Data processing method and device, program product, storage medium and electronic equipment

Abstract

The application discloses a data processing method and device, a program product, a storage medium and an electronic device. It relates to the field of information flow processing. The method comprises the following steps: in the case that an access request of a user for an information flow page is received, determining a test group corresponding to the access request in multiple test partitions; determining a target control coefficient based on a control coefficient associated with the test group corresponding to the access request in the multiple test partitions; determining a recommendation value of a candidate commodity based on the target control coefficient, a commodity type of the candidate commodity and a recommendation value of the candidate commodity, and determining a commodity to be displayed in the information flow page from multiple candidate commodities based on the recommendation value. The method solves the problem that in the related art, when grouping comparison tests are performed in an advertisement and recommendation mixed arrangement scene, the proportion between advertisement commodities and recommendation commodities displayed by different test groups is difficult to remain consistent, thereby affecting the experimental results of the grouping comparison tests.

Description

Technical Field

[0001] This application relates to the field of information flow processing, and more specifically, to a data processing method, apparatus, program product, storage medium, and electronic device. Background Technology

[0002] In the digital advertising ecosystem, Page View Ratio (PVR) is a key metric for balancing user experience and advertising revenue. Taking the "You May Also Like" feed on a shopping website's homepage as an example, this metric is measured by the proportion of ad impressions to total impressions. Assuming relatively stable traffic within the feed, PVR directly determines the supply of ad traffic; too high a PVR may lead to a poor user experience, while too low a PVR will negatively impact the group's advertising revenue. Therefore, ensuring the stability of PVR is not only a crucial breakthrough for technological innovation but also essential for the platform's sustainable development.

[0003] In a unified approach to advertising and recommendation, the system typically focuses on maintaining overall ad coverage near a predetermined target value, neglecting micro-level balance. Due to the traffic optimization effect of mixed placement, when the value of ad placement differs across traffic pools, the system tends to maximize overall efficiency. This leads to a biased distribution of traffic among test groups in comparative group testing. This imbalance manifests as differences in ad coverage among smaller traffic groups, making it difficult to balance the PVR (Player View Rate) of different test groups. Consequently, it directly impacts the experimental results of comparative group testing, resulting in low accuracy.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This application provides a data processing method, apparatus, program product, storage medium, and electronic device to at least solve the technical problem that, when conducting group comparison tests in a mixed advertising and recommendation scenario, the proportion of advertised and recommended products displayed in different test groups is difficult to maintain, thus affecting the experimental results of the group comparison test.

[0006] According to one aspect of the embodiments of this application, a data processing method is provided, comprising: upon receiving a user's access request for an information flow page, determining a test group corresponding to the access request in multiple test partitions, wherein the multiple test partitions include a recommendation partition and an advertising partition, and the test group is used for group comparison testing of advertising product display and recommended product display; determining a target control coefficient based on the control coefficient associated with the test group corresponding to the access request in the multiple test partitions, wherein the control coefficient is determined based on the target advertising coverage of advertising products; determining the recommendation value of candidate products based on the target control coefficient, the product type of candidate products, and the recommendation value of candidate products, and determining the product to be displayed on the information flow page from multiple candidate products based on the recommendation value.

[0007] Furthermore, multiple test partitions are divided into multiple dimensions. Based on the control coefficient associated with the test group corresponding to the access request in the multiple test partitions, the target control coefficient is determined by: for the target dimension among the multiple dimensions, based on the control coefficient of the test group corresponding to the access request in the test partition of the target dimension, the sub-control coefficient of the target dimension is determined, where the target dimension is any one of the multiple dimensions; based on the sub-control coefficients of the multiple dimensions, the target control coefficient is determined.

[0008] Furthermore, the control coefficient of the test group is determined in the following way: based on the product display results corresponding to the access requests of the test group within the first time range, the advertising coverage of the test group within the first time range is determined, where the first time range refers to the time range between the current moment and the previous moment; based on the target advertising coverage, the advertising coverage of the test group within the first time range, and the control coefficient, the control coefficient of the test group within the second time range is determined, where the second time range refers to the time range between the current moment and the next moment.

[0009] Furthermore, multiple test partitions are divided into multiple dimensions, including: recommendation dimension, advertising dimension, and global dimension. When a test group belongs to the recommendation dimension or advertising dimension, the control coefficient of the test group in the second time period is determined based on the target advertising coverage, the advertising coverage of the test group in the first time period, and the control coefficient. This includes: determining the initial control coefficient of the test group in the second time period based on the target advertising coverage, the advertising coverage of the test group in the first time period, and the control coefficient; and determining the control coefficient of the test group in the second time period based on other test groups and the initial control coefficient of each test group in the second time period. Here, other test groups refer to test groups other than the test group in the test partition to which the test group belongs.

[0010] Furthermore, based on the target advertising coverage, the advertising coverage of the test group within the first time period, and the adjustment coefficient, determining the initial adjustment coefficient of the test group within the second time period includes: calculating the difference between the advertising coverage and the target advertising coverage to obtain a first value; determining the adjustment value based on the first value and the preset step size of the coefficient adjustment; and determining the initial adjustment coefficient of the test group within the second time period based on the adjustment value and the adjustment coefficient of the test group within the first time period.

[0011] Furthermore, based on the other test groups and their respective initial control coefficients within the second time range, determining the control coefficient of the test group within the second time range includes: calculating the product between the initial control coefficient of the other test groups within the second time range and the flow ratio of the other test groups to obtain a second value; calculating the product between the initial control coefficient of the test group within the second time range and the flow ratio of the test group to obtain a third value; and determining the control coefficient of the test group within the second time range based on the initial control coefficient, the second value, and the third value.

[0012] Furthermore, before determining the control coefficient of the test group in the second time range, the method also includes: obtaining the advertising coverage of other test groups in the first time range, wherein other test groups refer to test groups other than the test group in the test partition where the test group is located; and determining the target advertising coverage of the test group in the first time range based on the advertising coverage of other test groups in the first time range and the advertising coverage of the test group in the first time range.

[0013] Furthermore, determining the target control coefficient based on the control coefficient associated with the test group corresponding to the access request in multiple test partitions includes: obtaining configuration information, wherein the configuration information includes group information of the test group to be subject to advertising coverage control; determining the test group recorded in the configuration information from the test groups corresponding to the access request in multiple test partitions to obtain the target test group; and determining the target control coefficient based on the control coefficient associated with the target test group corresponding to the access request.

[0014] Furthermore, the method also includes: detecting the task execution status and task indicator information of the target task instance, wherein the target task instance is used to determine the control coefficient of the test group; judging whether there is an anomaly in the target task instance based on the task execution status and task indicator information; and generating a first warning message if it is determined that there is an anomaly in the target task instance.

[0015] Furthermore, the method also includes at least one of the following steps: generating a second warning message when the amount of data of the product display results is detected to be lower than a preset threshold; prohibiting the calculation of advertising coverage based on the product display results when the time difference between the collection time of the product display results and the current time is detected to be greater than a preset time length; and generating a third warning message when the difference between the average advertising coverage of the test partition and the historical reference advertising coverage of the test partition is detected to be greater than a preset difference.

[0016] According to another aspect of the embodiments of this application, a data processing method is also provided, comprising: acquiring user access requests for an information flow page uploaded by a client; determining test groups corresponding to the access requests in multiple test partitions in a cloud server, wherein the multiple test partitions include recommendation partitions and advertising partitions, and the test groups are used for group comparison tests of advertising product display and recommended product display; determining a target control coefficient based on the control coefficient associated with the test groups corresponding to the access requests in the multiple test partitions, wherein the control coefficient is determined based on the target advertising coverage of advertising products; determining the recommendation value of candidate products based on the target control coefficient, the product type of candidate products, and the recommendation value of candidate products, and determining the products to be displayed on the information flow page from multiple candidate products based on the recommendation value; and feeding back the products to be displayed to the client.

[0017] According to another aspect of the embodiments of this application, a data processing apparatus is also provided, comprising: a first determining unit, configured to, upon receiving a user's access request for an information flow page, determine a test group corresponding to the access request in multiple test partitions, wherein the multiple test partitions include a recommendation partition and an advertising partition, and the test group is used for group comparison testing of advertising product display and recommended product display; a second determining unit, configured to determine a target control coefficient based on the control coefficient associated with the test group corresponding to the access request in the multiple test partitions, wherein the control coefficient is determined based on the target advertising coverage of advertising products; and a third determining unit, configured to determine the recommendation value of candidate products based on the target control coefficient, the product type of candidate products, and the recommendation value of candidate products, and determine the product to be displayed in the information flow page from multiple candidate products based on the recommendation value.

[0018] Furthermore, the multiple test partitions are divided into multiple dimensions, and the second determining unit includes: a first determining subunit, used to determine the sub-control coefficient of the target dimension based on the control coefficient of the test group corresponding to the access request in the test partition of the target dimension, wherein the target dimension is any one of the multiple dimensions; and a second determining subunit, used to determine the target control coefficient based on the sub-control coefficients of the multiple dimensions.

[0019] Furthermore, the data processing device also includes: a fourth determining unit, used to determine the advertising coverage of the test group within a first time range based on the product display results corresponding to the access requests of the test group within a first time range, wherein the first time range refers to the time range between the current moment and the previous moment; and a fifth determining unit, used to determine the control coefficient of the test group within a second time range based on the target advertising coverage, the advertising coverage of the test group within the first time range, and the control coefficient, wherein the second time range refers to the time range between the current moment and the next moment.

[0020] Furthermore, multiple test partitions are divided into multiple dimensions, including: recommendation dimension, advertising dimension, and global dimension. When a test group belongs to the recommendation dimension or advertising dimension, the fifth determining unit includes: a third determining subunit, used to determine the initial control coefficient of the test group in the second time range based on the target advertising coverage, the advertising coverage of the test group in the first time range, and the control coefficient; and a fourth determining subunit, used to determine the control coefficient of the test group in the second time range based on other test groups and the initial control coefficient of each test group in the second time range, wherein other test groups refer to test groups other than the test group in the test partition to which the test group belongs.

