Data processing method, device and equipment and computer readable storage medium

Abstract

The embodiment of the invention discloses a data processing method and device, equipment and a computer readable storage medium. The method comprises the steps that first media data and scenarized information sent by data providing equipment are acquired; according to a first playing probability of the first exposure period, performing exposure processing on the first media data, and obtaining return data returned by the data providing equipment according to an exposure result; determining a first initial consumption speed of the first media data according to the return data and the first playing probability; determining an available budget of the first media data in the second exposure period according to the first budget; and determining a second playing probability of the first media data in the second exposure period according to the available budget, the consumption speed and the residual delivery duration, and performing exposure processing on the first media data in the second exposure period according to the second playing probability. By adopting the application, the accuracy of the playing probability can be improved, and the utilization rate of the available budget and the delivery effect of the first media data can also be improved.

Description

Technical Field

[0001] This application relates to the field of Internet technology, and in particular to a data processing method, apparatus, device, and computer-readable storage medium. Background Technology

[0002] Data delivery devices deliver media data in two main ways: conventional delivery and contextual delivery. Contextual delivery refers to delivery that includes information about the context in which the media data belongs. Typical examples of contextual delivery include major game promotions, e-commerce promotions, short drama promotions, and car launch promotions.

[0003] In existing technologies, data delivery devices use high playback probability to expose media data. However, high probability exposure of media data does not guarantee the delivery effect of media data and will increase the budget for media data. Since the delivery effect of media data cannot be guaranteed, the budget is wasted, that is, the utilization rate of the budget is reduced. Summary of the Invention

[0004] This application provides a data processing method, apparatus, device, and computer-readable storage medium, which can not only improve the accuracy of playback probability, but also improve the utilization rate of available budget and the delivery effect of first media data.

[0005] One embodiment of this application provides a data processing method, the method comprising:

[0006] The data acquisition provides the first media data sent by the device and the contextual information of the first media data; the contextual information of the first media data includes the first budget of the first media data and the delivery time of the first media data;

[0007] Based on the first playback probability of the first exposure period, the first media data is exposed, and the feedback data of the first media data returned by the data providing device based on the exposure results is obtained; the exposure results are generated based on the exposure processing of the first exposure period; the first exposure period belongs to the delivery time;

[0008] Based on the returned data and the first playback probability, a first initial consumption rate of the first media data is determined; the first initial consumption rate is used to determine the consumption rate that matches the first budget.

[0009] The available budget for the first media data in the second exposure period is determined based on the first budget; the second exposure period refers to the next exposure period after the first exposure period, and the second exposure period is the campaign period;

[0010] Based on the available budget, consumption rate, and remaining campaign duration, determine the second playback probability of the first media data in the second exposure period, and perform exposure processing on the first media data in the second exposure period according to the second playback probability; the remaining campaign duration is related to the first budget.

[0011] One embodiment of this application provides a data processing apparatus, the apparatus comprising:

[0012] The acquisition module is used to acquire the first media data sent by the data providing device and the contextual information of the first media data; the contextual information of the first media data includes the first budget of the first media data and the delivery time of the first media data;

[0013] The acquisition module is also used to perform exposure processing on the first media data based on the first playback probability of the first exposure period, and to acquire the feedback data of the first media data returned by the data providing device based on the exposure result; the exposure result is generated based on the exposure processing of the first exposure period; the first exposure period belongs to the delivery time;

[0014] The determination module is used to determine the first initial consumption rate of the first media data based on the returned data and the first playback probability; the first initial consumption rate is used to determine the consumption rate that matches the first budget.

[0015] The determination module is also used to determine the available budget for the first media data in the second exposure period based on the first budget; the second exposure period refers to the next exposure period after the first exposure period, and the second exposure period is within the delivery time.

[0016] The determination module is also used to determine the second playback probability of the first media data in the second exposure period based on the available budget, consumption rate and remaining delivery time, and to perform exposure processing on the first media data in the second exposure period according to the second playback probability; the remaining delivery time is associated with the first budget.

[0017] In one possible implementation, the determining module determines the first initial consumption rate of the first media data based on the returned data and the first playback probability, and performs the following operations:

[0018] Obtain the consumed data from the returned data, and perform non-return data estimation based on the scenario domain and conversion goal of the first media data to obtain the non-return data consumption of the first media data in the first exposure period;

[0019] The summation of the reflowed and non-reflowed consumption yields the media data consumption z of the first media data in the first exposure cycle.

[0020] The initial consumption rate of the first media data is determined based on the media data consumption z and the first playback probability.

[0021] In one possible implementation, the determining module determines the first initial consumption rate of the first media data based on the media data consumption z and the first playback probability, and performs the following operations:

[0022] Get the first exposure from the returned data, and perform weighted processing on the media data consumption z based on the first exposure to obtain the first weighted media data consumption;

[0023] Based on the first exposure, the first playback probability is weighted to obtain the first weighted playback probability;

[0024] The initial consumption rate of the first media data is determined based on the first weighted media data consumption and the first weighted playback probability.

[0025] In one possible implementation, the determining module determines the first initial consumption rate of the first media data based on the first weighted media data consumption and the first weighted playback probability, and performs the following operations:

[0026] If the first exposure period is not the first exposure period of the campaign time, then obtain the second weighted playback probability of the first media data in the third exposure period and the second weighted media data consumption; the timestamp corresponding to the third exposure period is earlier than the timestamp corresponding to the first exposure period, and the third exposure period is adjacent to the first exposure period; the third exposure period belongs to the campaign time.

[0027] The first weighted media data consumption and the second weighted media data consumption are summed to obtain the total weighted media data consumption;

[0028] The first weighted playback probability and the second weighted playback probability are summed to obtain the total weighted playback probability;

[0029] The first initial consumption rate of the first media data is obtained by comparing the weighted total consumption of media data with the weighted total playback probability.

[0030] In one possible implementation, the acquisition module acquires the first media data sent by the data providing device and the contextual information of the first media data, and uses it to perform the following operations:

[0031] The data acquisition device sends a plan consisting of one or more media data, and contextual information of the media data included in each of the a plans; a is a positive integer; the a plans include plan c, and plan c includes the first media data;

[0032] The module is also used to perform the following operations:

[0033] If the first budget includes a media data budget, then the first initial consumption rate is determined as the media data consumption rate; the media data budget refers to the budget for the first media data.

[0034] If the first budget does not include the media data budget, but includes the planning budget, then the second initial consumption rate of the second media data is obtained; the second media data includes the remaining media data in plan c other than the first media data; the planning budget refers to the budget for plan c;

[0035] The first initial consumption rate and the second initial consumption rate are summed to obtain the consumption rate of the first media data.

[0036] If the first budget does not include the media data budget and the plan budget, and the first budget includes the account budget, or the first budget is an invalid budget, then the third initial consumption rate of the third media data is obtained; the third media data includes the remaining media data in plan a excluding the first media data; the account budget refers to the budget for plan a.

[0037] The first initial consumption rate and the third initial consumption rate are summed to obtain the consumption rate of the first media data.

[0038] In one possible implementation, the determining module determines the available budget for the first media data in the second exposure cycle based on the first budget, and performs the following operations:

[0039] If the first budget includes at least one of the media data budget, planning budget, or account budget, then the budget consumption of the first media data that matches the first budget is calculated.

[0040] Based on the first budget and budget expenditure, determine the available budget for the first media data in the second exposure cycle;

[0041] If the first budget is invalid, the remaining budget in the platform account provided by the data providing device is determined, and the remaining budget in the account is determined as the available budget for the first media data in the second exposure period.

[0042] In one possible implementation, the determining module determines the available budget for the first media data in the second exposure cycle based on the first budget and budget consumption, and performs the following operations:

[0043] If the first budget includes the media data budget, then the available budget data algorithm is obtained. The first budget and budget consumption are respectively input into the available budget data algorithm. By inputting the available budget data algorithm with the first budget and budget consumption, the available budget for the first media data in the second exposure period is determined.

[0044] If the first budget does not include the media data budget, but includes the planning budget, then obtain the available budget planning algorithm, input the first budget and budget consumption into the available budget planning algorithm respectively, and determine the available budget for the first media data in the second exposure period by inputting the available budget planning algorithm with the first budget and budget consumption;

[0045] If the first budget does not include the media data budget and the planning budget, but includes the account budget, then obtain the available budget account algorithm, input the first budget and budget consumption into the available budget account algorithm respectively, and determine the available budget for the first media data in the second exposure period by inputting the available budget account algorithm with the first budget and budget consumption.

[0046] In one possible implementation, the determining module calculates the budget consumption of the first media data that matches the first budget, and performs the following operations:

[0047] If the first budget includes a media data budget, then the media data budget consumption of the first media data will be calculated.

[0048] If the first budget includes the planned budget, then the planned budget consumption of the first media data is statistically analyzed;

[0049] If the first budget includes the account budget, then the account budget consumption of the first media data will be calculated.

[0050] The algorithm then determines the available budget for the first media data in the second exposure period by inputting the first budget and the available budget data after budget consumption, including:

[0051] In the available budget data algorithm, the difference between the media data budget and the media data budget consumption is calculated to obtain the remaining media data budget.

[0052] The difference between the planned budget and the planned budget consumption is used to obtain the remaining planned budget.

[0053] The difference between the account budget and the account budget consumption is calculated to obtain the remaining account budget.

[0054] The remaining budget among the media data budget, the planned budget, the account budget, and the account budget is determined as the available budget for the first media data in the second exposure period.

[0055] In one possible implementation, the determining module determines the second playback probability of the first media data in the second exposure period based on the available budget, consumption rate, and remaining delivery time, for the following operations:

[0056] Determine the remaining delivery time for the first media data, and multiply the consumption rate by the remaining delivery time to obtain the remaining delivery consumption.

[0057] The ratio of available budget to remaining ad spend is used to obtain the initial playback probability of the first media data in the second exposure period;

[0058] The minimum probability between the initial playback probability and the constrained playback probability is determined as the playback probability to be optimized.

[0059] Determine the scene domain to which the first media data belongs, and obtain the scene probability corresponding to the scene domain;

[0060] The scene probability and the playback probability to be optimized are summed to obtain the second playback probability of the first media data in the second exposure period.

[0061] In one possible implementation, the acquisition module acquires the first media data sent by the data providing device and the contextual information of the first media data, and uses it to perform the following operations:

[0062] The data acquisition device sends a plan consisting of one or more media data, and contextual information of the media data included in each of the a plans; a is a positive integer; the a plans include plan c, and plan c includes the first media data;

[0063] The data acquisition device sends a plan consisting of one or more media data, and contextual information of the media data included in each of the a plans; a is a positive integer; the a plans include plan c, and plan c includes the first media data;

[0064] The module then determines the remaining delivery time of the first media data and performs the following operations:

[0065] The difference between the delivery time and the time corresponding to the first exposure period is calculated to obtain the initial remaining delivery time of the first media data.

[0066] If the first budget includes a media data budget, then the initial remaining delivery time of the first media data will be determined as the remaining delivery time of the first media data.

[0067] If the first budget does not include the media data budget, but includes the planning budget, then the remaining campaign duration of plan c is determined as the remaining campaign duration of the first media data; the remaining campaign duration refers to the maximum initial remaining campaign duration among the initial remaining campaign durations corresponding to all media data included in plan c.

[0068] If the first budget does not include the media data budget and the plan budget, and the first budget includes the account budget, or the first budget is an invalid budget, then obtain the remaining campaign duration for each of the a plans;

[0069] The remaining duration of the longest remaining campaign duration among the a campaigns is determined as the remaining duration of the first media data campaign.

