Resource recommendation method, electronic device, and storage medium

Abstract

The application relates to the computer technical field and provides a resource recommendation method, which comprises the following steps: selecting a first quantity of to-be-recommended resources; distributing the to-be-recommended resources to corresponding recommendation positions for recommendation and starting to calculate a first recommendation duration; determining to-be-recommended resources that have completed recommendation as target resources; determining a unit-time exposure income value corresponding to each target resource; sequencing each target resource based on the income value to obtain a first sequence; and in the case that the first recommendation duration reaches a first duration threshold, re-distributing the recommendation positions of each target resource based on the first sequence, starting to calculate the first recommendation duration again, and repeatedly executing the step of determining the target resources, so that the problem that the promotion effect is poor when information is put in based on user types can be solved; since the recommendation positions are distributed to the target resources based on the unit-time exposure income, the promotion effect can be improved. In addition, an electronic device and a storage medium are also provided.

Description

Technical Field

[0001] This application relates to the field of computer technology, specifically to a resource recommendation method, electronic device, and storage medium. Background Technology

[0002] With the continuous development of internet technology, information promotion channels have gradually shifted from offline to online. How to conduct targeted content operation and advertising is a question that platform operators need to consider.

[0003] In related technologies, users are classified by analyzing user data such as preferences and behavioral patterns, target users are located, and then targeted information is delivered based on the type of target user.

[0004] However, the variability and uncontrollability of users have a significant impact on the accuracy of user analysis data. This can lead to problems such as low accuracy in recommendation placement allocation and poor promotion results when information is delivered based on user type. Summary of the Invention

[0005] To help improve the poor promotion effect when delivering information based on user type, this application provides a resource recommendation method, an electronic device, and a storage medium.

[0006] Firstly, this application provides a resource recommendation method, which adopts the following technical solution:

[0007] A resource recommendation method, the method comprising:

[0008] Select the first number of resources to be recommended;

[0009] The resources to be recommended are assigned to the corresponding recommendation slots for recommendation, and the first recommendation duration is calculated.

[0010] The resource to be recommended that has been recommended is identified as the target resource.

[0011] Determine the revenue per unit of exposure for each of the target resources;

[0012] Based on the unit exposure revenue value corresponding to the target resource, each target resource is sorted to obtain a first sequence;

[0013] If it is determined that the first recommendation duration has reached the first duration threshold, recommendation positions are reallocated for each of the target resources based on the first sequence, the first recommendation duration is recalculated, and the step of determining the target resources is repeated.

[0014] By adopting the above technical solution, the problem of poor promotion effect when information is delivered based on user type can be solved. Since the recommendation slots are allocated to target resources based on revenue per unit of exposure, the impact of the accuracy of user analysis data on the allocation of recommendation slots can be avoided. Therefore, the accuracy of recommendation slot allocation can be improved, and the promotion effect can be enhanced.

[0015] Optionally, the step of reallocating recommendation bits for each of the target resources based on the first sequence includes:

[0016] The order in which each of the target resources is located in the recommendation positions is determined to obtain the second sequence;

[0017] The recommendation slots are reallocated for each of the target resources based on the first sequence and the second sequence.

[0018] By adopting the above technical solution, recommendation positions are assigned to each target resource based on the ranking of the target resources and the ranking of the recommendation positions corresponding to the target resources. Therefore, high-quality recommendation positions can be assigned to target resources with high revenue per unit exposure, thereby improving the accuracy of recommendation allocation and thus improving the promotion effect.

[0019] Optionally, the step of reallocating the recommendation position for each of the target resources based on the first sequence includes: determining the expected recommendation position corresponding to each of the target resources based on the first sequence;

[0020] In accordance with the order indicated by the first sequence, determine in turn whether each of the target resources meets the expected recommendation conditions of the corresponding expected recommendation position;

[0021] If the target resource does not meet the expected recommendation conditions, then, in the order indicated by the first sequence, determine whether the resources following the target resource meet the expected recommendation conditions.

[0022] If a resource that meets the expected recommendation conditions is identified, the resource that meets the expected recommendation conditions is taken out from the first sequence and assigned to the expected recommendation position. The expected recommendation positions corresponding to the target resource and the resources in the first sequence that are ordered after the expected recommendation position are shifted one position forward. Then, it is determined again whether the target resource meets the expected recommendation conditions of the corresponding expected recommendation position.

[0023] If the target resource meets the expected recommendation conditions, the target resource is removed from the first sequence and assigned to the expected recommendation position.