[0021] Furthermore, the third determining subunit includes: a first calculation module for calculating the difference between the advertising coverage rate and the target advertising coverage rate to obtain a first value; a first determining module for determining an adjustment value based on the first value and a preset step size for coefficient adjustment; and a second determining module for determining an initial adjustment coefficient for the test group in a second time range based on the adjustment value and the adjustment coefficient of the test group in a first time range.

[0022] Furthermore, the fourth determining subunit includes: a second calculation module, used to calculate the product between the initial control coefficient of other test groups and the flow ratio of other test groups within the second time range, to obtain a second value; a third calculation module, used to calculate the product between the initial control coefficient of test groups and the flow ratio of test groups within the second time range, to obtain a third value; and a third determining module, used to determine the control coefficient of test groups within the second time range based on the initial control coefficient of test groups, the second value, and the third value.

[0023] Furthermore, the data processing device further includes: an acquisition unit, configured to acquire the advertising coverage of other test groups within a first time range, wherein other test groups refer to test groups other than the test group in the test partition where the test group is located; and a sixth determination unit, configured to determine the target advertising coverage corresponding to the test group within the first time range based on the advertising coverage of other test groups within the first time range and the advertising coverage of the test group within the first time range.

[0024] Further, the second determining unit includes: an acquisition subunit for acquiring configuration information, wherein the configuration information includes grouping information of the test group to be subject to advertising coverage adjustment; a fifth determining subunit for determining the test group recorded in the configuration information from the test groups corresponding to the access request in multiple test partitions, thereby obtaining the target test group; and a sixth determining subunit for determining the target adjustment coefficient based on the adjustment coefficient associated with the target test group corresponding to the access request.

[0025] Furthermore, the data processing device also includes: a detection unit for detecting the task execution status and task indicator information of the target task instance, wherein the target task instance is used to determine the control coefficient of the test group; a judgment unit for judging whether the target task instance has any abnormality based on the task execution status and task indicator information; and a first generation unit for generating a first warning message when it is determined that the target task instance has an abnormality.

[0026] Furthermore, the data processing device also includes at least one of the following modules: a second generation unit, used to generate a second warning message when the amount of data of the product display results is detected to be lower than a preset threshold; a processing unit, used to prohibit the calculation of advertising coverage based on the product display results when the time difference between the collection time of the product display results and the current time is detected to be greater than a preset time length; and a third generation unit, used to generate a third warning message when the difference between the average advertising coverage of the test partition and the historical reference advertising coverage of the test partition is detected to be greater than a preset difference.

[0027] According to another aspect of the present invention, an electronic device is also provided, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the data processing method described above during runtime.

[0028] According to another aspect of the present invention, a computer-readable storage medium is also provided, wherein the storage medium stores a program, wherein the program controls the device where the storage medium is located to execute any of the above-described data processing methods during runtime.

[0029] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the data processing method described above.

[0030] In this embodiment, upon receiving a user's access request for an information feed page, the system determines the test group corresponding to the access request in multiple test partitions. These test partitions include recommendation partitions and advertising partitions. The test groups are used for comparative testing of advertising and recommended product display. A target control coefficient is determined based on the control coefficient associated with the test group corresponding to the access request in the multiple test partitions. This control coefficient is determined based on the target advertising coverage of the advertising products. The recommendation value of the candidate products is determined based on the target control coefficient, the product type of the candidate products, and the recommendation value of the candidate products. Based on the recommendation value, the system determines the method of displaying the product on the information feed page from among the multiple candidate products. By setting control coefficients for the test groups in the multiple test partitions, the system determines the product recommendation based on the target control coefficient calculated from the control coefficients of the test groups. The system uses a value to determine the products to be displayed, and achieves control over the display ratio of advertised and recommended products in the test group based on a control coefficient. This controls the advertising coverage of the test group. The control coefficient is determined based on the target advertising coverage of the advertised products. The control coefficient can adjust the advertising coverage of the test group to be close to the target advertising coverage, thus ensuring that the advertising coverage of each test group remains the same or similar. This achieves the goal of balancing the advertising coverage between test groups based on the control coefficients of test groups in multiple test partitions, thereby improving the accuracy of experimental results in group comparison tests. Furthermore, it solves the technical problem that when conducting group comparison tests in scenarios with mixed advertising and recommendation displays, the ratio of advertised and recommended products displayed in different test groups is difficult to maintain consistently, thus affecting the experimental results of group comparison tests. Attached Figure Description

[0031] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0032] Figure 1 This is a schematic diagram of a computer terminal provided according to Embodiment 1 of this application;

[0033] Figure 2 This is a flowchart of the data processing method provided according to Embodiment 1 of this application;

[0034] Figure 3 This is a schematic diagram of an optional test partition provided according to Embodiment 1 of this application;

[0035] Figure 4 This is an engineering link diagram that can be independently controlled according to the optional recommended test groups provided in Embodiment 1 of this application;

[0036] Figure 5 This is an engineering flow diagram for independently controlling optional advertising test groups according to Embodiment 1 of this application;

[0037] Figure 6 This is a flowchart of the data processing method provided according to Embodiment 2 of this application;

[0038] Figure 7 This is a schematic diagram of a data processing apparatus provided according to Embodiment 3 of this application;

[0039] Figure 8 This is a structural block diagram of an electronic device provided according to Embodiment 4 of this application. Detailed Implementation

[0040] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0041] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0042] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant regions, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0043] Example 1

[0044] According to an embodiment of this application, a data processing method is also provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0045] The method embodiment provided in Embodiment 1 of this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing a data processing method is shown. Figure 1 As shown, the computer terminal (or mobile device) 10 may include a processor set 102 (the processor set 102 may include, but is not limited to, processing devices such as microprocessors (MCUs) or field-programmable gate arrays (FPGAs), and the processor set 102 may include a processor set, Figure 1 (Illustrated using 102a, 102b, ..., 102n), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may include: a display, an input / output interface (I / O interface), a Universal Serial Bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0046] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be wholly or partially embodied in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be wholly or partially integrated into any other element within the computer terminal 10 (or mobile device).

[0047] The memory 104 can be used to store software programs and modules of application software, such as program instructions / data storage devices corresponding to the data processing method in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby implementing the aforementioned data processing method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0048] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0049] The display may be a touchscreen LCD display that allows the user to interact with the user interface of the computer terminal 10 (or mobile device).

[0050] In the digital advertising ecosystem, Page View Ratio (PVR) is a key metric for balancing user experience and advertising revenue. Taking the "You May Also Like" feed on a shopping website's homepage as an example, this metric is measured by the proportion of ad impressions to total impressions. Assuming relatively stable traffic within the feed, PVR directly determines the supply of ad traffic; too high a PVR may lead to a poor user experience, while too low a PVR will negatively impact the group's advertising revenue. Therefore, ensuring the stability of PVR is not only a crucial breakthrough for technological innovation but also essential for the platform's sustainable development.

[0051] Initially, advertising placement followed a "fixed slot" model, where the specific location and proportion of ad display were predetermined. While this model ensured stability and predictability, it struggled to adapt to dynamic changes in user needs, severely hindering the optimal allocation of traffic resources. With continuous technological advancements, advertising placement shifted from a "fixed slot" model to a "dynamic slot" model. This model dynamically adjusts ad display strategies based on the commercial value of traffic, ensuring a stable overall commercialization rate while achieving market-oriented and intelligent allocation of advertising resources.

[0052] However, technological innovation is often accompanied by new challenges. "Dynamic placement," as a global optimization strategy, breaks the assumption of traffic independence in traditional group comparison tests with its cross-traffic decision-making mechanism, bringing new difficulties to the evaluation of experimental results. Currently, in the context of unified mixed ranking of advertising and recommendation, the "dynamic placement" model typically only controls the overall advertising coverage to remain near the agreed target value, while ignoring the balance at the micro level. Due to the traffic optimization effect of mixed placement, when the value of mixed placement of ads in different traffic pools differs, the system tends to pursue the maximization of overall efficiency. This leads to a "favoritism" phenomenon in the traffic allocation of test groups in group comparison tests. This imbalance manifests as differences in ad coverage in the comparative test results of some small traffic groups. In other words, it is difficult to balance the PVR of different test groups. For example, group comparison test can be understood as A / B testing. Suppose that in A / B testing, there is a baseline bucket A and a test bucket B (A and B are equivalent to different test groups). In order to ensure the reliability of the experimental results of group comparison test, it is necessary to keep the PVR of bucket A and bucket B the same (e.g., 10%). However, since the strategy adopted by bucket B is adjusted relative to bucket A (e.g., bucket B uses an optimized ad click-through rate prediction model, which can improve the ad placement effect), the PVR of bucket B may become 12%, while bucket A remains at 10%. This makes it difficult to keep the ratio of advertised products and recommended products displayed in experimental buckets A and B consistent, directly affecting the experimental results of group comparison test and leading to the problem of low accuracy of the experimental results of group comparison test.

[0053] Against the above-mentioned technical background, this application provides as follows Figure 2 The data processing method shown. Figure 2 This is a flowchart of a data processing method according to Embodiment 1 of this application. The method includes:

[0054] Step S201: Upon receiving a user's access request for the information flow page, determine the test group corresponding to the access request in multiple test partitions. The multiple test partitions include recommendation partitions and advertising partitions. The test group is used to conduct group comparison tests of advertising product display and recommended product display.

[0055] Optionally, electronic devices, application systems, servers, and other devices can be used as the execution subject of this application. In this embodiment, the target processing system is used as the execution subject to execute the data processing method.

[0056] Optionally, the information feed page is a page used to display products, and it may display multiple products. For example, the information feed page could be a "You May Also Like" page on a shopping website, with multiple products including recommended products and advertised products. Recommended products can be understood as products that users might be interested in, calculated through algorithms based on personalized data such as user interests, behavior, and browsing history. Their goal is to improve user experience, increase user dwell time on the platform, promote product sales, and thus enhance the platform's overall activity and user stickiness. Advertised products can be understood as products that merchants or brands pay to display in the information feed. Their goal is to attract user attention, promote brand exposure, and ultimately achieve sales conversion. Unlike recommended products, the display of advertised products is not entirely based on users' personal preferences but may be based on broader market positioning and advertising strategies.