[0070] In one possible implementation, it is determined that the first media data is exposed in the second exposure period according to the second playback probability, for the purpose of performing the following operations:

[0071] During the second exposure cycle, acquire media data acquisition requests sent by the terminal device, and determine the request object information based on the media data acquisition requests;

[0072] In the contextual information, obtain the set of target information for the first media data. If there is target information in the target information set that matches the target information, then add the first media data to the candidate media data set corresponding to the target information according to the second playback probability.

[0073] Determine the first optimized benefit per thousand impressions between the request object information and the first media data, and determine the second optimized benefit per thousand impressions between the fourth media data and the request object information; the fourth media data includes media data in the candidate media data set other than the first media data;

[0074] If the revenue from the first optimization per thousand impressions is greater than or equal to the revenue from the second optimization per thousand impressions, then the first media data will be exposed to the terminal device.

[0075] In one possible implementation, the determining module determines the first optimized thousand-view yield between the request object information and the first media data, and performs the following operations:

[0076] Determine the basic revenue per thousand impressions between the request object information and the primary media data;

[0077] Determine the scene domain to which the first media data belongs, and obtain the revenue per thousand impressions corresponding to the scene domain;

[0078] The basic revenue per thousand impressions and the revenue per thousand impressions in the scenario are summed to obtain the first optimized revenue per thousand impressions between the request object information and the first media data.

[0079] In one possible implementation, the module is also used to perform the following operations:

[0080] Acquire conversion data from the first media platform during the first time period; the conversion data is provided by the data provider device; the first time period refers to the ad placement period;

[0081] Based on the conversion data, generate the actual conversion rate of the first media data in the first time period;

[0082] Obtain the estimated conversion rate from the first media data in the first time period, and determine the actual conversion rate and the conversion rate prediction deviation between the actual and estimated conversion rates;

[0083] If the conversion rate prediction deviation is less than the deviation threshold, then a conversion sample is generated based on the conversion data;

[0084] The parameters in the conversion rate prediction model are adjusted based on the conversion samples to obtain the optimized conversion rate prediction model. The optimized conversion rate prediction model is used to predict the conversion rate of the first media data in the second time period, which is later than the first time period.

[0085] If the conversion rate prediction deviation is greater than or equal to the deviation threshold, the conversion data will be deleted.

[0086] One aspect of this application provides a computer program product, which includes a computer program stored in a computer-readable storage medium; a processor of a computer device reads the computer program from the computer-readable storage medium and executes the computer program, causing the computer device to perform the method described in this application.

[0087] In this embodiment, after exposing the first media data in the first exposure period based on the first playback probability, the consumption rate of the first media data can be determined based on the feedback data and the first playback probability. Then, based on the remaining delivery time, consumption rate, and available budget, the second playback probability of the first media data in the second exposure period can be determined. The first media data is then exposed in the second exposure period based on the second playback probability. As can be seen above, in this embodiment, when delivering the first media data (i.e., exposing the first media data), the playback probability of the first media data in the next exposure period is dynamically adjusted based on real-time feedback data and the playback probability of the previous exposure period. Therefore, the accuracy of the playback probability can be improved. Based on the accurate playback probability for each period, the utilization rate of the available budget and the delivery effect of the first media data can be improved. Attached Figure Description

[0088] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0089] Figure 1This is a schematic diagram of a system architecture provided in an embodiment of this application;

[0090] Figure 2 This is a flowchart illustrating a data processing method provided in an embodiment of this application. Figure 1 ;

[0091] Figure 3 This is a schematic diagram of a data processing scenario provided in an embodiment of this application. Figure 1 ;

[0092] Figure 4 This is a schematic diagram of a data processing scenario provided in an embodiment of this application. Figure 2 ;

[0093] Figure 5 This is a flowchart illustrating a data processing method provided in an embodiment of this application. Figure 2 ;

[0094] Figure 6 This is a schematic diagram of a data processing scenario provided in an embodiment of this application. Figure 3 ;

[0095] Figure 7 This is a flowchart illustrating a data processing method provided in an embodiment of this application. Figure 3 ;

[0096] Figure 8 This is a schematic diagram of a data processing scenario provided in an embodiment of this application. Figure 4 ;

[0097] Figure 9 This is a schematic diagram of a data processing scenario provided in an embodiment of this application. Figure 5 ;

[0098] Figure 10 This is a schematic diagram of the structure of a data processing device provided in an embodiment of this application;

[0099] Figure 11 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0100] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0101] For details, please see Figure 1 , Figure 1This is a schematic diagram of a system architecture provided in an embodiment of this application. For example... Figure 1 As shown, the network architecture may include service server 2000, service server 5000, terminal device cluster 3000, and terminal device cluster 4000. Specifically, terminal device cluster 3000 may include one or more terminal devices; the number of terminal devices in terminal device cluster 3000 is not limited here. Figure 1 As shown, the multiple terminal devices may specifically include terminal device 3000a, terminal device 3000b, terminal device 3000c, ..., terminal device 3000n; terminal device 3000a, terminal device 3000b, terminal device 3000c, ..., terminal device 3000n can be directly or indirectly connected to the business server 2000 via wired or wireless communication, so that each terminal device can interact with the business server 2000 through the network connection.

[0102] Specifically, the terminal device cluster 4000 may include one or more terminal devices; the number of terminal devices in the terminal device cluster 4000 will not be limited here. For example... Figure 1 As shown, the multiple terminal devices may specifically include terminal device 4000a, terminal device 4000b, terminal device 4000c, ..., terminal device 4000n; terminal device 4000a, terminal device 4000b, terminal device 4000c, ..., terminal device 4000n can be directly or indirectly connected to the business server 5000 via wired or wireless communication, so that each terminal device can interact with the business server 5000 through the network connection.

[0103] Each terminal device in terminal device cluster 3000 and terminal device cluster 4000 may include: smartphones, tablets, laptops, desktop computers, intelligent voice interaction devices, smart home appliances (e.g., smart TVs), wearable devices, vehicle terminals, aircraft, and other intelligent terminals with data processing capabilities.

[0104] Specifically, business server 2000 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Business server 5000 can also be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0105] The business server 2000 and terminal device cluster 3000 can be provided by a data provider, such as a platform that provides media data, like an advertiser. The business server 5000 and terminal device cluster 4000 can be provided by a data delivery provider, such as a platform that delivers media data to other platforms, like an advertising platform. The data providing device described in this embodiment refers to the device corresponding to the data provider, which can be one or more terminal devices in the business server 2000 or terminal device cluster 3000, or one or more terminal devices in the business server 2000 and terminal device cluster 3000. The data delivery device described in this embodiment refers to the device corresponding to the data delivery provider, which can be one or more terminal devices in the business server 5000 or terminal device cluster 4000, or one or more terminal devices in the business server 5000 and terminal device cluster 4000.

[0106] Data providing equipment and data delivery equipment jointly participate in the contextualized delivery of media data. For example, business server 2000 acts as the data providing equipment, and business server 5000 acts as the data delivery equipment. The process of contextualized delivery of media data is as follows: The data delivery equipment obtains the first media data and the contextualized information of the first media data sent by the data providing equipment. In this embodiment, the first media data refers to the data to be contextualized, and the data type of the first media data is not limited, including but not limited to text, images, videos, and audio. The contextualized information of the first media data includes the first budget and the delivery time of the first media data. The first budget refers to the budget set for the first media data within the delivery time.

[0107] Based on the first playback probability of the first exposure period, the data delivery device can perform exposure processing on the first media data and obtain the feedback data returned by the data provider based on the exposure results for the first media data. The data delivery device divides the exposure process of the media data into exposure periods, where each exposure period corresponds to a different time period, but the duration of each exposure period is the same. For example, in this embodiment, the period duration is 1 minute, the first exposure period is 12:05, and the next exposure period, i.e., the second exposure period in this embodiment, is 12:06. This embodiment does not limit the duration of the first exposure period and can be set according to the actual application scenario. The exposure result is generated based on the exposure processing of the first exposure period; the first exposure period belongs to the delivery time. Based on the feedback data and the first playback probability, the data delivery device can determine the first initial consumption rate of the first media data; the first initial consumption rate is used to determine the consumption rate matching the first budget; the consumption rate refers to the budget consumption of the first media data per unit time.

[0108] Based on the first budget, the data delivery device can determine the available budget for the first media data in the second exposure period; the second exposure period refers to the next exposure period after the first exposure period, and the second exposure period belongs to the delivery time. Furthermore, based on the available budget, consumption rate, and remaining delivery time, the data delivery device can determine the second playback probability of the first media data in the second exposure period, and perform exposure processing on the first media data in the second exposure period according to the second playback probability; the remaining delivery time is related to the first budget.

[0109] It is understood that in the specific implementation of this application, data related to user information (such as request object information) is involved. When the embodiments in this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with relevant laws, regulations and standards.

[0110] It is understood that the methods provided in this application embodiment can be executed by computer devices, including but not limited to terminal devices or business servers. The business server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud databases, cloud services, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. Terminal devices include but are not limited to mobile phones, computers, intelligent voice interaction devices, smart home appliances, vehicle terminals, and aircraft. The terminal devices and business servers can be directly or indirectly connected via wired or wireless means, and this application embodiment does not impose any limitations on this connection.

[0111] Further, please see Figure 2 , Figure 2 This is a flowchart illustrating a data processing method provided in an embodiment of this application. Figure 1 The embodiments of this application can be applied to various scenarios, including but not limited to cloud technology, artificial intelligence, smart transportation, assisted driving, audio and video, etc.

[0112] This data processing method can be executed by a data providing device, a data delivery device, or interactively by both. For ease of understanding, this application's embodiments illustrate this method as being executed by a data delivery device, where the data delivery device acts as a computer device executing the data processing method. Figure 2 As shown, the data processing method may include at least the following steps S101-S105.

[0113] Step S101: Obtain the first media data sent by the data providing device and the contextual information of the first media data; the contextual information of the first media data includes the first budget of the first media data and the delivery time of the first media data.

[0114] Specifically, the data acquisition device sends a plan that includes one or more media data, as well as contextual information of the media data included in each of the a plans; a is a positive integer; the a plans include plan c, and plan c includes the first media data.

[0115] Please see also Figure 3 , Figure 3 This is a schematic diagram of a data processing scenario provided in an embodiment of this application. Figure 1 . Figure 3 The media data described in the embodiments of this application (including the first media data described below) is exemplified as an advertisement. Figure 3 As shown, the data providing device 20a can be the device corresponding to the game advertiser. The game advertiser sends game advertisements and contextual information of the game advertisements to the data delivery device 5000 through the data providing device 20a.

[0116] Data provider device 20a holds 3 accounts. Account 1 is bound to 3 plans: Plan 1, Plan 2, and Plan 3. Figure 3 Let's take "a" as an example and consider it as three plans. The budget set for Plan 1, or the budget set for the ads in Plan 1, is provided by Account 1. Each plan can include one or more plans. Figure 3 Example plan 1 includes multiple ads, such as ad 1, ad 2, ..., ad n. The first media data can be... Figure 3 Either Ad 1 or Ad 2, as shown in the example.

[0117] Figure 3 By only describing Plan 1 and Account 1, it can be understood that the applications corresponding to Account 2 and Account 3 are the same as those corresponding to Account 1, and the applications corresponding to Plan 2 and Plan 3 are the same as those corresponding to Plan 1.

[0118] If we take Ad 1 as an example of the first media data, the contextual information of the first media data can include the ad name, ad placement time, target audience information of Ad 1, initial budget of Ad 1, ad attributes, etc. The specific settings can be made according to the actual application scenario.

[0119] Step S102: Based on the first playback probability of the first exposure period, perform exposure processing on the first media data, and obtain the feedback data of the first media data returned by the data providing device based on the exposure result; the exposure result is generated based on the exposure processing of the first exposure period; the first exposure period belongs to the delivery time.

