this invention all comply with the provisions of relevant laws and regulations, and require notification to the user and obtaining the user's consent or authorization. Where applicable, user personal information has undergone de-identification and / or anonymization and / or encryption technical processing. In addition, a corresponding operation entry is provided for the user to choose to agree to or reject the automated decision result; if the user chooses to reject, the process proceeds to the expert decision-making process.

[0172] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A data push method, characterized in that, The method includes: Obtain activity data, target push effect data, and pre-generated user sets corresponding to the promotional activities; Based on the number of users to be pushed to the activity to be promoted and the user set, target users are determined from the user set, the activity data is pushed to the target users, and the actual push effect data of the activity data within a preset time period is statistically analyzed. A preset closed-loop control algorithm is used to determine the adjustment coefficient for the number of push users based on the actual push effect data and the target push effect data; The number of users to be pushed to is updated based on the adjustment coefficient of the number of users to be pushed to, and the operation of determining the target user from the user set based on the number of users to be pushed to is returned.

2. The method according to claim 1, characterized in that, The preset closed-loop control algorithm is a PID control algorithm; the step of using the preset closed-loop control algorithm to determine the adjustment coefficient for the number of push users based on the actual push effect data and the target push effect data includes: The current deviation value is determined based on the difference between the actual push effect data and the target push effect data, and the proportion item data is determined based on the current deviation value and the preset proportion coefficient. The current accumulated deviation value is determined based on the current deviation value and the previously obtained accumulated deviation value, and the integral term data is determined based on the current accumulated deviation value and the preset integral coefficient. The current deviation rate of change is determined based on the current deviation value and the previous deviation value, and the differential term data is determined based on the current deviation rate of change and the preset differential coefficient. The adjustment coefficient for the number of users pushed is determined based on the proportional term data, the integral term data, and the differential term data.

3. The method according to claim 2, characterized in that, The method for determining the preset proportional coefficient, the preset integral coefficient, and the preset differential coefficient includes: The current promotion stage is determined based on the current total number of clicks corresponding to the activity to be promoted. Based on the pre-set correspondence between promotion stages and PID parameters, the PID parameters corresponding to the current promotion stage are determined. The PID parameters include a preset proportional coefficient, a preset integral coefficient, and a preset derivative coefficient. The PID parameters for later promotion phases are smaller than those for earlier promotion phases.

4. The method according to claim 1, characterized in that, The method for determining the target push effect data includes: The current promotion stage is determined based on the current total number of clicks for the activity to be promoted. Based on the pre-set correspondence between promotion stages and target push performance data, determine the target push performance data corresponding to the current promotion stage; Among them, the target push effect data corresponding to the later promotion stage is greater than the target push effect data corresponding to the earlier promotion stage.

5. The method according to claim 1, characterized in that, The users in the user set are sorted in descending order of their predicted interest in the promotional activity; the predicted interest is obtained by inputting the feature information of the corresponding user and the feature information of the promotional activity into a preset neural network model, and based on the output information of the preset neural network model. The step of determining target users from the user set based on the number of users to be pushed to, corresponding to the promotional activity and the user set, includes: Identify the first user in the user set whose position number corresponds to the number of users to be pushed to, and determine the first user and all users before the first user as target users.

6. The method according to claim 1, characterized in that, The method for determining the preset time period includes: Determine the number of unique visitors corresponding to the user set, and determine whether the preset data volume shortage condition is met based on the number of unique visitors; If the insufficient data condition is met, the current window offset is determined, and a time extension operation is performed on the default statistical time window based on the current window offset. The time period corresponding to the extended statistical time window is then used as the preset time period.

7. The method according to claim 6, characterized in that, Determining the current window offset includes: Query the database for the historical window offset corresponding to the promotion activity and the user set; If a historical window offset is found, the current window offset is determined based on the historical window offset and a preset time step, and the historical window offset is updated based on the current window offset. If no historical window offset is found, the current window offset is determined based on a preset time step and written to the database.

8. A data push device, characterized in that, include: The acquisition module is used to acquire activity data, target push effect data, and pre-generated user sets corresponding to the promotion activity; The determination module is used to determine target users from the user set based on the number of users to be pushed to the user set corresponding to the promotion activity and the user set, push the activity data to the target users, and count the actual push effect data of the activity data within a preset time period. The push module is used to determine the adjustment coefficient for the number of push users based on the actual push effect data and the target push effect data using a preset closed-loop control algorithm. The adjustment module is used to adjust the coefficient according to the number of push users to update the number of users to be pushed, and return to perform the operation of determining the target user from the user set based on the number of users to be pushed.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the data push method as described in any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the data push method as described in any one of claims 1 to 7.

Images