[0057] Optionally, upon receiving a user's access request for an information feed page, the target processing system determines the test group corresponding to the access request in multiple test partitions. For example, when a user clicks a button such as "You May Also Like" or "Recommended" on a shopping website (or shopping application) in hopes of being redirected to the corresponding page, the front end generates a corresponding access request and sends it to the target processing system. The target processing system receives the access request and determines the test group corresponding to the access request in multiple test partitions.

[0058] Optionally, a test partition is a hierarchical division of system functions according to different strategies or purposes. Each test partition includes multiple test groups (e.g., two, six, or more) used to implement specific testing strategies to evaluate the impact of different settings on user behavior. In this embodiment, the multiple test partitions include at least a recommendation partition and an advertising partition, and there can be one or more recommendation partitions and one or more advertising partitions.

[0059] In an optional embodiment, commonly speaking, group comparison testing can also be called A / B testing, test partitions can also be called experimental stratification (recommendation partitions can also be called recommendation stratification, and advertising partitions can also be called advertising stratification), and test groups can also be called buckets. In this embodiment, the concepts of the two mentioned above can be considered equivalent to each other.

[0060] Optionally, a test group is a subdivision within a test partition. A test group can represent a specific set of parameter settings or strategy implementations. Test groups (i.e., buckets) can be divided into a first-class group (i.e., bucket A) and a second-class group (i.e., bucket B). The first-class group is equivalent to the control group in a group comparison test, using the current version or default solution as a baseline. The second-class group is equivalent to the experimental group in a group comparison test, using the new version or the solution to be tested, used to evaluate the effect. Test groups in the recommendation partition are used for group comparison tests of recommended product displays, and test groups in the advertising partition are used for group comparison tests of advertised product displays. The test content of group comparison tests includes, but is not limited to, testing recommendation strategies, recommendation value calculation strategies, and product display strategies. For example, a first-class group in the recommendation partition might use a deep learning-based interest matching algorithm as the recommendation algorithm, and a second-class group in the recommendation partition might use a collaborative filtering-based real-time recommendation algorithm. A first-class group in the advertising partition might use a geographic targeting strategy, adjusting the ad display format based on the user's geographical location, and a second-class group in the advertising partition might use a mobile device type targeting strategy, adjusting the ad display format for different device types.

[0061] Optionally, the multiple test partitions maintain an orthogonal relationship, meaning that the test partitions are independent and do not interfere with each other in the experimental design. Upon receiving a user's access request, the target processing system can precisely map the access request to test groups in multiple test partitions for processing based on random allocation or other specific strategies. Moreover, the access request is mapped to only a single test group within a single test partition, that is, it is associated with only a single test group within a single test partition, to ensure the consistency of test conditions and the accuracy of experimental results.

[0062] Step S202: Determine the target control coefficient based on the control coefficient associated with the test group corresponding to the access request in multiple test partitions, wherein the control coefficient is determined based on the target advertising coverage of the advertising product.

[0063] Optionally, at least some test groups in the recommended partition and at least some test groups in the advertising partition have corresponding adjustment coefficients. These adjustment coefficients are used to balance the PVR between test groups within the same test partition. The adjustment coefficients for test groups in different layers are the same but independent. The adjustment coefficients can be dynamically determined based on the target advertising coverage rate of the advertised product. The target advertising coverage rate refers to the expected value of the advertising coverage rate of the advertised product. Advertising coverage rate can be understood as the proportion of advertised product exposure to total product exposure, where total product exposure is the sum of advertised product exposure and recommended product exposure. For example, advertising coverage rate can be expressed as the following formula:

[0064]

[0065] Here, PVR represents ad reach, ad_pv represents ad impressions, and total_pv represents total product impressions. When a product card is loaded and displayed in the visible area of ​​a specific user's feed page, the system records an impression event for that product.

[0066] The target processing system can determine the target control coefficient based on the control coefficients associated with the test groups corresponding to the access request in multiple test partitions. For example, the control coefficients of each test group can be added together directly or calculated using other mathematical formulas to obtain the target control coefficient. Another example is that multiple test partitions are divided into multiple dimensions, including recommendation, advertising, and global dimensions. The advertising partition belongs to the advertising dimension, and the recommendation partition belongs to the recommendation dimension. Based on the control coefficient of the test group corresponding to the access request in a test partition of a certain dimension, the sub-control coefficient of that dimension is determined. Therefore, based on the sub-control coefficients of multiple dimensions, the target control coefficient is determined.

[0067] Optionally, the target control coefficient is used to determine the recommendation value of products (i.e., recommended products and advertised products). By influencing the target control coefficient through the control coefficient, the recommendation value of products can be affected, thereby affecting the display ratio of recommended products and advertised products on the information flow page. This achieves the effect of controlling the advertising coverage of the test group through the control coefficient.

[0068] Step S203: Determine the recommendation value of the candidate product based on the target control coefficient, the product type of the candidate product, and the recommendation value of the candidate product, and determine the product to be displayed on the information flow page from multiple candidate products based on the recommendation value.

[0069] In related technologies, a control variable K is typically calculated based on the aggregation of a mixed-rank test partition, and then applied to the additive ranking formula of the mixed-rank test to control the global advertising coverage in the group comparison test. The process of calculating the recommendation value of candidate products through the additive ranking formula of the mixed-rank test is as follows:

[0070] rankscore=rec_value+α*ecpm-K*I(is_ad)

[0071] Here, rankscore represents the recommendation value of candidate products calculated by relevant technologies, which determines the display position of products (including advertised products and recommended products) in the user's information flow; rec_value represents the recommendation value, reflecting the recommendation value of recommended products or the degree of user interest in recommended products; ecpm represents the expected revenue per thousand impressions of advertised products, used to measure the commercial value of advertised products (i.e., the recommendation value of advertised products); α represents the preset adjustment coefficient, used to adjust the weight of the value of advertised products in the ranking; and I(is_ad) represents the indicator function, used to determine whether the current card is an advertised product. If the current card is an advertised product, then (I(is_ad) = 1); if the current card is not an advertised product, then (I(is_ad) = 0).

[0072] However, the above method can only guarantee the global PVR for group comparison tests. It cannot support PVR balancing between groups. For example, assuming the expected PVR is 10%, in the AB test, the PVR of bucket A is 10% and the PVR of bucket B is 12%. After balancing the global PVR with K, the PVR of bucket A may become 9% and the PVR of bucket B may become 11%. In this case, although the global PVR is adjusted to 10%, the PVR of buckets A and B is not balanced, which will affect the experimental results.

[0073] In an optional embodiment, to support PVR balancing between groups in the advertising and recommendation testing partitions, a "master-slave separation" control algorithm is designed. This involves setting control coefficients for test groups in both the recommendation and advertising partitions, using three dimensions of control variables to ensure global PVR control, independent control of the recommendation test group, and independent control of the advertising test group, respectively. For example, the target processing system can calculate the recommendation value of candidate products using the following formula:

[0074] new_rankscore = rec value +α*ecpm-(K+f1(w rec )+f2(w ad ))*I(is_ad)

[0075] Where new_rankscore represents the recommended value of the candidate item, (K+f1(w rec )+f2(w ad )) is equivalent to the target control coefficient, f1(w rec ) represents the regulation function of the recommendation dimension, which is used to calculate the sub-regulation coefficients of the recommendation dimension based on the regulation coefficients of the test groups in the recommendation partition, f2(w ad() represents the regulation function for the advertising dimension, which is used to calculate the sub-regulation coefficients for the advertising dimension based on the regulation coefficients of the test groups in the advertising partition. Since a traffic stream may pass through multiple layers of experimental test groups to be regulated, the regulation coefficient for the recommendation partition is denoted as... The control coefficient for advertising zones is: in, This represents the control coefficient for the test group corresponding to the traffic in the nth recommendation partition. This represents the control coefficient of the test group corresponding to the traffic in the m-th ad partition. Optionally, the multiple test partitions also include a global partition (also known as a global layer). The global partition belongs to the global dimension and is equivalent to the mixed test partition mentioned above. The test groups in this global partition do not involve group comparison testing, but are only used to calculate the control coefficient K to control the global PVR. That is, K represents the sub-control coefficient of the global dimension calculated based on the control coefficient of the test groups in the global partition.

[0076] Optionally, the multiple candidate products are products retrieved from the product database by the target processing system based on the user's access request. After determining the candidate products, the target processing system calculates the recommendation value of the candidate products based on the above formula, and then sorts the multiple candidate products in descending order of recommendation value to obtain the sorted multiple candidate products. Thus, the candidate products ranked first (e.g., the first N, where N is a positive integer) are determined as the products to be displayed on the information flow page.

[0077] After identifying the products to be displayed, the target processing system loads product cards into the information flow page, making them visual and interactive elements for users to browse. During browsing, users may perform various interactive actions, including but not limited to clicking product cards, lingering on product cards, completing purchases, adding products to favorites, and liking products. The target processing system automatically records and analyzes this user interaction information, extracting key metrics and data corresponding to the test groups, such as click-through rate, conversion rate, and user dwell time, as the results of group comparison tests. These group comparison test results are used to evaluate the product display effectiveness under different test group strategies. Based on these results, the system can determine which test group's strategy is superior and continuously optimize related algorithms and strategies, expanding the application scope to improve user experience and platform revenue.