[0120] Specifically, if the first exposure period is the first exposure period of the campaign, for example, the campaign period is from September 1, 2024 to September 3, 2024, and the duration of each exposure period is 1 minute, then the first exposure period is from 0:00 on September 1, 2024 to 0:01 on September 1, 2024. In this case, the first playback probability is a hyperparameter that can be set according to the scene field (i.e., industry field) to which the first media data belongs. For example, if the first media data is a game advertisement, then an initial playback probability can be set according to the game advertisement industry.

[0121] If the first exposure period is not the first exposure period of the campaign, for example, the campaign period is from September 1, 2024 to September 3, 2024, and the duration of each exposure period is 1 minute, the first exposure period is from 0:59 on September 1, 2024 to 1:00 on September 1, 2024. In this case, the process of determining the first playback probability is the same as the process of determining the second playback probability in step S105, that is, it is determined based on the available budget corresponding to the previous exposure period of the first exposure period (the available budget in step S104 is for the first exposure period), the consumption rate corresponding to the previous exposure period, and the remaining campaign duration corresponding to the previous exposure period.

[0122] Exposure processing refers to the process of presenting primary media data to the target audience (i.e., the intended recipients). At the end of the first exposure period, the data providing device can obtain the exposure results, which reflect the effectiveness of the re-run of the primary media data during the first exposure period. The data providing device integrates and analyzes the exposure results to obtain feedback data. The feedback data is of great significance for evaluating and optimizing exposure processing; therefore, the data providing device sends the feedback data to the data delivery device.

[0123] Step S103: Determine the first initial consumption rate of the first media data based on the returned data and the first playback probability; the first initial consumption rate is used to determine the consumption rate that matches the first budget.

[0124] Specifically, the process involves obtaining the consumed data from the returned data, performing non-return data estimation based on the scenario domain and conversion target of the first media data, and obtaining the non-return data consumption of the first media data in the first exposure period; summing the consumed data from the returned data and the non-return data consumption to obtain the media data consumption z of the first media data in the first exposure period; and determining the first initial consumption rate of the first media data based on the media data consumption z and the first playback probability.

[0125] The specific process of determining the first initial consumption rate of the first media data based on the first media data consumption and the first playback probability may include: if the first exposure period is not the first exposure period of the delivery time, then obtain the third playback probability and media data consumption y of the first media data in the third exposure period; the time corresponding to the third exposure period is earlier than the time corresponding to the first exposure period, and the third exposure period is adjacent to the first exposure period; the third exposure period belongs to the delivery time; sum the media data consumption z and the media data consumption y to obtain the total media data consumption; sum the first playback probability and the third playback probability to obtain the total playback probability; and perform ratio processing on the total media data consumption and the total playback probability to obtain the first initial consumption rate of the first media data.

[0126] The data delivery device determines the scene domain to which the first media data belongs. For example, if the first media data is a game advertisement, then its scene domain can be the game domain; if the first media data is a car advertisement, then its scene domain can be the car domain; and if the first media data is an e-commerce advertisement, then its scene domain can be the e-commerce domain.

[0127] Conversion goals are set by the data provider, such as downloads, registrations, activations, and lead generation. The cost of repeat purchases equals the product of the number of conversions and the bid for a single conversion action on the first media data. For example, if the bid for a single conversion action is 10 digital resources, then for each conversion goal achieved on the first media data, the data provider pays the data delivery device 10 digital resources. The number of conversions refers to the number of conversion goals achieved within the first exposure period. Assuming the e-commerce advertiser's conversion action is set to download, and 10 end users download the first media data (e.g., an e-commerce ad) during the first exposure period, the number of conversions is 10.

[0128] To facilitate understanding, let's take advertising as an example of media data. Calculating the consumption of the data-providing device (which can be understood as the advertiser) on the first media data (i.e., an advertisement) requires summing the consumption of returned traffic and the consumption of non-returned traffic. The estimation of non-returned traffic is related to the combination of industry (games, e-commerce, automobiles, etc.) and conversion goals (activation, registration, payment, etc.). Assuming that the returned traffic consumption of both short drama ads and game ads is 6, but the activation return traffic of short drama ads is faster, while the payment return traffic of game ads is slower, the non-returned traffic consumption of short drama ads may be 6, and the non-returned traffic consumption of game ads may be 4. Therefore, the total consumption of short drama ads (i.e., the media data consumption z of short drama ads) is 12, and the total consumption of game ads (i.e., the media data consumption z of game ads) is 10.

[0129] Taking the first media data as an example of an advertisement, the first feasible way to determine the first initial consumption rate of the first media data is as shown in the following formula (1):

[0130]

[0131] In formula (1), t′ represents the first exposure period; return_ad_cost(t′) represents the reflowed consumption of the first media data in the first exposure period; unreturn_ad_cost(t′) represents the unreflowed consumption of the first media data in the first exposure period; prot(t′) represents the first playback probability of the first media data in the first exposure period; return_ad_cost(t′)+unreturn_ad_cost(t′) represents the media data consumption z of the first media data in the first exposure period; and ad_cost_speed represents the first initial consumption speed of the first media data in the first exposure period.

[0132] Assuming return_ad_cost(t′) = 1000, unreturn_ad_cost(t′) = 500, and prot(t′) = 0.05, the data delivery device can obtain return_ad_cost(t′) + unreturn_ad_cost(t′) = 1500. Based on prot(t′) and return_ad_cost(t′) + unreturn_ad_cost(t′) = 1500, the data delivery device obtains ad_cost_speed = 30000.

[0133] To eliminate the influence of outlier data from individual exposure cycles, the first initial consumption rate is calculated using data from the most recent D (positive integers greater than 1) exposure cycles; therefore, the second feasible method for determining the first initial consumption rate of the first media data is shown in the following formula (2):

[0134]

[0135] In the above formula, t represents the t-th exposure period out of D exposure periods. The first exposure period can be represented by the D-th exposure period. The third exposure period includes the first exposure period, the second exposure period, ..., the (D-1)-th exposure period out of D exposure periods. return_ad_cost(t) represents the reflowed cost of the first media data in the t-th exposure period, and unreturn_ad_cost(t) represents the unreflowed cost of the first media data in the t-th exposure period. prot(t) represents the playback probability of the first media data in the t-th exposure period. The above return_ad_cost(t) + unreturn_ad_cost(t) represents the media data consumption of the first media data in the t-th exposure period, including the media data consumption z of the first media data in the first exposure period (i.e., the D-th exposure period mentioned above), and the media data consumption y of the first media data in the third exposure period (i.e., the first exposure period, the second exposure period, ..., the (D-1)-th exposure period mentioned above).

[0136] Assuming D=5, the first exposure period in this embodiment is represented as the fifth exposure period out of the last five exposure periods; the third exposure period in this embodiment is the first to fourth exposure periods out of the last five exposure periods. The media data consumption y of the third exposure period includes the media data consumption y of the first exposure period, the media data consumption y of the second exposure period, the media data consumption y of the third exposure period, and the media data consumption y of the fourth exposure period; similarly, the third playback probability of the third exposure period includes the third playback probability of the first exposure period, the third playback probability of the second exposure period, the third playback probability of the third exposure period, and the third playback probability of the fourth exposure period.

[0137] Step S104: Determine the available budget for the first media data in the second exposure period based on the first budget; the second exposure period refers to the next exposure period after the first exposure period, and the second exposure period belongs to the campaign period.

[0138] Specifically, if the first budget includes at least one of media data budget, planning budget, or account budget, then the budget consumption of the first media data that matches the first budget is calculated; based on the first budget and budget consumption, the available budget for the first media data in the second exposure period is determined; if the first budget is an invalid budget, then the remaining budget in the platform account provided by the data providing device is determined, and the remaining budget in the account is determined as the available budget for the first media data in the second exposure period.

[0139] The specific process of determining the available budget for the first media data in the second exposure period based on the first budget and budget consumption may include: if the first budget includes the media data budget, then an available budget data algorithm is obtained, and the first budget and budget consumption are respectively input into the available budget data algorithm. By inputting the available budget data algorithm with the first budget and budget consumption, the available budget for the first media data in the second exposure period is determined; if the first budget does not include the media data budget, but the first budget includes the planning budget, then an available budget planning algorithm is obtained, and the first budget and budget consumption are respectively input into the available budget planning algorithm. By inputting the available budget planning algorithm with the first budget and budget consumption, the available budget for the first media data in the second exposure period is determined; if the first budget does not include the media data budget and the planning budget, but the first budget includes the account budget, then an available budget account algorithm is obtained, and the first budget and budget consumption are respectively input into the available budget account algorithm. By inputting the available budget account algorithm with the first budget and budget consumption, the available budget for the first media data in the second exposure period is determined.

[0140] The specific process of calculating the budget consumption of the first media data matching the first budget may include: if the first budget includes the media data budget, then calculate the media data budget consumption of the first media data; if the first budget includes the planned budget, then calculate the planned budget consumption of the first media data; if the first budget includes the account budget, then calculate the account budget consumption of the first media data; then, by inputting an algorithm with available budget data containing the first budget and budget consumption, the available budget for the first media data in the second exposure period is determined, including: in the available budget data algorithm, subtracting the media data budget and the media data budget consumption to obtain the remaining media data budget; subtracting the planned budget and the planned budget consumption to obtain the remaining planned budget; subtracting the account budget and the account budget consumption to obtain the remaining account budget; and determining the minimum remaining budget among the remaining media data budget, the remaining planned budget, the remaining account budget, and the remaining account budget as the available budget for the first media data in the second exposure period.

[0141] Based on the initial budget, the available budget for the first media data in the second exposure cycle can be determined in the following four ways:

[0142] Taking the first media data as an example of advertising, if the first budget includes the media data budget (which can be understood as the advertising budget), then the available budget data algorithm is used, as shown in formula (3):

[0143]

[0144]

[0145] Among them, the media data budget can be at least one of the daily advertising limit or the total advertising limit in the above formula (3); the daily advertising limit refers to the maximum amount of money that can be spent by an advertisement in a day during the contextualized advertising campaign; the total advertising limit refers to the maximum amount of money that can be spent by an advertisement in total during the campaign period; the daily advertising expenditure refers to the expenditure of advertising in a day during the contextualized advertising campaign; the total advertising expenditure refers to the cumulative expenditure of advertising in total during the campaign period; the planned daily limit refers to the sum of the maximum amount of money that can be spent by all planned advertisements in a day during the contextualized advertising campaign; the planned daily expenditure refers to the sum of the expenditure of all planned advertisements in a day during the contextualized advertising campaign; the planned total limit refers to the sum of the maximum amount of money that can be spent by all planned advertisements in total during the contextualized advertising campaign; the planned total expenditure refers to the sum of the cumulative expenditure of all planned advertisements in total during the contextualized advertising campaign. Account daily limit refers to the maximum amount of money that can be spent on all ads in an account within a day during the contextualized ad delivery process; account daily consumption refers to the sum of the costs consumed by all ads in an account within a day during the contextualized ad delivery process; account balance refers to the remaining usable resources in the account during the contextualized ad delivery process; shared fees refer to the fees that multiple accounts under the data provider's name jointly own, manage, and use during the contextualized ad delivery process.

[0146] It should be emphasized that the scenario in which formula (3) is applied refers to the first budget including the media data budget. Under this premise, the first budget may include at least one of the planned budget or the account budget, or may not include the planned budget and the account budget.

[0147] If the first budget does not include the media data budget, but includes the planning budget, the available budget planning algorithm is used, as shown in formula (4):

[0148]

[0149] The planned budget can be at least one of the planned daily limit or the planned total limit in formula (4) above. For the meaning of each item in formula (4), please refer to the description in formula (3) above, which will not be repeated here.

[0150] It should be emphasized that the scenario in which formula (4) is applied refers to the first budget not including the media data budget and the first budget including the planning budget. Under this premise, the first budget may include the account budget or may not include the account budget.