[0024] By adopting the above technical solution, when the target resource does not meet the expected recommendation conditions corresponding to the expected recommendation position, the resources that meet the expected recommendation conditions and are ranked after the target resource are allocated to the expected recommendation position. Therefore, resources that meet the expected recommendation conditions can be allocated to the recommendation positions ranked first, which can improve the matching degree between the recommendation position allocation result and the expectation, thereby improving the promotion effect.

[0025] Optionally, before recalculating the first recommendation duration and repeating the step of determining the target resource when the first recommendation duration is determined to have reached a first duration threshold, the method further includes:

[0026] Resources that have not been fully recommended will be retained in their original recommendation slots.

[0027] By adopting the above technical solution, the problem of inaccurate allocation of recommendation slots due to inaccurate calculation of revenue per unit exposure for resources that have not been fully recommended can be solved. Therefore, the accuracy of recommendation slot allocation can be improved, thereby improving the promotion effect.

[0028] Optionally, the method further includes:

[0029] In response to a manual intervention command, the target recommendation position indicated by the manual intervention command and the resource allocation recommendation positions in the recommendation positions after the target recommendation position are sequentially shifted by one position based on the order corresponding to each recommendation position;

[0030] The resources indicated by the manual intervention instruction are allocated to the target recommendation slot.

[0031] By adopting the above technical solution, the recommendation position for resource allocation can be adjusted based on manual intervention instructions, thus better adapting to the needs of different recommendation scenarios and broadening the applicable scenarios of resource recommendation methods.

[0032] Optionally, after assigning the resource to be recommended to the corresponding recommendation slot, the method further includes:

[0033] Start calculating the second recommended duration;

[0034] If it is determined that the second recommendation duration has reached the second duration threshold, the unit exposure revenue value corresponding to each resource that has not been recommended is determined;

[0035] Based on the unit exposure revenue value corresponding to each resource, recommendation positions are redistributed for each resource, the first recommendation duration and the second recommendation duration are recalculated, and the steps of determining the target resource and determining the second recommendation duration are repeated.

[0036] By adopting the above technical solutions, the problem of mismatch between the revenue per unit of exposure and the recommendation slot due to the lack of updates to resources in low-exposure recommendation slots can be solved. Therefore, the accuracy of recommendation slot allocation can be improved, thereby enhancing the promotion effect.

[0037] Optionally, the step of determining that the second recommended duration has reached the second duration threshold further includes:

[0038] Select the second number of new resources to be recommended;

[0039] The new resources to be recommended are assigned to the corresponding reserved recommendation slots for recommendation, and / or, the new resources to be recommended are assigned to the corresponding recommendation slots for recommendation through manual intervention.

[0040] By adopting the above technical solution, the quality of recommendation slots allocated to new resources to be recommended can be improved. This helps to solve the problem that the calculated revenue per unit exposure is inaccurate when new resources to be recommended are allocated to poor-quality recommendation slots, which makes it impossible to allocate suitable recommendation slots to new resources to be recommended. Therefore, the accuracy of recommendation slot allocation can be improved.

[0041] Optionally, the method further includes:

[0042] Select the third number of test resources;

[0043] The test resources are assigned to the corresponding recommendation slots for pre-recommendation, and the pre-recommendation duration is calculated.

[0044] If it is determined that the pre-recommendation duration has reached the preset recommendation duration, the first duration threshold is determined based on the preset recommendation duration and the actual exposure count of each recommendation position; or, if it is determined that the pre-recommendation is completed, the pre-recommendation duration is determined as the first duration threshold.

[0045] By adopting the above technical solution, since the first duration threshold is determined based on the preset recommended duration and the actual number of exposures for each recommended position, the determined first duration threshold can reflect the time required for each recommended position to complete a unit number of exposures as a whole, thereby improving the accuracy of the determined first duration threshold.

[0046] Secondly, this application provides an electronic device that adopts the following technical solution:

[0047] An electronic device comprising:

[0048] At least one processor;

[0049] Memory;

[0050] At least one application, wherein the at least one application is stored in memory and configured to be executed by at least one processor, the at least one application being configured to: execute any of the resource recommendation methods provided in the first aspect.

[0051] Thirdly, this application provides a computer-readable storage medium, which adopts the following technical solution:

[0052] A computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform any of the resource recommendation methods provided in the first aspect.

[0053] In summary, this application includes at least one of the following beneficial technical effects:

[0054] 1. It can solve the problem of poor promotion effect when information is delivered based on user type. Since the recommendation slots are allocated to target resources based on revenue per unit of exposure, the impact of the accuracy of user analysis data on the allocation of recommendation slots can be avoided. Therefore, the accuracy of recommendation slot allocation can be improved and the promotion effect can be improved.