[0078] In this solution, by setting control coefficients for test groups in multiple test partitions, and determining the recommended value of a product based on the target control coefficient calculated from the control coefficient of the test group, the solution identifies the products to be displayed. This achieves the control of the display ratio between advertised and recommended products in the test group based on the control coefficient, i.e., controlling the advertising coverage of the test group. By setting the control coefficient based on the target advertising coverage of advertised products, the control coefficient can adjust the advertising coverage of the test group to be close to the target advertising coverage, thereby ensuring that the advertising coverage of each test group remains the same or similar. This achieves the goal of balancing the advertising coverage between test groups based on the control coefficients of test groups in multiple test partitions, thus improving the accuracy of experimental results in group comparison tests. Furthermore, it solves the technical problem that when conducting group comparison tests in scenarios where advertising and recommendation are mixed, the ratio of advertised and recommended products displayed in different test groups is difficult to maintain consistently, thus affecting the experimental results of group comparison tests.

[0079] Calculating the target control coefficient is crucial. Therefore, in the data processing method provided in Embodiment 1 of this application, multiple test partitions are divided into multiple dimensions. The target control coefficient is determined based on the control coefficient associated with the test group corresponding to the access request in the multiple test partitions. This includes: for the target dimension among the multiple dimensions, determining the sub-control coefficient of the target dimension based on the control coefficient of the test group corresponding to the access request in the test partition of the target dimension, wherein the target dimension is any one of the multiple dimensions; and determining the target control coefficient based on the sub-control coefficients of the multiple dimensions.

[0080] Optionally, the advertising partition in the multiple test partitions belongs to the advertising dimension, the recommendation partition belongs to the recommendation dimension, and the multiple test partitions also include a global partition, which belongs to the global dimension. It is equivalent to the mixed test partition mentioned above. The test groups in this global partition do not involve group comparison tests, but are only used to calculate the control coefficient K to control the global PVR.

[0081] For a target dimension among multiple dimensions, sub-control coefficients for the target dimension are determined based on the control coefficients of the test groups corresponding to the access requests in the test partition of the target dimension. For example, when the target dimension is the recommendation dimension, the sub-control coefficients for the recommendation dimension are determined based on the control function corresponding to the recommendation dimension and the control coefficients of the test groups corresponding to the access requests in the recommendation partition (i.e., the test partition of the recommendation dimension). When the target dimension is the advertising dimension, the sub-control coefficients for the advertising dimension are determined based on the control function corresponding to the advertising dimension and the control coefficients of the test groups corresponding to the access requests in the advertising partition (i.e., the test partition of the advertising dimension). When the target dimension is the global dimension, the sub-control coefficients for the global dimension are determined based on the control function corresponding to the global dimension and the control coefficients of the test groups corresponding to the access requests in the global partition (i.e., the test partition of the global dimension).

[0082] In order for the modulation coefficients of the test groups in the recommendation and advertising partitions to achieve the effect of balancing the PVR among the test groups, the modulation functions f of the recommendation and advertising partitions need to satisfy the following property constraints:

[0083] 1. The average global value score remains unchanged: This ensures that the average value score remains unchanged from a global perspective before and after regulation.

[0084] a. That is, E(new_rankscore) = E(rankscore), and therefore E(f1(w) = E(rankscore). rec ))+E(f2(w ad ))=0, where E() represents the expected value.

[0085] b. Because the internal regulation of recommendations and advertising is completely independent, for example, there is no f2(w) ad ) term, E(f1(w rec The condition )) = 0 still needs to be true.

[0086] c. Therefore, E(f1(w) rec ))=0 and E(f2(w ad ))=0, which means that the mean of the independent control variable f(w) is guaranteed to be 0.

[0087] 2. Decoupling of zone control: Ensure that PVR control between different zones does not affect each other.

[0088] a. That is, the control variable w of the currently recommended test group b (or advertising test group). rec The mean value is only related to the control coefficient w of the barrel itself. b It is relevant to, but irrelevant to, other layers. The moderating variable w for the recommended test group (or advertising test group) recIt can be understood as a set of control coefficients for the test group corresponding to the traffic of the test group in different recommendation (or advertising) partitions.

[0089] b. That is, E(f1(w) rec ))=g1(w b ), where g1() represents the control variable w rec The mean and w b The functional relationship between them is not required to examine the specific content of the function.

[0090] Optionally, Figure 3 This is a schematic diagram of an optional test partition provided according to Embodiment 1 of this application, such as... Figure 3 As shown, assume there are n test partitions, where the i-th partition contains l... i There are 1 test group, and the traffic ratio of each test group is 1. And satisfy The control coefficient for the j-th test group in the i-th partition is

[0091] Let the vector of control coefficients set of all test packets traversed by the k-th traffic (i.e., access request) be denoted as . Here This indicates the test group number into which the k-th traffic flow falls in the i-th test partition.

[0092] Based on the two aforementioned property constraints, we can determine that the following solution satisfies the formal constraints described above:

[0093] Solution 1: And for any partition i,

[0094] Solution 2: And for any partition i,

[0095] That is, both the "multi-level multiplication" and "multi-level addition" forms satisfy the above property constraints.

[0096] In an optional embodiment, a multi-level summation method is used as the control function for the recommendation and advertising partitions to calculate the sub-control coefficients for the recommendation dimension and the advertising dimension. For example, the control coefficients of the test group corresponding to the access request in the recommendation partition are summed to obtain the sub-control coefficients for the recommendation dimension; the control coefficients of the test group corresponding to the access request in the advertising partition are summed to obtain the sub-control coefficients for the advertising dimension.

[0097] In an optional embodiment, the number of global partitions is 1, and the target processing system can directly determine the control coefficient of the test group corresponding to the access request in the global partition as the sub-control coefficient of the global dimension.

[0098] Optionally, after determining the sub-control coefficients for multiple dimensions, the target processing system can sum the sub-control coefficients for multiple dimensions to obtain the target control coefficient. For example, the target control coefficient can be obtained by summing the sub-control coefficients for the global dimension, the advertising dimension, and the recommendation dimension, i.e., the (K+f1(w) above. rec )+f2(w ad )).

[0099] It should be noted that the above method enables the effective calculation of the target control coefficient, thereby improving the accuracy of PVR balancing between groups and thus improving the accuracy of experimental results in group comparison tests.

[0100] The calculation of the control coefficient is crucial. Therefore, in the data processing method provided in Embodiment 1 of this application, the control coefficient of the test group is determined in the following way: based on the product display results corresponding to the access requests of the test group within a first time range, the advertising coverage of the test group within the first time range is determined, where the first time range refers to the time range between the current moment and the previous moment; based on the target advertising coverage, the advertising coverage of the test group within the first time range, and the control coefficient, the control coefficient of the test group within the second time range is determined, where the second time range refers to the time range between the current moment and the next moment.

[0101] In an optional embodiment, the control coefficient of the test group can be dynamically adjusted based on the PID (Proportional-Integral-Derivative) algorithm, according to the real-time test group exposure data and the expected exposure effect (i.e., the target ad coverage). For example, the control coefficient required for the test group in the next time period can be determined based on the real-time test group exposure data and expected exposure effect in the previous time period, as well as the ad coverage of the test group in the previous time period.

[0102] Optionally, the real-time exposure data of the test group in the previous time period is also known as the ad coverage of the test group within the first time range. When the test group belongs to the recommendation dimension or the global dimension, the ad coverage of test group j within the first time range can be expressed as... in, This represents the total number of advertised products returned for access requests from test group j within the time range [t, t+1) (i.e., the total number of advertised product exposures). This represents the total number of products (i.e., total product exposures) returned by access requests from test group j within the time range [t, t+1). Since display advertising scenarios are extremely diverse (including "You May Also Like," post-purchase, and over 200 smaller scenarios), it's difficult to obtain a unified and stable bucket of exposure data for each scenario. Therefore, when the test group belongs to the advertising dimension, the goal of the adjustment task can be transformed from leveling PVR to leveling PVS (adPV per session, ad impressions per request), that is, representing the ad coverage of test group j within the first time range as... in, This represents the total number of access requests for test group j within the time range [t, t+1). For example, the target processing system can send the information "ad test group ID, scene ID, time window" to the system used for data detection and analysis. The system used for data detection and analysis will then provide feedback on the exposure data of the corresponding test group within the time window, i.e., the number of ad impressions and the number of requests, so that the target processing system can determine the ad coverage of the test group.

[0103] In some embodiments, the target ad coverage for different partitions (i.e., tiers) can be the same fixed value, for example, 10%.

[0104] In some embodiments, the target ad coverage differs across partitions, and the target ad coverage of a partition can be fixed or dynamically changing. For example, the adjustment coefficient for the global dimension test group is used to adjust the global PVR. Therefore, in determining the adjustment coefficient for the global dimension test group, a fixed value specified by the administrator is used as the target ad coverage, such as a fixed value agreed upon in advance with the scene operator. The adjustment coefficients for the recommendation dimension and ad dimension test groups are used to balance the PVR between test groups within the same partition, ensuring the experimental effect of the comparative test of the recommendation dimension and ad dimension groups. Therefore, in determining the adjustment coefficients for the recommendation dimension and ad dimension test groups, the target ad coverage of a partition can be dynamically determined based on the average real-time PVR data of all test groups in a partition (i.e., a layer).

[0105] Optionally, after determining the target ad coverage and the ad coverage of the test group in the first time range, the target processing system can determine the control coefficient of the test group in the second time range based on the idea of ​​the PID algorithm, according to the target ad coverage, the ad coverage of the test group in the first time range, and the control coefficient of the test group in the first time range.

[0106] It should be noted that through the above process, the adjustment coefficient of the test group is dynamically determined based on the real-time advertising exposure data and expected exposure effect of the test group. This improves the accuracy of the determined adjustment coefficient, thereby improving the PVR balancing effect between different test groups in the group comparison test and improving the accuracy of the experimental results of the group comparison test.

[0107] To more accurately determine the control coefficient of the test group, in the data processing method provided in Embodiment 1 of this application, multiple test partitions are divided into multiple dimensions, including: recommendation dimension, advertising dimension, and global dimension. When the test group belongs to the recommendation dimension or advertising dimension, the control coefficient of the test group in the second time range is determined based on the target advertising coverage, the advertising coverage of the test group in the first time range, and the control coefficient. This includes: determining the initial control coefficient of the test group in the second time range based on the target advertising coverage, the advertising coverage of the test group in the first time range, and the control coefficient; and determining the control coefficient of the test group in the second time range based on other test groups and the initial control coefficient of each test group in the second time range. Here, other test groups refer to test groups other than the test group in the test partition where the test group is located.