[0151] If the first budget does not include the media data budget and the planning budget, but includes the account budget, the available budget account algorithm shall be used, as shown in formula (5):

[0152] account_left_budget = min(account daily limit - account daily spending),

[0153] Account balance + sharing fees)(5)

[0154] The account budget can be the daily account limit in formula (5) above. For the meaning of each item in formula (5), please refer to the description in formula (3) above, which will not be repeated here.

[0155] If the first budget is an invalid budget, the calculation method is shown in formula (6):

[0156] account_left_budget = account balance + sharing fee (6)

[0157] When the data provider has only one account, there is no sharing fee, and the available budget is the account balance.

[0158] Step S105: Based on the available budget, consumption rate, and remaining campaign duration, determine the second playback probability of the first media data in the second exposure period, and perform exposure processing on the first media data in the second exposure period according to the second playback probability; the remaining campaign duration is related to the first budget.

[0159] Specifically, the remaining delivery time of the first media data is determined, and the consumption rate and the remaining delivery time are multiplied to obtain the remaining delivery consumption. The ratio of the available budget and the remaining delivery consumption is calculated to obtain the initial playback probability of the first media data in the second exposure period. The minimum probability between the initial playback probability and the constrained playback probability is determined as the playback probability to be optimized. The scene domain to which the first media data belongs is determined, and the scene probability corresponding to the scene domain is obtained. The scene probability and the playback probability to be optimized are summed to obtain the second playback probability of the first media data in the second exposure period.

[0160] The specific process for determining the remaining delivery time of the first media data may include: subtracting the delivery time from the time corresponding to the first exposure cycle to obtain the initial remaining delivery time of the first media data; if the first budget includes the media data budget, then the initial remaining delivery time of the first media data is determined as the remaining delivery time of the first media data; if the first budget does not include the media data budget, but includes the plan budget, then the remaining delivery time of plan c is determined as the remaining delivery time of the first media data; the remaining delivery time of the plan refers to the maximum initial remaining delivery time among the initial remaining delivery times corresponding to all media data included in plan c; if the first budget does not include the media data budget and the plan budget, but includes the account budget, or the first budget is an invalid budget, then the remaining delivery time of each of the a plans is obtained; the maximum remaining delivery time among the remaining delivery times of the a plans is determined as the remaining delivery time of the first media data.

[0161] For details, please refer to the following: Figure 4 , Figure 4 This application provides an illustration of a data processing scenario. Figure 2 If the available budget will not be exhausted within the remaining campaign duration, the risk of over-playing is considered low, and no over-play control is needed (the playback probability is 100%). Figure 4 As shown, there are three ways to calculate the probability of the second playback, as follows:

[0162] If the first budget includes the media data budget, then the formula (7) for calculating the second playback probability is as follows:

[0163] prob=min(ad_left_budget / ad_cost_speed / k1,1.0)+p_alpha (7)

[0164] In formula (7), ad_left_budget represents the available budget determined by formula (3), which refers to the resources that can still be used in the remaining delivery time; ad_cost_speed represents the media data consumption rate, which is the consumption rate of the first media data under the condition that the first budget includes the media data budget; prob is the second playback probability, which refers to the probability that the first media data will be played in the next exposure cycle during the contextualized ad delivery process. k1 represents the remaining delivery time, which refers to the remaining time from the current moment to the end of the first media data delivery during the contextualized ad delivery process. p_alpha represents the scenario probability value, which refers to the scenario parameter value set by the data delivery platform according to the industry and scenario field of the advertisement under the contextualized ad delivery, and its value range is [0, 1].

[0165] If the first budget does not include the media data budget, and the first budget includes the planning budget, then the formula (8) for calculating the second playback probability is as follows:

[0166] prob=min(campaign_left_budget / campaign_cost_speed / k2, 1.0)+p_alpha (8)

[0168] In formula (8), campaign_left_budget represents the available budget determined by formula (4); campaign_cost_speed represents the planned consumption speed, which is the sum of the initial consumption speeds of all ads in the plan. Under the condition that the first budget does not include the media data budget but includes the plan budget, the planned consumption speed is the consumption speed of the first media data; k2 represents the remaining delivery time, which is the maximum value of the remaining time length from the current time to the end time of delivery for each ad in the plan during the contextualized delivery of ads.

[0169] For the relationship between campaign_cost_speed in formula (8) and ad_cost_speed in formulas (1) and (2), please refer to the following text. Figure 5 The description in the corresponding embodiments.

[0170] If the first budget does not include the media data budget and the planning budget, and the first budget includes the account budget, or the first budget is an invalid budget, the formula (9) for calculating the second playback probability is as follows:

[0171] prob=min(account_left_budget / account_cost_speed / k3,1.0)+p_alpha (9)

[0172] In formula (9), account_left_speed represents the available budget determined by formula (5); account_cost_speed represents the account consumption speed, which is the sum of the initial consumption speeds of all ads in the account. Under the condition that the first budget does not include the media data budget and the plan budget, but includes the account budget, or the first budget is an invalid budget, the account consumption speed is the consumption speed of the first media data; k3 represents the remaining delivery time, which is the maximum value of the remaining time length from the current time to the end time of delivery for each ad in the ad scenario delivery process.

[0173] For the relationship between account_cost_speed in formula (9) and ad_cost_speed in formulas (1) and (2), please refer to the following text. Figure 5 The description in the corresponding embodiments.

[0174] As described above, the embodiments of this application calculate the playback probability based on available budget, consumption rate, and remaining delivery time. Controlling the exposure processing of media data through the playback probability can help the data delivery device allocate resources more rationally and recommend content that better matches the interests and needs of the terminal users. Controlling the exposure of media data through the playback probability can also provide a more accurate basis for media data delivery, accurately control the resource consumption of media data during the delivery process, improve resource utilization, and thus further enhance the revenue of both the data provider and the data delivery party.

[0175] Please see Figure 5 , Figure 5 This is a flowchart illustrating a data processing method provided in an embodiment of this application. Figure 2 This data processing method can be executed by a data providing device, a data delivery device, or interactively by both. For ease of understanding, this application's embodiments illustrate the method as being executed by a data delivery device. Figure 5 As shown, the data processing method includes steps S201-S210.

[0176] Step S201: Obtain the first media data sent by the data providing device and the contextual information of the first media data; the contextual information of the first media data includes the first budget of the first media data and the delivery time of the first media data.

[0177] Specifically, in this application embodiment, the first media data is exemplified as an advertisement, and the contextualized delivery can be a square-based contextualized delivery. For ease of understanding, the following terms are explained first:

[0178] Performance-based advertising: Performance-based advertising is a pay-per-performance online advertising model. Its core objective is to achieve specific behaviors desired by advertisers (referred to as data providers in this application embodiment), such as clicks, downloads, registrations, activations, and lead generation. In optimized cost per action (oCPA) performance advertising, advertisers bid for their desired conversion actions. The advertising platform (referred to as data providers in this application embodiment) uses recommendation capabilities to expose the advertiser's ads to suitable end-users on appropriate media traffic, achieving ad exposure and results within the cost contract stipulated by the advertiser's bid.

[0179] The need for contextualized advertising: Advertisers place ads on advertising platforms in two ways: regular placement and contextualized placement. Contextualized placement refers to ad placement that incorporates industry-specific patterns and occurs within a specific context. Typical examples of contextualized advertising include large-scale game promotions, e-commerce promotions, short drama promotions, and car launch promotions.

[0180] Ad Conversion Rate Prediction Component: A model used in performance advertising systems to predict the probability that an end-user will convert from an ad. See below for an explanation of how to use this component. Figure 7 The description in the corresponding embodiments.

[0181] Ad Playback Probability Control Component: Based on the ad's budget settings, and according to preset rules and real-time data, this component dynamically adjusts the ad playback probability to achieve precise targeting. It integrates machine learning algorithms for real-time calculation and decision-making, improving ad effectiveness.

[0182] Recall and coarse ranking scenario channel component: Based on the information of <ad and audience> provided by advertisers in scenario-based delivery, the component in the ad recommendation system recalls and coarsely ranks the audience for specific ads to pass through the channel, so that the <ad and audience> of scenario-based delivery can enter the calculation queue of fine ranking.

[0183] Fine-ranking commercial value calculation component: In the fine-ranking calculation of the commercial value of advertisements, this application embodiment adds raised eCPM to the original basic effective cost per mille (basic eCPM).

[0184] Step S202: Based on the first playback probability of the first exposure period, perform exposure processing on the first media data and obtain the feedback data of the first media data returned by the data providing device based on the exposure result; the exposure result is generated based on the exposure processing of the first exposure period; the first exposure period belongs to the delivery time.

[0185] For the specific implementation process of step S202, please refer to the above text. Figure 2 Step S102 in the corresponding embodiment will not be described in detail here.

[0186] Step S203: Determine the first initial consumption rate of the first media data based on the returned data and the first playback probability.

[0187] Specifically, the process involves obtaining the consumed data from the returned data, performing non-return data estimation based on the scenario domain and conversion target of the first media data, and obtaining the non-return data consumption of the first media data in the first exposure period; summing the consumed data from the returned data and the non-return data consumption to obtain the media data consumption z of the first media data in the first exposure period; and determining the first initial consumption rate of the first media data based on the media data consumption z and the first playback probability.

[0188] The specific process of determining the first initial consumption rate of the first media data based on the media data consumption z and the first playback probability may include: obtaining the first exposure in the returned data; weighting the first media data consumption based on the first exposure to obtain the first weighted media data consumption; weighting the first playback probability based on the first exposure to obtain the first weighted playback probability; and determining the first initial consumption rate of the first media data based on the first weighted media data consumption and the first weighted playback probability.

[0189] The specific process for determining the first initial consumption rate of the first media data, including the first weighted media data consumption and the first weighted playback probability, may include: if the first exposure period is not the first exposure period of the delivery time, then obtaining the second weighted playback probability and the second weighted media data consumption of the first media data in the third exposure period; the timestamp corresponding to the third exposure period is earlier than the timestamp corresponding to the first exposure period, and the third exposure period is adjacent to the first exposure period; the third exposure period belongs to the delivery time; summing the first weighted media data consumption and the second weighted media data consumption to obtain the total weighted media data consumption; summing the first weighted playback probability and the second weighted playback probability to obtain the total weighted playback probability; and performing ratio processing on the total weighted media data consumption and the total weighted playback probability to obtain the first initial consumption rate of the first media data.

[0190] Step S103 above describes two methods for calculating the first initial consumption rate. The first method is to calculate the first initial consumption rate based on the data of the first exposure cycle (including reflow consumption, non-reflow consumption, and the first playback probability). The second method is to calculate the first initial consumption rate based on the data of the most recent D exposure cycles.

[0191] To improve the confidence level, this application embodiment calculates the first initial consumption rate by weighting the exposure amount of each exposure cycle based on the second method. Therefore, the third feasible method for determining the first initial consumption rate of the first media data is as follows: (10)

[0192]

[0193] Among them, the above formula (10) has the same terms as formula (2), so it will not be repeated here. Please refer to the explanation of formula (2) above; e(t) represents the exposure amount of the first media data in the t-th exposure period. The exposure amount is obtained from the feedback data and refers to the number of times the advertisement is shown to the audience in the contextualized advertising. [return_ad_cost(t)+unreturn_ad_cost(t)]*e(t) represents the weighted media data consumption of the first media data in the t-th exposure period, including the first weighted media data consumption of the first media data in the first exposure period (i.e., the D-th exposure period mentioned above) and the second weighted media data consumption of the first media data in the third exposure period (i.e., the first exposure period, the second exposure period, ..., the D-1-th exposure period mentioned above).

[0194] Step S204: If the first budget includes the media data budget, then the first initial consumption rate is determined as the media data consumption rate; the media data budget refers to the budget for the first media data.