[0055] 2. Since the ranking of target resources and the ranking of their corresponding recommendation slots allocate recommendation slots to each target resource, high-quality recommendation slots can be assigned to target resources with high revenue per unit of exposure. Therefore, the accuracy of recommendation allocation can be improved, thereby enhancing the promotion effect. Attached Figure Description

[0056] Figure 1 This is a flowchart illustrating a resource recommendation method provided in an embodiment of this application;

[0057] Figure 2 This is a flowchart illustrating another resource recommendation method provided in an embodiment of this application;

[0058] Figure 3 This application provides a schematic diagram of the structure of an electronic device. Detailed Implementation

[0059] To make the purpose, technical solution, and advantages of this application clearer, the following description is provided in conjunction with the appendix. Figure 1-3 The present application will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the application.

[0060] This application discloses a resource recommendation method executed by an electronic device. The electronic device can be a server, personal computer, tablet computer, or other device with computing capabilities. This embodiment does not limit the type of electronic device. Figure 1 As shown, the method includes at least the following steps:

[0061] Step 101: Select the first number of resources to be recommended.

[0062] The initial quantity is pre-stored in the electronic device and can be set based on actual needs.

[0063] Optionally, the resource to be recommended can be a digital resource that can be used for recommendation. Specifically, the resource to be recommended can be books, music, comics, games, etc. This embodiment does not limit the type of resource to be recommended.

[0064] In one example, selecting a first number of resources to be recommended includes: selecting a first number of resources from a pool of candidate resources to obtain the resources to be recommended. The resources in the pool of candidate resources may include only one type, or they may include multiple different types; this embodiment does not limit this.

[0065] Optionally, the number of candidate resource pools can be one, or more than one. When the number of candidate resource pools is more than one, different candidate resource pools can be divided according to the type of resource, or different candidate resource pools can be divided according to the level of resource. This embodiment does not limit this.

[0066] Optionally, the method for selecting resources to be recommended can be random selection, or selection based on the type and / or level of the resources in proportion. This embodiment does not limit the method for selecting resources to be recommended.

[0067] Step 102: Assign the resources to be recommended to the corresponding recommendation slots for recommendation, and start calculating the first recommendation duration.

[0068] Optionally, the method of assigning the resources to be recommended to the corresponding recommendation positions can be random allocation, or allocation based on the type and / or level of the resources. This embodiment does not limit the method of assigning the resources to be recommended to the corresponding recommendation positions.

[0069] Optionally, different recommendation slots may have different qualities. Specifically, the quality of a recommendation slot is determined based on factors that affect its recommendation effect, such as the exposure of the recommendation slot and the size of the recommendation area. This embodiment does not limit the method for determining the quality of recommendation slots.

[0070] In this embodiment, the number of recommendation slots is greater than or equal to the number of resources to be recommended. Optionally, if the number of recommendation slots is greater than the number of resources to be recommended, priority is given to allocating high-quality recommendation slots to resources.

[0071] In one example, the process of recording the first recommended duration includes: recording the time when the calculation of the first recommended duration begins; and determining the time difference between the current time and the time when the calculation of the first recommended duration begins as the first recommended duration.

[0072] In another example, starting to record the first recommended duration includes: starting a timer and determining the duration indicated by the current timer as the first recommended duration.

[0073] In actual implementation, the first recommended duration can also be calculated in other ways. This embodiment does not limit the calculation method of the first recommended duration.

[0074] Step 103: Determine the recommended resources that have been completed as target resources.

[0075] In this context, "recommendation completion" refers to the resource's exposure reaching an exposure threshold within the continuous recommendation period of the current recommendation slot. This exposure threshold is pre-stored in the electronic device and can be set and adjusted according to actual needs. In one example, the exposure threshold is 1000 times. In actual implementation, other methods can also be used to determine whether a resource's recommendation is complete, such as determining completion when the target resource's exposure revenue reaches a preset revenue threshold. This embodiment does not limit the method used to determine whether a resource's recommendation is complete.

[0076] In this embodiment, after a resource is allocated to a recommendation position for exposure, the continuous recommendation duration of the resource in that recommendation position is calculated until a recommendation position is reallocated to the resource. At this point, the calculation of the continuous recommendation duration of the resource in the current recommendation position is stopped, and the calculation of the continuous recommendation duration of the resource in the newly allocated recommendation position begins.

[0077] Optionally, the continuous recommendation duration of the target resource in the current recommendation position can be equal to the first duration, or it can be greater than or less than the first duration. This embodiment does not limit this.

[0078] Optionally, step 103 can be executed after the recommended resource has been recommended, or it can be executed after the first recommendation duration has reached the first duration threshold. This embodiment does not limit the timing of determining the recommended resource as the target resource.