[0108] Since the regulation functions for the recommendation and advertising dimensions need to satisfy the two property constraints mentioned above, when using a multi-level additive form as the regulation function, it is necessary to ensure that the regulation coefficients of the test groups in the same test partition under the recommendation and advertising dimensions satisfy the zeroing condition, that is, for any partition i, we have Therefore, the target processing system can first calculate the initial control coefficient of the test group in the second time range based on the idea of ​​PID algorithm, and then update the initial control coefficient of the test group in the second time range through zeroing operation to obtain the control coefficient of the test group in the second time range, so as to satisfy the aforementioned zeroing condition, and thus satisfy the aforementioned two formal constraints.

[0109] Optionally, during the zeroing operation, the target processing system can update the initial control coefficient of the current test group based on the initial control coefficients of other test groups and the current test group in the second time range, so as to obtain the control coefficient of the test group in the second time range.

[0110] It should be noted that by adjusting the control coefficient based on the PID algorithm and combining it with the initial control coefficients of other test groups in the same partition to adjust the initial control coefficient of the current test group, the control coefficient of the test group can meet the constraints of "unchanged global value score mean" and "partition control decoupling". This effectively improves the accuracy of the determined control coefficient. It not only supports independent PVR control of test groups within the advertisement, but also supports PVR leveling of test groups within the recommendation. This avoids the need to establish and maintain complex experimental penetration relationships and can effectively improve the accuracy of the experimental results of group comparison tests.

[0111] To more accurately determine the initial control coefficient of the test group, in the data processing method provided in Embodiment 1 of this application, determining the initial control coefficient of the test group in the second time range based on the target advertising coverage, the advertising coverage of the test group in the first time range, and the control coefficient includes: calculating the difference between the advertising coverage and the target advertising coverage to obtain a first value; determining an adjustment value based on the first value and a preset step size for coefficient control; and determining the initial control coefficient of the test group in the second time range based on the adjustment value and the control coefficient of the test group in the first time range.

[0112] In an optional embodiment, the initial adjustment coefficient for the test group of the recommendation dimension in the second time range can be calculated based on the following formula:

[0113]

[0114] in, This represents the initial control coefficient for test group j within the second time range. β represents the control coefficient of test group j within the first time range, and β represents the preset step size. This represents the target ad coverage corresponding to test group j. Under the recommendation dimension, test groups within the same test partition have the same target ad coverage within the same time range. Equivalent to the first value, This is equivalent to adjusting the value.

[0115] In an optional embodiment, the initial control coefficient for the test group of the advertising dimension within the second time range can be calculated based on the following formula:

[0116]

[0117] Within the advertising dimension, test groups in the same test partition have the same target ad coverage within the same time frame. Equivalent to the first value, This is equivalent to adjusting the value.

[0118] It should be noted that, through the above process, the control coefficient of the test group is adjusted based on the deviation between the actual advertising exposure data and the expected exposure effect of the test group, thereby achieving accurate determination of the initial control coefficient.

[0119] To more accurately determine the control coefficient of the test group, in the data processing method provided in Embodiment 1 of this application, determining the control coefficient of the test group in the second time range based on other test groups and the initial control coefficient of each test group in the second time range includes: calculating the product between the initial control coefficient of other test groups in the second time range and the flow ratio of other test groups to obtain a second value; calculating the product between the initial control coefficient of the test group in the second time range and the flow ratio of the test group to obtain a third value; and determining the control coefficient of the test group in the second time range based on the initial control coefficient, the second value, and the third value.

[0120] In an optional embodiment, when the test group belongs to the recommendation dimension or the advertising dimension, the target processing system can calculate the adjustment coefficient of the test group in the second time range using the following formula:

[0121]

[0122] in, This represents the control coefficient of the j-th test group in the i-th layer within the second time range. This represents the initial control coefficient of test group j within the second time range. Let represent the product of the initial control coefficient of the b-th test group in the i-th layer and the flow ratio of the b-th test group, where b = j. This is equivalent to the third value, in the case where b≠j. Equivalent to the second value, This represents the summation of the products of all test groups and the traffic ratio in test partition i, where the j-th test group is located.

[0123] It should be noted that the above method enables the inter-layer zeroing operation of the control coefficients of the test groups, thereby achieving decoupling of the partition control and fundamentally ensuring that the control between different test partitions does not affect each other, thus improving the accuracy of the experimental results of group comparison tests.

[0124] To more accurately determine the target advertising coverage corresponding to the test group, the data processing method provided in Embodiment 1 of this application further includes, before determining the adjustment coefficient of the test group in the second time range: obtaining the advertising coverage of other test groups in the first time range, wherein other test groups refer to test groups other than the test group in the test partition where the test group is located; and determining the target advertising coverage corresponding to the test group in the first time range based on the advertising coverage of other test groups in the first time range and the advertising coverage of the test group in the first time range.

[0125] Optionally, the adjustment coefficients for the test groups in the recommendation and advertising dimensions are used to balance the PVR between test groups within the same partition, ensuring the experimental effect of the comparative test of the recommendation and advertising dimensions. Therefore, in determining the adjustment coefficients for the test groups in the recommendation and advertising dimensions, the target advertising coverage of a partition can be dynamically adjusted based on the average real-time PVR data of all test groups in that partition. For example, the target advertising coverage of the current test group within the first time range is calculated by averaging the advertising coverage of all other test groups in the same partition within the first time range. In other words, the target advertising coverage of the current test group within the first time range is calculated by averaging the advertising coverage of all test groups in the partition where the current test group is located. Furthermore, the target advertising coverage of test groups within the same recommendation partition (or advertising partition) is the same within the same time range.

[0126] Optionally, the adjustment coefficient for the global test group is used to adjust the global PVR. Therefore, in determining the adjustment coefficient for the global test group, the target ad coverage rate is determined to be a fixed value specified by the administrator, such as a fixed value agreed upon in advance with the scene operator. Furthermore, for the global dimension, the test groups in the global partition are not used for group comparison testing; that is, they do not need to balance the PVR between groups. Therefore, the aforementioned zeroing operation is unnecessary. In this case, the adjustment coefficient for the global test group in the second time range can be calculated based on the following formula:

[0127]

[0128] Among them, PVR target This represents the target ad coverage rate at the global level, and it is a fixed value.

[0129] It should be noted that, through the above method, the target ad coverage of the recommendation partition and the ad partition can be dynamically determined based on the ad coverage of the test group in the same layer, thereby improving the accuracy of the determined target ad coverage.

[0130] To enable more flexible balancing between groups, the data processing method provided in Embodiment 1 of this application determines the target control coefficient based on the control coefficient associated with the test group corresponding to the access request in multiple test partitions. This includes: obtaining configuration information, wherein the configuration information includes group information of the test group to be subject to advertising coverage control; determining the test group recorded in the configuration information from the test groups corresponding to the access request in multiple test partitions to obtain the target test group; and determining the target control coefficient based on the control coefficient associated with the target test group corresponding to the access request.

[0131] Optionally, the configuration information can be preset by the administrator and stored in the control configuration table. In practical applications, the recommendation-side operations (group comparison testing, PVR control, etc.) and advertising-side operations (group comparison testing, PVR control, etc.) of the target processing system are implemented based on different engineering links, and these engineering links are independent of each other. Therefore, the grouping information of the test groups to be controlled for ad coverage on the recommendation and advertising sides can be recorded in the independent control configuration tables. The grouping information can be test group identifiers, etc. Only when the grouping information of a certain test group is recorded in the configuration information is it determined that PVR control needs to be performed on that test group, that is, the grouping between groups in the same test partition needs to be balanced. Conversely, if the grouping information of that test group is not recorded in the configuration information, it is determined that PVR control does not need to be performed on that test group.

[0132] Therefore, the target processing system can determine the test group recorded in the configuration information from the test groups corresponding to the access request in multiple test partitions, obtain the target test group, and determine the target control coefficient based on the control coefficient associated with the target test group corresponding to the access request. In an optional embodiment, the control coefficient is set only for the target test group, that is, the control coefficient of the target test group is dynamically adjusted only based on the PID algorithm.

[0133] In an optional embodiment, the configuration information can be dynamically updated based on administrator needs. The configuration information is updated based on the administrator's target configuration request, which is a configuration request approved by an auditor. The configuration request includes updated information for the configuration information, such as adding or deleting test group information. For example, for global control tasks, which are deployed long-term in a global partition and typically require no changes to ensure the stability and consistency of the overall (i.e., global) PVR, for independent control needs of recommended test groups, the administrator can submit an independent control application through the recommendation job platform in the target processing system. After approval, the administrator manually registers the test group bucket number in the control configuration table so that the offline control end and the online consumer end can automatically detect configuration changes and respond accordingly. For independent control needs of advertising test groups, the administrator can directly click the "Advertising Coverage Control" button on the front-end page of the advertising job platform in the target processing system to initiate an independent control application. After approval, all necessary information is automatically synchronized to the offline data task and the online configuration table. The entire process can be completed seamlessly without additional manual intervention. By independently controlling all new or modified recommendation / advertising test group requests, and having them approved online before being officially launched, not only has the transparency and controllability of the entire process been improved, but the collaboration efficiency between the operators and the maintenance team has also been greatly enhanced.

[0134] It should be noted that the above method can be used to control the advertising coverage of specific test groups in the test partition, thereby improving the flexibility and applicability of this application.