[0195] Step S205: If the first budget does not include the media data budget, and the first budget includes the planning budget, then obtain the second initial consumption rate of the second media data; the second media data includes the remaining media data in the plan c other than the first media data; the planning budget refers to the budget for the plan c.

[0196] Specifically, for ease of understanding, assume that plan c includes 5 advertisements, namely advertisement 1, advertisement 2, advertisement 3, advertisement 4, and advertisement 5. The first media data can be advertisement 1, then the second media data includes advertisement 2, advertisement 3, advertisement 4, and advertisement 5; the second initial consumption rate includes the second initial consumption rate of advertisement 2, the second initial consumption rate of advertisement 3, the second initial consumption rate of advertisement 4, and the second initial consumption rate of advertisement 5; the process of determining the second initial consumption rate is the same as the process of determining the first initial consumption rate.

[0197] Step S206: Summate the first initial consumption rate and the second initial consumption rate to obtain the consumption rate of the first media data.

[0198] Specifically, as can be seen from step S205 above, if the first budget does not include the media data budget, and the first budget includes the planning budget, then the planned consumption rate is calculated as shown in the following formula (11):

[0199] campaign_cost_speed=∑ ad∈campaign ad_cost_speed (11)

[0200] Among them, ∑ in the above formula (11) ad∈campaignad_cost_speed represents the sum of the initial cost-speeds (including the first initial cost-speed and the second initial cost-speed) of all ads in campaign c during the first exposure period; campaign_cost_speed represents the campaign cost-speed of campaign c, where the first media data cost-speed is the same as the campaign cost-speed of campaign c, provided that the first budget does not include the media data budget and the first budget includes the campaign budget.

[0201] Step S207: If the first budget does not include the media data budget and the plan budget, and the first budget includes the account budget, or the first budget is an invalid budget, then obtain the third initial consumption rate of the third media data; the third media data includes the remaining media data in a plans other than the first media data; the account budget refers to the budget for a plans.

[0202] Specifically, assume that account 1 provided by the data providing device is bound to 5 plans, i.e., a=5; the 5 plans are Plan 1, Plan 2, Plan 3, Plan 4, and Plan 5; assume that there are a total of 20 ads in the 5 plans, the first media data can be ad 1 in Plan 1, the third media data includes the remaining 19 ads excluding the first media data; the third initial consumption rate includes the third initial consumption rate of the remaining 19 ads excluding the first media data; the process of determining the third initial consumption rate is the same as the process of determining the first initial consumption rate.

[0203] Step S208: Summing the first initial consumption rate and the third initial consumption rate to obtain the consumption rate of the first media data.

[0204] Specifically, as can be seen from step S207 above, if the first budget does not include the media data budget and the planning budget, and the first budget includes the account budget, or the first budget is an invalid budget, then the account consumption rate is calculated as shown in the following formula (12):

[0205] account_cost_speed=∑ ad∈account ad_cost_speed (12)

[0206] Among them, the ∑ of formula (12) ad∈account ad_cost_speed represents the sum of the initial cost-speeds (including the first initial cost-speed and the third initial cost-speed) of all ads in account 1 during the first exposure period; account_cost_speed represents the account cost-speed of account 1. Under the condition that the first budget does not include the media data budget and the planning budget, and the first budget includes the account budget, the first media data cost budget is the same as the account cost-speed of account 1.

[0207] Step S209: Determine the available budget for the first media data in the second exposure period based on the first budget; the second exposure period refers to the next exposure period after the first exposure period, and the second exposure period belongs to the campaign period.

[0208] For the specific implementation process of step S209, please refer to the above text. Figure 2 Step S104 in the corresponding embodiment will not be described in detail here.

[0209] Step S210: Determine the second playback probability of the first media data in the second exposure period based on the available budget, consumption rate, and remaining delivery time; and perform exposure processing on the first media data in the second exposure period based on the second playback probability; the remaining delivery time is related to the first budget.

[0210] Specifically, during the second exposure cycle, a media data acquisition request sent by the terminal device is obtained, and the request object information is determined based on the media data acquisition request; the set of target object information for the first media data is obtained from the contextual information; if there is target object information in the target object information set that matches the request object information, the first media data is added to the candidate media data set corresponding to the request object information according to the second playback probability; the first optimized thousand-view benefit between the request object information and the first media data is determined, and the second optimized thousand-view benefit between the fourth media data and the request object information is determined; the fourth media data includes media data in the candidate media data set other than the first media data; if the first optimized thousand-view benefit is greater than or equal to the second optimized thousand-view benefit, the first media data is exposed to the terminal device.

[0211] The specific process of determining the first optimized revenue per thousand impressions between the request object information and the first media data may include: determining the basic revenue per thousand impressions between the request object information and the first media data; determining the scene domain to which the first media data belongs and obtaining the scene revenue per thousand impressions corresponding to the scene domain; summing the basic revenue per thousand impressions and the scene revenue per thousand impressions to obtain the first optimized revenue per thousand impressions between the request object information and the first media data.

[0212] Specifically, in the process of contextualized media data delivery (such as large-scale game promotions, e-commerce promotions, short drama promotions, and automobile promotions), the data providing device inputs contextualized media data information into the data delivery device. The data delivery device then uses the target audience information within this contextualized information to prioritize the delivery of contextualized ads to terminal devices that match the target audience information, thus bringing positive effects to contextualized delivery. For example, when an ad request occurs on a terminal device, the ad delivery device will check whether the request target information of that terminal device matches the target audience information. If a match is found, the configured ads will be retrieved and coarsely ranked, then entered into a fine-ranking queue for competition calculation.

[0213] Please see also Figure 6 , Figure 6 This is a schematic diagram of a data processing scenario provided in an embodiment of this application. Figure 3 .like Figure 6 As shown, the data delivery device 5000 acquires the first media data and contextual information of the first media data sent by the data provider 20a. The data provider 20a can be a game advertising provider. The contextual information 201a of the first media data includes the game name, game delivery time, game target information, game budget (referred to as the first budget in this embodiment), game attributes, etc.

[0214] The data delivery device 5000 receives a media data acquisition request sent by the terminal device 10. Based on the media data acquisition request, the data delivery device 5000 obtains the request object information of the terminal device 10. This embodiment does not limit the method by which the data delivery device 5000 obtains the request object information. One feasible method is to include the request object information, such as an Internet Protocol address (IP address) or application account, in the media data acquisition request. Another feasible method is to obtain basic information about the terminal object, such as gender, age, region, and browsing content, provided the terminal object authorizes the request.

[0215] Assuming the target audience for the game ad is aged 18-40, the ad type is competitive, and the platform is mobile; and assuming the target audience is aged 25 and wants to view content including shooting games and competitive games, the data delivery device 5000 compares the target audience information with the target audience information. If the age in the target audience information falls within the age range of the target audience information, and the ad attributes in the target audience information are similar to the viewing content attributes in the target audience information, then the target audience information can be considered a match.

[0216] When the target information matches the request information, the data delivery device uses the second playback probability as a random sampling probability to determine whether to add the first media data to the candidate media data set corresponding to the terminal device 10. The candidate media data set refers to the candidate media data used to expose to the terminal device 10. Therefore, the data delivery device needs to select a candidate media data from the candidate media data set and send the selected candidate media data to the terminal device 10.

[0217] For example, if the second playback probability is 1, it means the media data has a 1% chance of being played, so it will be added to the candidate media data set corresponding to terminal device 10. If the second playback probability is 0.5, it means the media data has a 0.5% chance of not being played, meaning there's a 50% probability it won't be played. Therefore, when there are 10 terminal device requests matching the target audience information, the media data will be randomly added to the candidate media data sets corresponding to 5 terminal devices, meaning the candidate media data sets corresponding to 5 terminal devices will not contain this media data. The above operations are performed in the coarse-grained ranking scenario pathway component. The following operations will be performed in the fine-grained ranking business value calculation component.

[0218] Please see again. Figure 6 , Figure 6 The example candidate media dataset contains 5 ads: game ad 11, ad 12, ad 13, ad 14, and ad 15. The first media data can be game ad 11. Data delivery device 5000 determines the request target information and the basic revenue per thousand impressions (RPM) between game ad 11 and the target audience. The basic RPM refers to the average revenue obtained by the ad for every thousand impressions, calculated as the conversion price of the first media data * pCTR * pCVR. The predicted click-through rate (pCTR) is the probability estimated by a terminal audience after seeing the ad, based on a series of algorithms and data analysis. The predicted conversion rate (pCVR) is the predicted probability of a terminal audience performing a conversion behavior (such as purchasing goods, registering an account, or downloading an application) after clicking the ad.

[0219] Data delivery device 5000 can generate a basic revenue of 500 per thousand impressions for game ad 11; similarly, it can generate a basic revenue of 400 per thousand impressions for ad 12; a basic revenue of 500 per thousand impressions for ad 13; a basic revenue of 500 per thousand impressions for ad 14; and a basic revenue of 300 per thousand impressions for ad 15.

[0220] The data delivery device 5000 determines the scene domain to which the first media data belongs and obtains the scene domain's revenue per thousand impressions. Scene domain revenue per thousand impressions refers to the revenue obtained for every thousand ad impressions during contextualized ad delivery. For example, if the first media data is a game ad, then its scene domain can be the game domain, and the scene domain revenue per thousand impressions corresponding to the game domain is obtained; if the first media data is a car ad, then its scene domain can be the car domain, and the scene domain revenue per thousand impressions corresponding to the game domain is obtained.

[0221] Data delivery device 5000 can generate a revenue of 400 per thousand impressions for game ad 11; similarly, it can generate a revenue of 350 per thousand impressions for ad 12; a revenue of 200 per thousand impressions for ad 13; a revenue of 300 per thousand impressions for ad 14; and a revenue of 350 per thousand impressions for ad 15.

[0222] Data delivery device 5000 sums the basic revenue per thousand impressions and the revenue per thousand impressions in the context of game ad 11, resulting in an optimized revenue per thousand impressions of 900; sums the basic revenue per thousand impressions and the revenue per thousand impressions in the context of ad 12, resulting in an optimized revenue per thousand impressions of 750; sums the basic revenue per thousand impressions and the revenue per thousand impressions in the context of ad 13, resulting in an optimized revenue per thousand impressions of 700; sums the basic revenue per thousand impressions and the revenue per thousand impressions in the context of ad 14, resulting in an optimized revenue per thousand impressions of 800; and sums the basic revenue per thousand impressions and the revenue per thousand impressions in the context of ad 15, resulting in an optimized revenue per thousand impressions of 650.

[0223] The data delivery device (5000) compared the revenue per thousand impressions (ROI) of five optimized ads to identify the ad with the highest ROI per thousand impressions. Figure 6 If the example is game advertisement 11, then game advertisement 11 will be sent to terminal device 10.

[0224] As can be seen from the above, when the first media data is delivered (i.e., the first media data is exposed) in the embodiments of this application, the playback probability of the first media data in the next exposure cycle is dynamically adjusted according to the real-time feedback data and the playback probability of the previous exposure cycle. Therefore, the accuracy of the playback probability can be improved. Based on the accurate playback probability of each cycle, the utilization rate of the available budget and the delivery effect of the first media data can be improved.

[0225] Please see Figure 7 , Figure 7 This is a flowchart illustrating a data processing method provided in an embodiment of this application. Figure 3This data processing method can be executed by a data providing device, a data delivery device, or interactively by both. For ease of understanding, this application's embodiments illustrate the method as being executed by a data delivery device. Figure 7 As shown, the data processing method includes steps S301-S311.

[0226] Step S301: Obtain the first media data sent by the data providing device and the contextual information of the first media data; the contextual information of the first media data includes the first budget of the first media data and the delivery time of the first media data.