[0079] In this embodiment, the first duration threshold is pre-stored in the electronic device and can be adjusted automatically or manually. In one example, the first duration threshold is two hours.

[0080] In one example, determining the recommended resource that has been recommended is the target resource, which includes: detecting whether each recommended resource has been recommended during the recommendation process; and if there is a recommended resource that has been recommended, determining the recommended resource that has been recommended is the target resource.

[0081] In another example, determining the recommended resource that has been recommended is the target resource, which includes: when the first recommendation duration reaches a first duration threshold, determining the recommended resource that has been recommended is the target resource among the various recommended resources.

[0082] Step 104: Determine the revenue per unit of exposure for each target resource.

[0083] The number of times per unit is pre-stored in the electronic device and can be set and adjusted according to actual needs.

[0084] Optionally, the Revenue Per Exposure (RPP) value indicates the revenue generated by the target resource during the exposure process. Specifically, RRP is calculated based on factors such as the target resource's click-through rate, view rate, conversion rate, and / or Average Revenue Per Paying User (ARPPU) during the exposure process.

[0085] In one example, the exposure threshold is equal to the number of exposures per unit.

[0086] In one instance, the number of times is 1000.

[0087] In one example, the revenue per exposure is the revenue value corresponding to the first unit of exposures of the target resource within the continuous recommendation duration of the current recommendation position. Specifically, determining the revenue per exposure value for each target resource includes: determining the revenue value corresponding to the first unit of exposures of each target resource within the continuous recommendation duration of the current recommendation position as the revenue per exposure value. For example, taking 1000 exposures as an example, the revenue value corresponding to the first 1000 exposures of the target resource is determined as the revenue per exposure value. At this time, step 104 can be executed when the target resource is determined, or it can be executed when the first recommendation duration reaches the first duration threshold; this embodiment does not limit this.

[0088] By defining the revenue value corresponding to the previous unit of exposure as the revenue value per unit of exposure, the impact of the number of exposures on the revenue per exposure of a resource can be reduced. In other words, by evaluating the revenue per unit of exposure of a resource under the same number of exposures, it can help solve the problem that the revenue per unit of exposure of a resource changes with the number of exposures due to the different exposures corresponding to different recommendation positions, and improve the accuracy of the determined revenue per unit of exposure.

[0089] In another example, the revenue per exposure is the average revenue per exposure for a target resource within the continuous recommendation period of the current recommendation position. Specifically, determining the revenue per exposure value for each target resource includes: multiplying the ratio of the total revenue of each target resource within the continuous recommendation period of the current recommendation position to the total number of exposures of the target resource within the continuous recommendation period of the current recommendation position by the unit number of exposures to determine the revenue per exposure value for the target resource. In this case, step 104 is executed when the first recommendation period reaches a first duration threshold.

[0090] Since the revenue per exposure is determined by the ratio of the total revenue of a resource to the total number of exposures during the continuous recommendation period of the current recommendation position, and the revenue per exposure changes with the total number of exposures, and different recommendation positions correspond to different amounts of exposure, the calculated revenue per exposure can reflect the degree of matching between the resource and the recommendation position, thereby improving the accuracy of the recommendation positions allocated to the resource.

[0091] Step 105: Sort each target resource based on the unit exposure revenue value corresponding to the target resource to obtain the first sequence.

[0092] Optionally, the first sequence can be implemented using data structures such as arrays, lists, and linked lists. The elements in the first sequence can be any information that can identify the target resource, such as the target resource's ID. This embodiment does not limit the implementation method and content of the first sequence.

[0093] In this embodiment, the example is to sort the target resources in descending order of revenue per unit exposure. In actual implementation, the target resources can also be sorted in ascending order of revenue per unit exposure. This embodiment does not limit this.

[0094] Optionally, if different target resources have the same revenue per unit of exposure, the target resources with the same revenue per unit of exposure are ranked based on Customer Lifetime Value (LTV). Specifically, the Lifetime Value of a target resource is calculated as follows: Lifetime Value = Average Purchase Value × Average Purchase Rate × Average Customer Lifetime.

[0095] In actual implementation, the ranking of target resources with the same unit exposure value can also be determined based on other methods, such as: determining the ranking of target resources based on the listing time of the target resources. This embodiment does not limit the method of determining the ranking of resources with the same unit exposure value.

[0096] In one example, each target resource is sorted based on its revenue per unit exposure value to obtain a first sequence, which includes: when the target resources are determined, the determined target resources are inserted into the first sequence in descending order of their revenue per unit exposure value to obtain the first sequence.