[0135] In an alternative embodiment, the following can be employed: Figure 4 The diagram shown illustrates the offline control and online consumption of the control coefficients for the test group in the recommended partition. Figure 4 This is an engineering link diagram that can be independently controlled according to the optional recommended test groups provided in Embodiment 1 of this application, such as... Figure 4As shown, the recommendation side includes offline control logic and online consumption logic. The offline control logic can be divided into two core components: the feature parsing task component and the offline control task component. The target processing system can obtain target logs generated based on online real-time data streams. The target logs include at least the relevant information of the products to be displayed to the user fed back by the target processing system, and may also include user interaction information on the displayed products (e.g., clicks, purchases, etc.). The target logs are transmitted to the feature parsing task component, which is responsible for periodically parsing the target logs to obtain the exposure data of advertising products and the total exposure data of products. It also determines the test groups to be subject to advertising coverage control from the control configuration table, and aggregates the parsed data according to the granularity of the test groups (e.g., determining the relevant exposure data of the test groups to be subject to advertising coverage control), and transmits the processed data downstream in units of partitions, triggering the execution of subsequent control tasks. The offline control task component is activated after receiving a message from the feature parsing service. This module is responsible for a series of control activities, including but not limited to data integrity verification, PID algorithm control (i.e., calculating the initial control coefficient), inter-layer zeroing processing (i.e., calculating the control coefficient based on the initial control coefficient), database writing, and control log persistence (e.g., ...). Figure 4 Operations such as writing work information into the data management platform, etc., allow the offline control task component to determine the test group information that needs PVR leveling from the control configuration table during operation. Figure 4 The data management platform in the system is used for real-time data change capture and distribution.

[0136] like Figure 4 As shown, the recommendation side consumes the control coefficients written in the database online through the advertising fusion prediction service. The advertising fusion prediction service can determine the recommendation test groups to be controlled by PVR from the control configuration table of the recommendation side, determine the control coefficients of these test groups from the database, and then calculate the recommendation value of candidate products based on this information. Based on the recommendation value, the service determines the products to be displayed on the information flow page from multiple candidate products to be displayed to users, and then obtains the group comparison test results based on user feedback information.

[0137] It should be noted that the above engineering process enables independent adjustment capabilities based on recommended test groups, directly supporting PVR balance for the corresponding recommended test groups without relying on cumbersome ad experiment penetration. Furthermore, since ad exposure calculations under the recommendation caliber are complex, and adjustment requirements vary across different scenarios (e.g., the "You May Like" section on the homepage requires separate adjustments for page 0 and page turning), breaking down the entire adjustment process into more granular functional modules not only improves the system's flexibility but also enhances its maintainability and scalability.

[0138] In an alternative embodiment, the following can be employed: Figure 5 The diagram shown illustrates the offline control and online consumption of the control coefficients for test groups within the advertising partition. Figure 5 This is an engineering link diagram for independently controlling optional advertising test groups provided in Embodiment 1 of this application, such as... Figure 5 As shown, the advertising side also includes offline control logic and online consumption logic. The offline control logic includes an offline control task component. Administrators can configure the control configuration table on the advertising operation platform, and then the advertising operation platform will synchronize the control configuration table to the database of the data management platform. The offline control task can obtain configuration information from the data management platform, and then send a request to the system used for data detection and analysis based on the configuration information to request the relevant exposure data of the test group. For example, the system can send the information "ad test group ID, scene ID, time window" to the system used for data detection and analysis through the data interface, and then the system used for data detection and analysis will return the exposure data of the corresponding test group within the time window, that is, the number of ad exposures and the number of requests. The aforementioned data detection and analysis system can be understood as a system that at least collects data displayed on the front end. Thanks to the design of this interface function, the offline control on the advertising side does not need to be aware of the cumbersome exposure data caliber and aggregation calculation process. It only needs to deploy an offline control task on the data management platform and periodically request the data interface of the aforementioned data detection and analysis system. In this way, the upstream data source upon which the adjustment task depends is centralized within the advertising system, which not only reduces the burden on the system's workflow and avoids redundant development work, but also gives the system better scalability for different scenarios. After obtaining the relevant exposure data of the test group, the offline adjustment task performs a series of adjustment activities, including but not limited to data integrity verification, PID algorithm adjustment (i.e., calculating the initial adjustment coefficient), inter-layer zeroing processing (i.e., calculating the adjustment coefficient based on the initial adjustment coefficient), database writing, and setting adjustment logs to disk (e.g., ...). Figure 5 Operations such as writing work information into the data management platform.

[0139] like Figure 5As shown, the advertising side consumes the control coefficients written in the database online through the display advertising engine, mixed display service, and advertising fusion prediction service. The display advertising engine can determine the advertising test groups to be controlled by PVR from the control configuration table, determine the control coefficients of these test groups from the database, and then calculate the sub-control coefficients of candidate products in the advertising dimension based on this information. Then, the sub-control coefficients of the advertising dimension are sent to the advertising fusion prediction service through the mixed display service. The advertising fusion prediction service calculates the recommendation value based on the sub-control coefficients of the advertising dimension, the sub-control coefficients of the recommendation dimension, and the sub-control coefficients of the global dimension. Based on the recommendation value, it determines the products to be displayed on the information flow page from multiple candidate products to be displayed to users. Finally, the group comparison test results are obtained based on user feedback information.

[0140] Optionally, the target processing system can calculate and consume the control coefficients of the test groups in the global dimension based on the same engineering link as the independent control engineering link of the recommended test groups, so it will not be elaborated here.

[0141] It should be noted that, through the above method, the PVR control module, which has been maintained by the algorithm team themselves, has been completely decoupled from the operational logic and engineering logic of PVR control in this system upgrade, thereby enhancing the system robustness of the control link.

[0142] To implement the detection and alarm mechanism, the data processing method provided in Embodiment 1 of this application further includes: detecting the task execution status and task indicator information of the target task instance, wherein the target task instance is used to determine the control coefficient of the test group; judging whether the target task instance is abnormal based on the task execution status and task indicator information; and generating a first warning message when it is determined that the target task instance is abnormal.

[0143] In this embodiment, when performing offline control tasks, job instances (i.e., target task instances) are generated at the partition level to calculate control coefficients for each partition. Therefore, the target processing system can detect the task execution status and task indicator information of the target task instances and determine whether there are any anomalies based on the task execution status and task indicator information. For example, the task execution status includes, but is not limited to, running, completed, failed, canceled, etc., and the task indicator information includes, but is not limited to, task execution time (e.g., start time, completion time, etc.), resource usage (e.g., CPU utilization, memory usage, etc.), and the number of retries (e.g., if the task encounters an error during execution, it may need to be retried).

[0144] The target processing system has pre-defined target detection rules, which include conditions that must be met to indicate that a target task instance is abnormal. The system can determine whether a target task instance is abnormal based on these rules, task execution status, and task metrics. For example, one rule in the target detection rules could be: "If a task exits abnormally M times consecutively, it is determined that the target task instance executing that task is abnormal." Optionally, if an abnormality is determined in a target task instance, a first warning message is generated. This first warning message includes instance information of the abnormal target task instance and the abnormality details, for relevant administrators to handle. Optionally, if no abnormality is determined in a target task instance, no first warning message is generated.

[0145] For example, when an offline control task malfunctions, the system can automatically send an alarm message to the administrator communication group. Furthermore, to prevent information overload, a mechanism can be introduced to trigger an alarm only after M consecutive abnormal task exits. The target processing system can also be configured with an automatic inspection service, which automatically checks the task execution status every half hour. If a task is found to be unresponsive for an extended period, an alarm is proactively triggered. The target processing system can also continuously detect and report any significant changes in the PVR value, enabling real-time year-on-year and month-on-month PVR monitoring.

[0146] In an optional embodiment, the target processing system can be configured with a real-time monitoring dashboard to allow users to quickly browse task status and key metrics. In addition, the aforementioned log system allows administrators to quickly query and adjust records for individual cases to quickly locate problems, and the control logs can be written to disk to support log analysis over a longer time span and in more dimensions.

[0147] It should be noted that the above method has enabled the automation of the detection and alarm mechanism, improved the reliability of the calculated control coefficient, and thus improved the reliability of the experimental results of the group comparison test.

[0148] To implement a fallback and disaster recovery mechanism, the data processing method provided in Embodiment 1 of this application further includes at least one of the following steps: generating a second warning message when the amount of data of the product display results is detected to be lower than a preset threshold; prohibiting the calculation of advertising coverage based on the product display results when the time difference between the collection time of the product display results and the current time is detected to be greater than a preset time length; and generating a third warning message when the difference between the average advertising coverage of the test partition and the historical reference advertising coverage of the test partition is detected to be greater than a preset difference.

[0149] Optionally, the target processing system can detect whether the amount of data corresponding to the product display results of the access requests of the test group within the first time range is lower than a preset threshold. If the amount of data collected within a specified time window is lower than the preset threshold (e.g., less than half of the time window), it is considered that this data is insufficient as a reliable basis for control. In this case, the target processing system can generate a second warning message, which indicates that the data accumulation time of the collected product display results is insufficient, for administrator processing. Conversely, if the amount of data of the product display results is not lower than the preset threshold, no second warning message is generated.

[0150] Optionally, the target processing system can detect whether the time difference between the collection time of the product display result and the current time is greater than a preset time length. If the data whose collection time differs from the current system time by more than the preset time length (e.g., the length of a complete time window, i.e., the time length between two adjacent moments), these product display results are identified as expired data, and the calculation of advertising coverage based on these product display results is prohibited. Conversely, if the time difference between the collection time of the product display result and the current time is less than or equal to the preset time length, the calculation of advertising coverage based on these product display results is allowed. This achieves the timeliness check of the data.

[0151] Optionally, to ensure the data quality meets expected standards, the target processing system can detect whether the difference between the average ad coverage rate (PVR) of the test partition and the historical reference ad coverage rate of the test partition is greater than a preset difference. This involves comparing the current average ad coverage rate of the test partition with its average ad coverage rate in the previous time period. The average ad coverage rate of the test partition is calculated by averaging the ad coverage rates of the test groups within the test partition. If the difference between the average PVR and the historical data for a partition is found to be greater than the preset difference (e.g., deviating from the upper or lower limits of the historical data (e.g., a difference percentage exceeding 50%)), an alarm is triggered, a third warning message is generated, and further operations using this suspicious data are prohibited. The third warning message includes at least the partition information of the test partition for administrator processing. Conversely, if the difference between the average ad coverage rate and the historical reference ad coverage rate of the test partition is less than or equal to the preset difference, no third warning message is generated.