[0227] For ease of description and understanding, this application uses advertisements as an example of media data in its embodiments. Please also refer to... Figure 8 , Figure 8 This is a schematic diagram of a data processing scenario provided in an embodiment of this application. Figure 4 .like Figure 8 As shown, the data providing device 20a can be the device corresponding to the game advertiser. The game advertiser sends game advertisements and contextual information 201a of the game advertisements to the data delivery device 5000 through the data providing device 20a. The contextual information 201a of the game advertisements includes the game name; the game delivery time; the game target audience information; the game budget; and the game attributes, etc.

[0228] The data providing device 20b can be the device corresponding to the short drama advertiser. The short drama advertiser sends the short drama advertisement and its contextual information 201b to the data delivery device 5000 through the data providing device 20b. The contextual information 201b of the short drama advertisement includes the short drama delivery time; the short drama name; the short drama target audience information; the short drama budget; and the short drama attributes, etc.

[0229] The data providing device 20c can be the device corresponding to the car advertiser. The car advertiser sends the car advertisement and its contextual information 201c to the data delivery device 5000 through the data providing device 20c. The contextual information 201c of the car advertisement includes information about the target audience, car budget, car name, car delivery time, and car attributes.

[0230] Data delivery device 5000 acquires contextual information 201a of game advertisements sent by game advertisers via channel 1; acquires contextual information 201b of short drama advertisements sent by short drama advertisers via channel 2; and acquires contextual information 201c of car advertisements sent by car advertisers via channel 3. This information can be obtained from... Figure 8It was discovered that the data delivery device 5000 acquired three types of advertising contextual information in different formats. Therefore, the data delivery device 5000 adopted a unified data format in its contextual data processing component to organize the contextual information of the three types of advertising, resulting in three types of structured information with the same format, as shown below: The structured information for game advertisements includes game name, game budget, game delivery time, game target audience information, and game attributes; the structured information for short drama advertisements includes short drama name, short drama budget, short drama delivery time, short drama target audience information, and short drama attributes; and the structured information for car advertisements includes car name, car budget, car delivery time, car target audience information, and car attributes.

[0231] The data delivery device 5000 inputs this structured information into the playback probability calculation and control component, the recall coarse ranking scenario path component + fine ranking commercial value calculation component, and the conversion rate prediction component for subsequent processing and calculation.

[0232] In the contextualized ad delivery process, advertisers input contextualized knowledge into the ad platform. This embodiment of the application, after obtaining the contextualized knowledge input by the advertiser, uses an automated contextualized data processing component to specifically adjust the calculation logic of downstream components such as playback probability calculation and control, coarse-grained scenario channel recall & fine-grained commercial value calculation, and conversion rate prediction. This improves the ad delivery effectiveness during contextualized ad delivery, resulting in better ad performance in terms of both ad spending and conversion rate prediction.

[0233] Step S302: Based on the first playback probability of the first exposure period, perform exposure processing on the first media data and obtain the feedback data of the first media data returned by the data providing device based on the exposure result; the exposure result is generated based on the exposure processing of the first exposure period; the first exposure period belongs to the delivery time.

[0234] For the specific implementation process of step S302, please refer to the above text. Figure 2 Step S102 in the corresponding embodiment will not be described in detail here.

[0235] Step S303: Based on the returned data and the first playback probability, determine the first initial consumption rate of the first media data; the first initial consumption rate is used to determine the consumption rate that matches the first budget.

[0236] For the specific implementation process of step S303, please refer to the above text. Figure 2 The description of step S102 in the corresponding embodiment and the above text Figure 5 Steps S203-S208 in the corresponding embodiments will not be described in detail here.

[0237] Step S304: Determine the available budget for the first media data in the second exposure period based on the first budget; the second exposure period refers to the next exposure period after the first exposure period, and the second exposure period belongs to the campaign period.

[0238] Step S305: Based on the available budget, consumption rate, and remaining campaign duration, determine the second playback probability of the first media data in the second exposure period, and perform exposure processing on the first media data in the second exposure period according to the second playback probability; the remaining campaign duration is related to the first budget.

[0239] For the specific implementation process of steps S304-S305, please refer to the above text. Figure 2 Steps S104-S105 in the corresponding embodiments will not be described in detail here.

[0240] Step S306: Obtain the conversion data of the first media data in the first time period; the conversion data is provided by the data providing device; the first time period belongs to the delivery time.

[0241] For details, please refer to the following: Figure 9 , Figure 9 This is a schematic diagram of a data processing scenario provided in an embodiment of this application. Figure 5 In this application embodiment, the media data can be advertising. During the contextualized media data delivery period (e.g., major game promotions, e-commerce promotions, short drama promotions, car promotions), the conversion rate will fluctuate drastically, potentially being underestimated on the first day and overestimated on the second. To ensure a relatively stable conversion rate during contextualized media data delivery, it is necessary to perform contextualized filtering on the samples of the conversion rate prediction model within a configured timeframe, thereby ensuring that the model's learning effect meets the needs of contextualized delivery.

[0242] like Figure 9 As shown, assuming that advertisement a is promoted at 0:00 on September 1, 2024, the first time period is from 0:00 on September 1, 2024 to 0:00 on September 2, 2024. Assuming that the cycle for obtaining conversion data is one hour, the first time period can be from 0:00 to 1:00 on September 1, or from 1:00 to 2:00, ..., or from 23:00 to 0:00 on September 2.

[0243] The data delivery platform acquires the conversion data of the first media platform during the first time period; it also acquires the conversion data of the first media platform during the third time period. The third time period is the period preceding the first time period and falls within the delivery timeframe.

[0244] Step S307: Based on the conversion data, generate the actual conversion rate of the first media data in the first time period.

[0245] Specifically, the conversion data for the first time period is obtained from step S306 above, and the actual conversion rate of the first media data for the first time period is generated based on the conversion data.

[0246] Step S308: Obtain the estimated conversion rate of the first media data in the first time period, and determine the conversion rate prediction deviation between the actual conversion rate and the estimated conversion rate.

[0247] Specifically, the conversion data for the third time period is obtained from step S306 above. The information of the conversion data for the third time period is as follows: For example, if there are 10 terminal objects that react to advertisement a, and the specific reactions of the 10 terminal objects are 10 clicks and 1 purchase, then 10 samples can be constructed for advertisement a, namely sample 1 (advertisement a, terminal object 1 clicks but does not purchase), sample 2 (advertisement a, terminal object 2 clicks but does not purchase), ..., sample 10 (advertisement a, terminal object 10 clicks and purchases). The information of the conversion data for the first time period is similar to that of the conversion data for the third time period.

[0248] The data providing device acquires the conversion data for the third time period, formats the conversion data in the contextualized data processing component, and then sends the formatted conversion data to the conversion rate prediction component for conversion rate prediction processing.

[0249] The data delivery device inputs the above 10 samples into the conversion rate pre-model. Each sample is classified according to terminal object characteristics, media data characteristics, statistical characteristics of terminal object and media data, and long-term conversion sequence characteristics of terminal object, resulting in data with different characteristics. The four types of feature data are input into the feature encoding layer for processing. In this layer, the high-dimensional, discrete feature data of the four types of feature data is converted into low-dimensional, dense vector representations. These feature vectors are then pooled, removing unimportant feature vectors and retaining important ones. The four types of feature vectors are then input into the feature concatenation layer. In the feature concatenation layer, feature vectors of different categories are concatenated according to the needs of the data delivery device to form feature vectors containing more information, and the format is standardized. The concatenated feature vectors are then input into the feature interaction layer. In the feature interaction layer, multiplication operations are performed on different features to capture their interactions. The interacting feature vectors are then input into the fully connected layer. In the fully connected layer, the various features extracted earlier are fused and abstracted to generate a higher-level, more abstract feature representation vector. After weighted summation and activation function processing of the feature representation vector, a predicted value can be output, which is the estimated conversion rate.

[0250] The difference between the estimated conversion rate and the actual conversion rate is calculated to obtain the conversion rate prediction deviation.

[0251] Step S309: If the conversion rate prediction deviation is less than the deviation threshold, then generate conversion samples based on the conversion data.

[0252] Specifically, the conversion rate prediction deviation in step S308 is compared with the deviation threshold. If the conversion rate prediction deviation is less than the deviation threshold, a conversion sample is generated based on the second conversion data in step S306.

[0253] Step S310: Adjust the parameters in the conversion rate prediction model based on the conversion samples to obtain the conversion rate prediction optimization model; the conversion rate prediction optimization model is used to predict the conversion rate of the first media data in the second time period; the second time period is later than the first time period.

[0254] Step S311: If the conversion rate prediction deviation is greater than or equal to the deviation threshold, then delete the conversion data.

[0255] Specifically, the conversion rate prediction deviation in step S309 is compared with the deviation threshold. If the conversion rate prediction deviation is greater than the deviation threshold, the conversion data is filtered and the conversion data of the first time period in step S308 is deleted.

[0256] As can be seen from the above, when the first media data is delivered (i.e., the first media data is exposed) in the embodiments of this application, the playback probability of the first media data in the next exposure cycle is dynamically adjusted according to the real-time feedback data and the playback probability of the previous exposure cycle. Therefore, the accuracy of the playback probability can be improved. Based on the accurate playback probability of each cycle, the utilization rate of the available budget and the delivery effect of the first media data can be improved.

[0257] Further, please see Figure 10 , Figure 10 This is a schematic diagram of a data processing apparatus provided in an embodiment of this application. The data processing apparatus 1 described above can be used to execute the corresponding steps in the method provided in the embodiment of this application. For example... Figure 10 As shown, the data processing device 1 may include an acquisition module 11 and a determination module 12.

[0258] The acquisition module 11 is used to acquire the first media data sent by the data providing device and the contextual information of the first media data; the contextual information of the first media data includes the first budget of the first media data and the delivery time of the first media data;

[0259] The acquisition module 11 is also used to perform exposure processing on the first media data according to the first playback probability of the first exposure period, and to acquire the feedback data of the first media data returned by the data providing device based on the exposure result; the exposure result is generated based on the exposure processing of the first exposure period; the first exposure period belongs to the delivery time;

[0260] The determining module 12 is used to determine the first initial consumption rate of the first media data based on the returned data and the first playback probability; the first initial consumption rate is used to determine the consumption rate that matches the first budget.

[0261] The determination module 12 is also used to determine the available budget for the first media data in the second exposure period based on the first budget; the second exposure period refers to the next exposure period after the first exposure period, and the second exposure period belongs to the delivery time;

[0262] The determination module 12 is also used to determine the second playback probability of the first media data in the second exposure period based on the available budget, consumption rate and remaining delivery time, and to perform exposure processing on the first media data in the second exposure period according to the second playback probability; the remaining delivery time is associated with the first budget.

[0263] In one possible implementation, the determining module 12 determines the first initial consumption rate of the first media data based on the returned data and the first playback probability, and performs the following operations:

[0264] Obtain the consumed data from the returned data, and perform non-return data estimation based on the scenario domain and conversion goal of the first media data to obtain the non-return data consumption of the first media data in the first exposure period;

[0265] The summation of the reflowed and non-reflowed consumption yields the media data consumption z of the first media data in the first exposure cycle.

[0266] The initial consumption rate of the first media data is determined based on the media data consumption z and the first playback probability.

[0267] In one possible implementation, the determining module 12 determines the first initial consumption rate of the first media data based on the media data consumption z and the first playback probability, and performs the following operations:

[0268] Get the first exposure from the returned data, and perform weighted processing on the media data consumption z based on the first exposure to obtain the first weighted media data consumption;

[0269] Based on the first exposure, the first playback probability is weighted to obtain the first weighted playback probability;

[0270] The initial consumption rate of the first media data is determined based on the first weighted media data consumption and the first weighted playback probability.

[0271] In one possible implementation, the determining module 12 determines the first initial consumption rate of the first media data based on the first weighted media data consumption and the first weighted playback probability, and performs the following operations:

[0272] If the first exposure period is not the first exposure period of the campaign time, then obtain the second weighted playback probability of the first media data in the third exposure period and the second weighted media data consumption; the timestamp corresponding to the third exposure period is earlier than the timestamp corresponding to the first exposure period, and the third exposure period is adjacent to the first exposure period; the third exposure period belongs to the campaign time.

[0273] The first weighted media data consumption and the second weighted media data consumption are summed to obtain the total weighted media data consumption;

[0274] The first weighted playback probability and the second weighted playback probability are summed to obtain the total weighted playback probability;

[0275] The first initial consumption rate of the first media data is obtained by comparing the weighted total consumption of media data with the weighted total playback probability.

[0276] In one possible implementation, the acquisition module 11 acquires the first media data sent by the data providing device and the contextual information of the first media data, and performs the following operations:

[0277] The data acquisition device sends a plan consisting of one or more media data, and contextual information of the media data included in each of the a plans; a is a positive integer; the a plans include plan c, and plan c includes the first media data;

[0278] Then module 12 is also used to perform the following operations:

[0279] If the first budget includes a media data budget, then the first initial consumption rate is determined as the media data consumption rate; the media data budget refers to the budget for the first media data.

[0280] If the first budget does not include the media data budget, but includes the planning budget, then the second initial consumption rate of the second media data is obtained; the second media data includes the remaining media data in plan c other than the first media data; the planning budget refers to the budget for plan c;

[0281] The first initial consumption rate and the second initial consumption rate are summed to obtain the consumption rate of the first media data.

[0282] If the first budget does not include the media data budget and the plan budget, and the first budget includes the account budget, or the first budget is an invalid budget, then the third initial consumption rate of the third media data is obtained; the third media data includes the remaining media data in plan a excluding the first media data; the account budget refers to the budget for plan a.

[0283] The first initial consumption rate and the third initial consumption rate are summed to obtain the consumption rate of the first media data.

[0284] In one possible implementation, the determining module 12 determines the available budget for the first media data in the second exposure cycle based on the first budget, and performs the following operations:

[0285] If the first budget includes at least one of the media data budget, planning budget, or account budget, then the budget consumption of the first media data that matches the first budget is calculated.

[0286] Based on the first budget and budget expenditure, determine the available budget for the first media data in the second exposure cycle;

[0287] If the first budget is invalid, the remaining budget in the platform account provided by the data providing device is determined, and the remaining budget in the account is determined as the available budget for the first media data in the second exposure period.

[0288] In one possible implementation, the determining module 12 determines the available budget for the first media data in the second exposure cycle based on the first budget and budget consumption, and performs the following operations:

[0289] If the first budget includes the media data budget, then the available budget data algorithm is obtained. The first budget and budget consumption are respectively input into the available budget data algorithm. By inputting the available budget data algorithm with the first budget and budget consumption, the available budget for the first media data in the second exposure period is determined.

[0290] If the first budget does not include the media data budget, but includes the planning budget, then obtain the available budget planning algorithm, input the first budget and budget consumption into the available budget planning algorithm respectively, and determine the available budget for the first media data in the second exposure period by inputting the available budget planning algorithm with the first budget and budget consumption;

[0291] If the first budget does not include the media data budget and the planning budget, but includes the account budget, then obtain the available budget account algorithm, input the first budget and budget consumption into the available budget account algorithm respectively, and determine the available budget for the first media data in the second exposure period by inputting the available budget account algorithm with the first budget and budget consumption.

[0292] In one possible implementation, the determining module 12 calculates the budget consumption of the first media data that matches the first budget, and performs the following operations:

[0293] If the first budget includes a media data budget, then the media data budget consumption of the first media data will be calculated.

[0294] If the first budget includes the planned budget, then the planned budget consumption of the first media data is statistically analyzed;

[0295] If the first budget includes the account budget, then the account budget consumption of the first media data will be calculated.

[0296] The algorithm then determines the available budget for the first media data in the second exposure period by inputting the first budget and the available budget data after budget consumption, including:

[0297] In the available budget data algorithm, the difference between the media data budget and the media data budget consumption is calculated to obtain the remaining media data budget.

[0298] The difference between the planned budget and the planned budget consumption is used to obtain the remaining planned budget.

[0299] The difference between the account budget and the account budget consumption is calculated to obtain the remaining account budget.

[0300] The remaining budget among the media data budget, the planned budget, the account budget, and the account budget is determined as the available budget for the first media data in the second exposure period.

[0301] In one possible implementation, the determining module 12 determines the second playback probability of the first media data in the second exposure period based on the available budget, consumption rate, and remaining delivery time, for the following operations:

[0302] Determine the remaining delivery time for the first media data, and multiply the consumption rate by the remaining delivery time to obtain the remaining delivery consumption.

[0303] The ratio of available budget to remaining ad spend is used to obtain the initial playback probability of the first media data in the second exposure period;

[0304] The minimum probability between the initial playback probability and the constrained playback probability is determined as the playback probability to be optimized.

[0305] Determine the scene domain to which the first media data belongs, and obtain the scene probability corresponding to the scene domain;

[0306] The scene probability and the playback probability to be optimized are summed to obtain the second playback probability of the first media data in the second exposure period.

[0307] In one possible implementation, the acquisition module 11 acquires the first media data sent by the data providing device and the contextual information of the first media data, and performs the following operations:

[0308] The data acquisition device sends a plan consisting of one or more media data, and contextual information of the media data included in each of the a plans; a is a positive integer; the a plans include plan c, and plan c includes the first media data;

[0309] The data acquisition device sends a plan consisting of one or more media data, and contextual information of the media data included in each of the a plans; a is a positive integer; the a plans include plan c, and plan c includes the first media data;

[0310] Then module 12 determines the remaining delivery time of the first media data and performs the following operations:

[0311] The difference between the delivery time and the time corresponding to the first exposure period is calculated to obtain the initial remaining delivery time of the first media data.

[0312] If the first budget includes a media data budget, then the initial remaining delivery time of the first media data will be determined as the remaining delivery time of the first media data.

[0313] If the first budget does not include the media data budget, but includes the planning budget, then the remaining campaign duration of plan c is determined as the remaining campaign duration of the first media data; the remaining campaign duration refers to the maximum initial remaining campaign duration among the initial remaining campaign durations corresponding to all media data included in plan c.

[0314] If the first budget does not include the media data budget and the plan budget, and the first budget includes the account budget, or the first budget is an invalid budget, then obtain the remaining campaign duration for each of the a plans;

[0315] The remaining duration of the longest remaining campaign duration among the a campaigns is determined as the remaining duration of the first media data campaign.

[0316] In one possible implementation, it is determined that the first media data is exposed in the second exposure period according to the second playback probability, for the purpose of performing the following operations:

[0317] During the second exposure cycle, acquire media data acquisition requests sent by the terminal device, and determine the request object information based on the media data acquisition requests;

[0318] In the contextual information, obtain the set of target information for the first media data. If there is target information in the target information set that matches the target information, then add the first media data to the candidate media data set corresponding to the target information according to the second playback probability.

[0319] Determine the first optimized benefit per thousand impressions between the request object information and the first media data, and determine the second optimized benefit per thousand impressions between the fourth media data and the request object information; the fourth media data includes media data in the candidate media data set other than the first media data;

[0320] If the revenue from the first optimization per thousand impressions is greater than or equal to the revenue from the second optimization per thousand impressions, then the first media data will be exposed to the terminal device.

[0321] In one possible implementation, the determining module 12 determines the first optimized thousand-view yield between the request object information and the first media data, and performs the following operations:

[0322] Determine the basic revenue per thousand impressions between the request object information and the primary media data;

[0323] Determine the scene domain to which the first media data belongs, and obtain the revenue per thousand impressions corresponding to the scene domain;

[0324] The basic revenue per thousand impressions and the revenue per thousand impressions in the scenario are summed to obtain the first optimized revenue per thousand impressions between the request object information and the first media data.

[0325] In one possible implementation, module 12 is further configured to perform the following operations:

[0326] Acquire conversion data from the first media platform during the first time period; the conversion data is provided by the data provider device; the first time period refers to the ad placement period;

[0327] Based on the conversion data, generate the actual conversion rate of the first media data in the first time period;

[0328] Obtain the estimated conversion rate from the first media data in the first time period, and determine the actual conversion rate and the conversion rate prediction deviation between the actual and estimated conversion rates;

[0329] If the conversion rate prediction deviation is less than the deviation threshold, then a conversion sample is generated based on the conversion data;

[0330] The parameters in the conversion rate prediction model are adjusted based on the conversion samples to obtain the optimized conversion rate prediction model. The optimized conversion rate prediction model is used to predict the conversion rate of the first media data in the second time period, which is later than the first time period.

[0331] If the conversion rate prediction deviation is greater than or equal to the deviation threshold, the conversion data will be deleted.

[0332] In this application, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of a larger module or unit that includes the functionality of that module or unit.

[0333] As can be seen from the above, when the first media data is delivered (i.e., the first media data is exposed) in the embodiments of this application, the playback probability of the first media data in the next exposure cycle is dynamically adjusted according to the real-time feedback data and the playback probability of the previous exposure cycle. Therefore, the accuracy of the playback probability can be improved. Based on the accurate playback probability of each cycle, the utilization rate of the available budget and the delivery effect of the first media data can be improved.

[0334] Further, please see Figure 11 , Figure 11 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. The computer device may be... Figure 1 At least one of the terminal devices or service servers shown. For example... Figure 11 As shown, the computer device 1000 may include: at least one processor 1001, such as a CPU, at least one network interface 1004, a user interface 1003, a memory 1005, and at least one communication bus 1002. The communication bus 1002 is used to enable communication between these components.

[0335] In some embodiments, the user interface 1003 may include a display screen and a keyboard, and the network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be high-speed RAM or non-volatile memory, such as at least one disk storage device. Optionally, the memory 1005 may also be at least one storage device located remotely from the aforementioned processor 1001.

[0336] like Figure 11 As shown, the memory 1005, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and a device control application program.

[0337] exist Figure 11In the computer device 1000 shown, the network interface 1004 provides network communication functionality; the user interface 1003 is mainly used to provide an input interface for the user; and the processor 1001 can be used to call the device control application stored in the memory 1005 to achieve:

[0338] The data acquisition provides the first media data sent by the device and the contextual information of the first media data; the contextual information of the first media data includes the first budget of the first media data and the delivery time of the first media data;

[0339] Based on the first playback probability of the first exposure period, the first media data is exposed, and the feedback data of the first media data returned by the data providing device based on the exposure results is obtained; the exposure results are generated based on the exposure processing of the first exposure period; the first exposure period belongs to the delivery time;

[0340] Based on the returned data and the first playback probability, a first initial consumption rate of the first media data is determined; the first initial consumption rate is used to determine the consumption rate that matches the first budget.

[0341] The available budget for the first media data in the second exposure period is determined based on the first budget; the second exposure period refers to the next exposure period after the first exposure period, and the second exposure period is the campaign period;

[0342] Based on the available budget, consumption rate, and remaining campaign duration, determine the second playback probability of the first media data in the second exposure period, and perform exposure processing on the first media data in the second exposure period according to the second playback probability; the remaining campaign duration is related to the first budget.

[0343] It should be understood that the computer device 1000 described in the embodiments of this application can perform the data processing methods or apparatus described in the preceding embodiments, and will not be repeated here. Furthermore, the beneficial effects of using the same methods will also not be repeated.

[0344] This application also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the data processing methods or apparatus described in the preceding embodiments, which will not be repeated here. Furthermore, the beneficial effects of using the same methods will also not be repeated.

[0345] The aforementioned computer-readable storage medium may be the data processing apparatus provided in any of the foregoing embodiments or the internal storage unit of the aforementioned computer device, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on the computer device.