[0152] It should be noted that the above methods enhance the ability to identify and process abnormal data streams during the control process, especially by implementing automatic filtering and circuit breaking measures for dirty data. This effectively prevents control failures caused by external factors such as upstream data delays, ensures the continuity and reliability of the overall service, and improves the accuracy of experimental results in group comparison tests.

[0153] In this embodiment, by setting control coefficients for test groups in multiple test partitions, and determining the recommended value of a product based on the target control coefficient calculated from the control coefficient of the test group, the product to be displayed is determined. This achieves the control of the display ratio between advertising and recommended products in the test group based on the control coefficient, that is, controlling the advertising coverage of the test group. By setting the control coefficient based on the target advertising coverage of advertising products, the control coefficient can adjust the advertising coverage of the test group to be close to the target advertising coverage, thereby ensuring that the advertising coverage of each test group remains the same or similar. This achieves the purpose of balancing the advertising coverage between test groups based on the control coefficients of test groups in multiple test partitions, thereby improving the accuracy of the experimental results of group comparison tests. This solves the technical problem that when conducting group comparison tests in scenarios where advertising and recommendation are mixed, the ratio between advertising and recommended products displayed in different test groups is difficult to maintain, thus affecting the experimental results of group comparison tests.

[0154] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0155] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of this application.

[0156] Example 2

[0157] According to embodiments of this application, a data processing method is also provided, such as... Figure 6 As shown, the method includes:

[0158] Step S601: Obtain the user's access request for the information flow page uploaded by the client.

[0159] Step S602: In the cloud server, determine the test group corresponding to the access request in multiple test partitions, including recommendation partitions and advertising partitions. The test group is used to conduct group comparison tests of advertising product display and recommended product display. Based on the control coefficient associated with the test group corresponding to the access request in multiple test partitions, determine the target control coefficient, wherein the control coefficient is determined based on the target advertising coverage of advertising products. Based on the target control coefficient, the product type of the candidate product, and the recommendation value of the candidate product, determine the recommendation value of the candidate product, and determine the product to be displayed on the information flow page from multiple candidate products based on the recommendation value.

[0160] Step S603: Feedback the products to be displayed to the client.

[0161] In this embodiment, by setting control coefficients for test groups in multiple test partitions, and determining the recommended value of a product based on the target control coefficient calculated from the control coefficient of the test group, the product to be displayed is determined. This achieves the control of the display ratio between advertising and recommended products in the test group based on the control coefficient, that is, controlling the advertising coverage of the test group. By setting the control coefficient based on the target advertising coverage of advertising products, the control coefficient can adjust the advertising coverage of the test group to be close to the target advertising coverage, thereby ensuring that the advertising coverage of each test group remains the same or similar. This achieves the purpose of balancing the advertising coverage between test groups based on the control coefficients of test groups in multiple test partitions, thereby improving the accuracy of the experimental results of group comparison tests. This solves the technical problem that when conducting group comparison tests in scenarios where advertising and recommendation are mixed, the ratio between advertising and recommended products displayed in different test groups is difficult to maintain, thus affecting the experimental results of group comparison tests.

[0162] The specific methods for data processing on the cloud server are the same as those in Example 1, and will not be repeated here.

[0163] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0164] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of this application.

[0165] Example 3

[0166] According to embodiments of this application, a data processing apparatus for implementing the above-described data processing method is also provided, such as... Figure 7 As shown, the device includes: a first determining unit 701, a second determining unit 702, and a third determining unit 703.

[0167] The first determining unit 701 is used to determine the test group corresponding to the access request in multiple test partitions when receiving a user's access request for an information flow page. The multiple test partitions include a recommendation partition and an advertising partition. The test group is used to conduct a group comparison test of advertising product display and recommended product display.

[0168] The second determining unit 702 is used to determine the target control coefficient based on the control coefficient associated with the test group corresponding to the access request in multiple test partitions, wherein the control coefficient is determined based on the target advertising coverage of the advertising product.

[0169] The third determining unit 703 is used to determine the recommended value of the candidate product based on the target control coefficient, the product type of the candidate product, and the recommended value of the candidate product, and to determine the product to be displayed on the information flow page from multiple candidate products based on the recommended value.

[0170] In the data processing apparatus provided in Embodiment 3 of this application, when the first determining unit 701 receives a user's access request for an information flow page, it determines the test group corresponding to the access request in multiple test partitions. The multiple test partitions include a recommendation partition and an advertising partition. The test group is used to conduct a group comparison test of advertising product display and recommended product display. The second determining unit 702 determines a target control coefficient based on the control coefficient associated with the test group corresponding to the access request in the multiple test partitions. The control coefficient is determined based on the target advertising coverage of advertising products. The third determining unit 703 determines the recommendation value of the candidate products based on the target control coefficient, the product type of the candidate products, and the recommendation value of the candidate products. Based on the recommendation value, it determines the product to be displayed on the information flow page from the multiple candidate products. In this solution, by setting control coefficients for test groups in multiple test partitions, and determining the recommended value of a product based on the target control coefficient calculated from the control coefficient of the test group, the solution identifies the products to be displayed. This achieves the control of the display ratio between advertised and recommended products in the test group based on the control coefficient, i.e., controlling the advertising coverage of the test group. By setting the control coefficient based on the target advertising coverage of advertised products, the control coefficient can adjust the advertising coverage of the test group to be close to the target advertising coverage, thereby ensuring that the advertising coverage of each test group remains the same or similar. This achieves the goal of balancing the advertising coverage between test groups based on the control coefficients of test groups in multiple test partitions, thus improving the accuracy of experimental results in group comparison tests. Furthermore, it solves the technical problem that when conducting group comparison tests in scenarios where advertising and recommendation are mixed, the ratio of advertised and recommended products displayed in different test groups is difficult to maintain consistently, thus affecting the experimental results of group comparison tests.

[0171] Optionally, in the data processing apparatus provided in Embodiment 3 of this application, multiple test partitions are divided into multiple dimensions, and the second determining unit includes: a first determining subunit, used to determine the sub-control coefficient of the target dimension based on the control coefficient of the test group corresponding to the access request in the test partition of the target dimension, wherein the target dimension is any one of the multiple dimensions; and a second determining subunit, used to determine the target control coefficient based on the sub-control coefficients of the multiple dimensions.

[0172] Optionally, in the data processing apparatus provided in Embodiment 3 of this application, the data processing apparatus further includes: a fourth determining unit, configured to determine the advertising coverage of the test group within a first time range based on the product display results corresponding to the access requests of the test group within a first time range, wherein the first time range refers to the time range between the current moment and the previous moment; and a fifth determining unit, configured to determine the control coefficient of the test group within a second time range based on the target advertising coverage, the advertising coverage of the test group within the first time range, and the control coefficient, wherein the second time range refers to the time range between the current moment and the next moment.

[0173] Optionally, in the data processing apparatus provided in Embodiment 3 of this application, multiple test partitions are divided into multiple dimensions, including: recommendation dimension, advertising dimension, and global dimension. When a test group belongs to the recommendation dimension or advertising dimension, the fifth determining unit includes: a third determining subunit, used to determine the initial control coefficient of the test group in the second time range based on the target advertising coverage, the advertising coverage of the test group in the first time range, and the control coefficient; and a fourth determining subunit, used to determine the control coefficient of the test group in the second time range based on other test groups and the initial control coefficient of each test group in the second time range, wherein other test groups refer to test groups other than the test group in the test partition where the test group is located.

[0174] Optionally, in the data processing apparatus provided in Embodiment 3 of this application, the third determining subunit includes: a first calculation module, used to calculate the difference between the advertising coverage rate and the target advertising coverage rate to obtain a first value; a first determining module, used to determine an adjustment value based on the first value and a preset step size for coefficient adjustment; and a second determining module, used to determine an initial adjustment coefficient for the test group in a second time range based on the adjustment value and the adjustment coefficient of the test group in a first time range.

[0175] Optionally, in the data processing apparatus provided in Embodiment 3 of this application, the fourth determining subunit includes: a second calculation module, used to calculate the product between the initial control coefficient of other test groups in the second time range and the flow ratio of other test groups to obtain a second value; a third calculation module, used to calculate the product between the initial control coefficient of test groups in the second time range and the flow ratio of test groups to obtain a third value; and a third determining module, used to determine the control coefficient of test groups in the second time range based on the initial control coefficient of test groups, the second value, and the third value.

[0176] Optionally, in the data processing apparatus provided in Embodiment 3 of this application, the data processing apparatus further includes: an acquisition unit, configured to acquire the advertising coverage of other test groups within a first time range, wherein other test groups refer to test groups other than the test group in the test partition where the test group is located; and a sixth determination unit, configured to determine the target advertising coverage corresponding to the test group within the first time range based on the advertising coverage of other test groups within the first time range and the advertising coverage of the test group within the first time range.

[0177] Optionally, in the data processing apparatus provided in Embodiment 3 of this application, the second determining unit includes: an acquisition subunit, used to acquire configuration information, wherein the configuration information includes grouping information of test groups to be subject to advertising coverage regulation; a fifth determining subunit, used to determine the test group recorded in the configuration information from the test groups corresponding to the access request in multiple test partitions, and obtain the target test group; and a sixth determining subunit, used to determine the target regulation coefficient based on the regulation coefficient associated with the target test group corresponding to the access request.

[0178] Optionally, in the data processing apparatus provided in Embodiment 3 of this application, the data processing apparatus further includes: a detection unit, used to detect the task execution status and task indicator information of the target task instance, wherein the target task instance is used to determine the control coefficient of the test group; a judgment unit, used to judge whether the target task instance is abnormal based on the task execution status and task indicator information; and a first generation unit, used to generate a first warning information when it is determined that the target task instance is abnormal.