[0346] Furthermore, the computer-readable storage medium may include both internal storage units and external storage devices of the computer device. The computer-readable storage medium is used to store the computer program and other programs and data required by the computer device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0347] This application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium and executes the computer program, enabling the computer device to perform the data processing methods or apparatus described in the preceding embodiments, which will not be repeated here. Furthermore, the beneficial effects of using the same method will also not be repeated here.

[0348] The terms "first," "second," etc., in the specification, claims, and drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the term "comprising," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or modules, but may optionally include steps or modules not listed, or may optionally include other step units inherent to these processes, methods, apparatuses, products, or devices.

[0349] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0350] The above-disclosed embodiments are merely preferred embodiments of this application and should not be construed as limiting the scope of this application. Therefore, any equivalent variations made in accordance with the claims of this application shall still fall within the scope of this application.

Claims

1. A data processing method, characterized by, The method comprises: obtaining first media data sent by a data providing device and scene information of the first media data; the scene information of the first media data comprises a first budget of the first media data and a delivery time of the first media data; performing exposure processing on the first media data according to a first play probability of a first exposure period, and obtaining return data of the first media data returned by the data providing device according to an exposure result; the exposure result is generated based on the exposure processing of the first exposure period; and the first exposure period belongs to the delivery time; determining a first initial consumption speed of the first media data according to the return data and the first play probability; the first initial consumption speed is used to determine a consumption speed matching the first budget; determining an available budget of the first media data in a second exposure period according to the first budget; the second exposure period refers to a next exposure period of the first exposure period, and the second exposure period belongs to the delivery time; determining a second play probability of the first media data in the second exposure period according to the available budget, the consumption speed and a remaining delivery time length, and performing exposure processing on the first media data in the second exposure period according to the second play probability; the remaining delivery time length is associated with the first budget.

2. The method of claim 1, wherein, The method further comprises: obtaining a returned consumption in the return data, performing non-return estimation processing on the returned consumption according to a scene field to which the first media data belongs and a conversion target, and obtaining a non-return consumption of the first media data in the first exposure period; performing summation processing on the returned consumption and the non-return consumption to obtain media data consumption z of the first media data in the first exposure period; determining the first initial consumption speed of the first media data according to the media data consumption z and the first play probability.

3. The method of claim 2, wherein, The method further comprises: obtaining a first exposure amount in the return data, performing weighting processing on the first media data consumption according to the first exposure amount to obtain first weighted media data consumption; performing weighting processing on the first play probability according to the first exposure amount to obtain first weighted play probability; determining the first initial consumption speed of the first media data according to the first weighted media data consumption and the first weighted play probability.

4. The method of claim 3, wherein, The method further comprises: if the first exposure period is not the first exposure period of the delivery time, obtaining a second weighted play probability of the first media data in a third exposure period and a second weighted media data consumption; the timestamp corresponding to the third exposure period is earlier than the timestamp corresponding to the first exposure period, and the third exposure period is adjacent to the first exposure period; the third exposure period belongs to the delivery time; performing sum processing on the first weighted media data consumption and the second weighted media data consumption to obtain a weighted media data consumption total sum; performing sum processing on the first weighted play probability and the second weighted play probability to obtain a weighted play probability total sum; performing ratio processing on the weighted media data consumption total sum and the weighted play probability total sum to obtain a first initial consumption speed of the first media data.

5. The method of claim 1, wherein, The first media data and the scenario information of the first media data sent by the data providing device include: obtaining a total of a plans each including one or more media data sent by the data providing device, and the scenario information of the media data included in each of the a plans; a is a positive integer; the a plans include a plan c, and the plan c includes the first media data; The method further includes: if the first budget includes a media data budget, determining the first initial consumption speed as a consumption speed of the media data; the media data budget refers to a budget for the first media data; if the first budget does not include the media data budget and includes a plan budget, obtaining a second initial consumption speed of a second media data; the second media data includes the remaining media data in the plan c except the first media data; the plan budget refers to a budget for the plan c; performing sum processing on the first initial consumption speed and the second initial consumption speed to obtain the consumption speed of the first media data; if the first budget does not include the media data budget and the plan budget and includes an account budget or is an invalid budget, obtaining a third initial consumption speed of a third media data; the third media data includes the remaining media data in the a plans except the first media data; the account budget refers to a budget for the a plans; performing sum processing on the first initial consumption speed and the third initial consumption speed to obtain the consumption speed of the first media data.

6. The method of claim 1, wherein, The determination of the available budget of the first media data in the second exposure period according to the first budget includes: if the first budget includes at least one of a media data budget, a plan budget or an account budget, counting a budget consumption of the first media data matching the first budget; determining the available budget of the first media data in the second exposure period according to the first budget and the budget consumption; If the first budget is an invalid budget, determining a remaining budget of an account in a platform account provided by the data providing device, and determining the remaining budget of the account as an available budget of the first media data in a second exposure period.

7. The method of claim 6, wherein, The determining of the available budget of the first media data in the second exposure period according to the first budget and the budget consumption includes: If the first budget includes the media data budget, obtaining an available budget data algorithm, inputting the first budget and the budget consumption into the available budget data algorithm, and determining the available budget of the first media data in the second exposure period through the available budget data algorithm input with the first budget and the budget consumption; If the first budget does not include the media data budget and includes the plan budget, obtaining an available budget plan algorithm, inputting the first budget and the budget consumption into the available budget plan algorithm, and determining the available budget of the first media data in the second exposure period through the available budget plan algorithm input with the first budget and the budget consumption; If the first budget does not include the media data budget and the plan budget and includes the account budget, obtaining an available budget account algorithm, inputting the first budget and the budget consumption into the available budget account algorithm, and determining the available budget of the first media data in the second exposure period through the available budget account algorithm input with the first budget and the budget consumption.

8. The method of claim 7, wherein, The counting of the budget consumption of the first media data matching the first budget includes: If the first budget includes the media data budget, counting the media data budget consumption of the first media data; If the first budget includes the plan budget, counting the plan budget consumption of the first media data; If the first budget includes the account budget, counting the account budget consumption of the first media data; The determining of the available budget of the first media data in the second exposure period through the available budget data algorithm input with the first budget and the budget consumption includes: In the available budget data algorithm, performing difference processing on the media data budget and the media data budget consumption to obtain a media data remaining budget; performing difference processing on the plan budget and the plan budget consumption to obtain a plan remaining budget; performing difference processing on the account budget and the account budget consumption to obtain an account remaining budget; determining the minimum remaining budget of the media data remaining budget, the plan remaining budget, the account remaining budget and the account remaining budget as the available budget of the first media data in the second exposure period.

9. The method of claim 1, wherein, The determining of the second play probability of the first media data in the second exposure period according to the available budget, the consumption speed and the remaining delivery time length includes: determining a remaining delivery time length of the first media data, and performing product processing on the consumption speed and the remaining delivery time length to obtain a remaining delivery consumption; performing ratio processing on the available budget and the remaining delivery consumption to obtain an initial play probability of the first media data in the second exposure period; determining the minimum probability between the initial play probability and a constraint play probability as a to-be-optimized play probability; determining a scene field to which the first media data belongs, and obtaining a scene probability corresponding to the scene field; performing summation processing on the scene probability and the to-be-optimized play probability to obtain a second play probability of the first media data in the second exposure period.

10. The method of claim 9, wherein, The first media data and the scene information of the first media data sent by the data providing device include: a plans each including one or more media data are sent by the data providing device, and the scene information of the media data included in each of the a plans; a is a positive integer; the a plans include a plan c, and the plan c includes the first media data; The determination of the remaining delivery duration of the first media data includes: performing difference processing on the delivery time and the time corresponding to the first exposure period to obtain an initial remaining delivery duration of the first media data; if the first budget includes a media data budget, the initial remaining delivery duration of the first media data is determined as the remaining delivery duration of the first media data; if the first budget does not include the media data budget and includes a plan budget, a plan remaining delivery duration of the plan c is determined as the remaining delivery duration of the first media data; the plan remaining delivery duration refers to the maximum initial remaining delivery duration among initial remaining delivery durations corresponding to all media data included in the plan c; if the first budget does not include the media data budget and the plan budget, and the first budget includes an account budget, or the first budget is an invalid budget, plan remaining delivery durations corresponding to the a plans are obtained; the maximum plan remaining delivery duration among the a plan remaining delivery durations is determined as the remaining delivery duration of the first media data.

11. The method of claim 1, wherein, The exposure processing of the first media data in the second exposure period according to the second play probability includes: in the second exposure period, obtaining a media data obtaining request sent by a terminal device, and determining request object information according to the media data obtaining request; obtaining a delivery object information set of the first media data in the scene information, and if there is delivery object information matching the request object information in the delivery object information set, adding the first media data to a candidate media data set corresponding to the request object information according to the second play probability; determining a first optimized thousand-time display revenue between the request object information and the first media data, and determining a second optimized thousand-time display revenue between fourth media data and the request object information; the fourth media data includes media data in the candidate media data set except the first media data. If the first optimized thousand impression yield is greater than or equal to the second optimized thousand impression yield, the first media data is exposed to the terminal device.

12. The method of claim 11, wherein, The first optimized thousand impression yield between the request object information and the first media data is determined. A basic thousand impression yield between the request object information and the first media data is determined. A scene field to which the first media data belongs is determined, and a scene thousand impression yield corresponding to the scene field is acquired. The basic thousand impression yield and the scene thousand impression yield are summed to obtain the first optimized thousand impression yield between the request object information and the first media data.

13. The method of claim 1, wherein, Further comprising: Conversion data of the first media data in a first period is acquired; the conversion data is provided by the data providing device; The first period belongs to the delivery time; According to the conversion data, a real conversion rate of the first media data in the first period is generated; An estimated conversion rate of the first media data in the first period is acquired, and a conversion rate estimation deviation between the real conversion rate and the estimated conversion rate is determined; If the conversion rate estimation deviation is less than a deviation threshold, a conversion sample is generated according to the conversion data; Parameters in a conversion rate estimation model are adjusted according to the conversion sample to obtain a conversion rate estimation optimization model; the conversion rate estimation optimization model is used to predict a conversion rate of the first media data in a second period; the second period is later than the first period; If the conversion rate estimation deviation is greater than or equal to the deviation threshold, the conversion data is deleted.

14. A data processing apparatus, characterized by Comprising: An acquisition module is configured to acquire first media data and scene information of the first media data sent by a data providing device; The scene information of the first media data comprises a first budget of the first media data and a delivery time of the first media data; The acquisition module is further configured to expose the first media data according to a first play probability of a first exposure period, and acquire return data for the first media data returned by the data providing device according to an exposure result; The exposure result is generated based on exposure processing of the first exposure period; the first exposure period belongs to the delivery time; A determination module is configured to determine a first initial consumption speed of the first media data according to the return data and the first play probability; the first initial consumption speed is used to determine a consumption speed matched with the first budget; The determination module is further configured to determine an available budget of the first media data in a second exposure period according to the first budget; the second exposure period refers to a next exposure period of the first exposure period, and the second exposure period belongs to the delivery time; The determining module is further configured to determine a second play probability of the first media data in the second exposure period according to the available budget, the consumption speed, and a remaining exposure duration, and perform exposure processing on the first media data in the second exposure period according to the second play probability, wherein the remaining exposure duration is associated with the first budget.

15. A computer device, comprising: The computer device comprises: a processor, a memory, and a network interface; the network interface is configured to provide a data communication function, the memory is configured to store a computer program, and the processor is configured to call the computer program to enable the computer device to perform the method in any one of claims 1-13.

16. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is adapted to be loaded and executed by the processor to enable a computer device having the processor to perform the method in any one of claims 1-13.

17. A computer program product, characterised in that, The computer program product comprises a computer program stored in a computer readable storage medium, and the computer program is adapted to be read and executed by the processor to enable a computer device having the processor to perform the method in any one of claims 1-13.

Images