[0179] Optionally, in the data processing apparatus provided in Embodiment 3 of this application, the data processing apparatus further includes at least one of the following modules: a second generation unit, configured to generate a second warning message when the amount of data of the product display results is detected to be lower than a preset threshold; a processing unit, configured to prohibit the calculation of advertising coverage based on the product display results when the time difference between the collection time point of the product display results and the current time is detected to be greater than a preset time length; and a third generation unit, configured to generate a third warning message when the difference between the average advertising coverage of the test partition and the historical reference advertising coverage of the test partition is detected to be greater than a preset difference.

[0180] It should be noted that the first determining unit 701, the second determining unit 702, and the third determining unit 703 mentioned above correspond to steps S201 to S203 in Embodiment 1. The instances and application scenarios implemented by the above units and corresponding steps are the same, but are not limited to the content disclosed in Embodiment 1. It should be noted that the above modules, as part of the device, can run in the computer terminal 10 provided in Embodiment 1.

[0181] It should be noted that the preferred implementation schemes involved in the above embodiments of this application are the same as the schemes, application scenarios and implementation processes provided in Embodiment 1, but are not limited to the schemes provided in Embodiment 1.

[0182] Example 4

[0183] Embodiments of this application may provide an electronic device, which may be any one of a group of electronic devices. Optionally, in this embodiment, the aforementioned electronic device may also be replaced by a terminal device such as a mobile terminal.

[0184] Optionally, in this embodiment, the aforementioned electronic device may be located in at least one of a plurality of network devices in a computer network.

[0185] In this embodiment, the electronic device can execute the program code corresponding to the steps in the data processing method provided in any of the above method embodiments.

[0186] Optionally, Figure 8 This is a structural block diagram of an electronic device according to an embodiment of this application. Figure 8 As shown, the electronic device 80 may include: one or more ( Figure 8 (Only one is shown) Processor 802 and memory 804. The electronic device 80 may also include a memory controller to control and manage the memory 804; the electronic device 80 may also include a peripheral interface to connect to a radio frequency module, an audio module, and a display screen, etc.

[0187] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the data processing method and apparatus in this application embodiment. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby implementing the aforementioned data processing method. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to the terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0188] The processor can access the information and application programs stored in the memory via the transmission device to execute the program code corresponding to the steps in the data processing method provided in any of the above method embodiments.

[0189] Those skilled in the art will understand that Figure 8The structure shown is for illustrative purposes only. Electronic devices can also be smartphones (such as Android phones, iOS phones, etc.), tablets, PDAs, mobile internet devices (MIDs), PADs, and other terminal devices. Figure 8 This does not limit the structure of the aforementioned electronic device. For example, electronic device 80 may also include components that are more... Figure 8 The more or fewer components shown (such as network interfaces, display devices, etc.), or having the same Figure 8 The different configurations shown.

[0190] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0191] Example 5

[0192] Embodiments of this application also provide a computer-readable storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the data processing method provided in Embodiment 1.

[0193] Optionally, in this embodiment, the storage medium may be located in any electronic device in a group of electronic devices in a computer network, or in any mobile terminal in a group of mobile terminals.

[0194] Example 6

[0195] Embodiments of this application also provide a computer program product. Optionally, in this embodiment, the computer program product may include a computer program that, when executed by a processor, implements the data processing method provided in Embodiment 1.

[0196] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0197] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0198] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of units or modules may be electrical or other forms.

[0199] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0200] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0201] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0202] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A data processing method, characterized in that, include: Upon receiving a user's access request for an information feed page, the test group corresponding to the access request in multiple test partitions is determined. The multiple test partitions include a recommendation partition and an advertising partition. The test group is used to conduct a group comparison test of advertising product display and recommended product display. Based on the control coefficient associated with the test group corresponding to the access request in multiple test partitions, a target control coefficient is determined, wherein the control coefficient is determined based on the target advertising coverage of the advertising product. The recommended value of the candidate product is determined based on the target control coefficient, the product type of the candidate product, and the recommended value of the candidate product. Based on the recommended value, the product to be displayed on the information flow page is determined from multiple candidate products.

2. The method according to claim 1, characterized in that, The multiple test partitions are divided into multiple dimensions. Based on the control coefficient associated with the test group corresponding to the access request in the multiple test partitions, the target control coefficient is determined, including: For the target dimension among the multiple dimensions, the sub-control coefficient of the target dimension is determined based on the control coefficient of the test group corresponding to the test partition of the target dimension in which the access request is located, wherein the target dimension is any one of the multiple dimensions. The target control coefficient is determined based on the sub-control coefficients of the multiple dimensions.

3. The method according to claim 1, characterized in that, The control coefficients for the test groups are determined in the following way: Based on the product display results corresponding to the access requests of the test group within the first time range, the advertising coverage of the test group within the first time range is determined, wherein the first time range refers to the time range between the current moment and the previous moment. Based on the target advertising coverage, the advertising coverage of the test group within the first time range, and the adjustment coefficient, the adjustment coefficient of the test group within the second time range is determined, wherein the second time range refers to the time range between the current moment and the next moment.

4. The method according to claim 3, characterized in that, The multiple test partitions are divided into multiple dimensions, including: recommendation dimension, advertising dimension, and global dimension. When a test group belongs to the recommendation dimension or the advertising dimension, the adjustment coefficient of the test group in the second time range is determined based on the target advertising coverage, the advertising coverage of the test group in the first time range, and the adjustment coefficient. This includes: Based on the target advertising coverage, the advertising coverage of the test group within the first time range, and the adjustment coefficient, the initial adjustment coefficient of the test group within the second time range is determined; Based on the other test groups and the initial control coefficients of each test group within the second time range, the control coefficients of the test group within the second time range are determined, wherein the other test groups refer to test groups other than the test group in the test partition where the test group is located.

5. The method according to claim 4, characterized in that, Based on the target ad coverage, the ad coverage of the test group within the first time range, and the adjustment coefficient, determining the initial adjustment coefficient of the test group within the second time range includes: Calculate the difference between the advertising coverage rate and the target advertising coverage rate to obtain a first value; Based on the first value and the preset step size of the coefficient adjustment, the adjustment value is determined; Based on the adjusted value and the control coefficient of the test group within the first time range, the initial control coefficient of the test group within the second time range is determined.

6. The method according to claim 4, characterized in that, Determining the control coefficient of the test group within the second time range based on other test groups and the initial control coefficient of each test group within the second time range includes: The second value is obtained by multiplying the initial control coefficient of the other test groups within the second time range with the flow ratio of the other test groups. The third value is obtained by multiplying the initial control coefficient of the test group within the second time range with the flow ratio of the test group. Based on the initial control coefficient of the test group, the second value, and the third value, the control coefficient of the test group within the second time range is determined.

7. The method according to claim 3, characterized in that, Before determining the modulation coefficient of the test group within the second time range, the method further includes: Obtain the advertising coverage of other test groups within the first time range, wherein the other test groups refer to test groups other than the test group in the test partition where the test group is located; Based on the advertising coverage of the other test groups within the first time range and the advertising coverage of the test group within the first time range, the target advertising coverage of the test group within the first time range is determined.

8. The method according to claim 1, characterized in that, Based on the control coefficient associated with the test group corresponding to the access request in multiple test partitions, the target control coefficient is determined as follows: Obtain configuration information, wherein the configuration information includes grouping information of the test group to be subject to advertising coverage adjustment; The target test group is obtained by determining the test group recorded in the configuration information from the test groups corresponding to the access request in multiple test partitions; The target control coefficient is determined based on the control coefficient associated with the target test group corresponding to the access request.

9. The method according to claim 1, characterized in that, The method further includes: The task execution status and task indicator information of the target task instance are detected, wherein the target task instance is used to determine the control coefficient of the test group; Based on the task execution status and the task indicator information, determine whether the target task instance is abnormal; If an anomaly is determined in the target task instance, a first warning message is generated.

10. The method according to claim 3, characterized in that, The method further includes at least one of the following steps: If the amount of data in the product display result is detected to be lower than a preset threshold, a second warning message is generated; If the time difference between the collection time of the product display result and the current time is greater than a preset time length, it is prohibited to calculate the advertising coverage based on the product display result; If the difference between the average ad coverage of the test partition and the historical reference ad coverage of the test partition is greater than a preset difference, a third warning message is generated.

11. A data processing method, characterized in that, include: Obtain user access requests for the information feed page uploaded by the client; In the cloud server, the test group corresponding to the access request in multiple test partitions is determined, wherein the multiple test partitions include recommendation partitions and advertising partitions, and the test group is used for group comparison testing of advertising product display and recommended product display; based on the control coefficient associated with the test group corresponding to the access request in the multiple test partitions, a target control coefficient is determined, wherein the control coefficient is determined based on the target advertising coverage of advertising products; based on the target control coefficient, the product type of the candidate product, and the recommendation value of the candidate product, the recommendation value of the candidate product is determined, and based on the recommendation value, the product to be displayed on the information flow page is determined from multiple candidate products; The product to be displayed is then sent back to the client.

12. A data processing apparatus, characterized in that, include: The first determining unit is configured to, upon receiving a user's access request for an information flow page, determine the test group corresponding to the access request in multiple test partitions, wherein the multiple test partitions include a recommendation partition and an advertising partition, and the test group is used to conduct a group comparison test of advertising product display and recommended product display; The second determining unit is used to determine a target control coefficient based on the control coefficient associated with the test group corresponding to the access request in multiple test partitions, wherein the control coefficient is determined based on the target advertising coverage of the advertising product. The third determining unit is used to determine the recommended value of the candidate product based on the target control coefficient, the product type of the candidate product, and the recommended value of the candidate product, and to determine the product to be displayed on the information flow page from multiple candidate products based on the recommended value.

13. An electronic device, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the data processing method of any one of claims 1 to 11.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is executed, it controls the device on which the storage medium is located to perform the data processing method of any one of claims 1 to 11.

15. A computer program product, characterized in that, It includes a computer program or instructions that, when executed by a processor, implement the data processing method according to any one of claims 1 to 11.

Images