[0097] In another example, each target resource is sorted based on its revenue per unit exposure value to obtain a first sequence, including: when the first recommendation duration reaches a first duration threshold, each target resource is sorted based on its revenue per unit exposure value to obtain a first sequence.

[0098] Step 106: If it is determined that the first recommendation duration has reached the first duration threshold, the recommendation positions are reallocated to each target resource based on the first sequence, the first recommendation duration is recalculated, and the step of determining the target resource is repeated, i.e., step 103 is executed.

[0099] In this way, recommendation slots can be reallocated to resources that have completed their recommendations, based on a first duration threshold. Therefore, the recommendation slots corresponding to resources that have completed their recommendations can be updated periodically based on the revenue per unit exposure, thereby improving the accuracy of recommendation slot allocation.

[0100] In this embodiment, when the recommendation slot is reassigned to the target resource, the continuous exposure of the target resource in the recommendation slot is also recalculated because the recommendation slot corresponding to the target resource has changed.

[0101] Optionally, recommendation slots may be reallocated to each target resource based on the first sequence, including: allocating high-quality recommendation slots to resources with high revenue per unit exposure based on the first sequence.

[0102] Optionally, the recommendation positions for each target resource are reallocated based on the first sequence, including: determining the order of the recommendation positions for each target resource to obtain a second sequence; and reallocating recommendation positions for each target resource based on the first and second sequences.

[0103] In this embodiment, the example is to sort the recommended positions of the target resources in descending order of quality. In actual implementation, the target resources can also be sorted in ascending order of income value. This embodiment does not limit this.

[0104] Since recommendation slots are assigned to each target resource based on the ranking of the target resources and the ranking of the recommendation slots corresponding to the target resources, high-quality recommendation slots can be assigned to target resources with high revenue per unit of exposure. Therefore, the accuracy of recommendation allocation can be improved, thereby improving the promotion effect.

[0105] In one example, the order of the recommendation positions of each target resource is determined to obtain a second sequence, which includes: determining the order of the recommendation positions of each target resource based on the sorting of each recommendation position to obtain a second sequence.

[0106] The ranking of each recommendation position is determined based on the quality of the recommendation position.

[0107] Optionally, the quality order of each recommendation position can be pre-stored in an electronic device and can be adjusted automatically or manually.

[0108] Optionally, the quality order of each recommendation position is determined based on the type of the recommended object. In one example, the electronic device stores the quality ranking of each recommendation position for different types of recommended objects. When using the resource recommendation method to make recommendations for a certain type of recommended object, the electronic device can obtain the quality order corresponding to that type of recommended object. In this way, the accuracy of the quality order of each recommendation position can be determined.

[0109] In another example, the order of the recommendation positions of each target resource is determined to obtain a second sequence, which includes: sorting the recommendation positions corresponding to each target resource based on the time taken to complete the recommendation of the target resource to obtain the second sequence.

[0110] Optionally, the recommendation positions for each target resource are reallocated based on the first sequence and the second sequence, including: assigning the nth resource in the first sequence to the nth recommendation position in the second sequence, where n is less than or equal to the number of target resources.

[0111] In one example, recalculating the first recommended duration includes: re-recording the time when the calculation of the first recommended duration begins; and determining the time difference between the current time and the time when the calculation of the first recommended duration begins as the first recommended duration.

[0112] In another example, recalculating the first recommended duration includes: resetting the timer to zero and restarting the timer, and determining the duration indicated by the current timer as the first recommended duration.

[0113] Optionally, before recalculating the first recommendation duration and repeating the step of determining the target resource after determining that the first recommendation duration has reached the first duration threshold, the method further includes: retaining the unrecommended resources in the original recommendation position.

[0114] In this embodiment, since the resource is retained in the original recommendation position, the continuous exposure of the resource in the original recommendation position is calculated cumulatively because the recommendation position corresponding to the resource has not changed.

[0115] If the first recommendation duration reaches the first duration threshold, resources that have not been recommended will not participate in the reallocation of recommendation slots in this round. This can help solve the problem of inaccurate calculation of the revenue per unit exposure of resources that have not been recommended, which leads to inaccurate allocation of recommendation slots. Therefore, it can improve the accuracy of recommendation slot allocation and thus improve the promotion effect.

[0116] The implementation principle of the resource recommendation method disclosed in this application is as follows: A first number of resources to be recommended are selected; the resources to be recommended are assigned to corresponding recommendation positions for recommendation, and a first recommendation duration is calculated; the recommended resources are determined as target resources; the unit exposure revenue value corresponding to each target resource is determined; each target resource is sorted based on its corresponding revenue value to obtain a first sequence; if the first recommendation duration reaches a first duration threshold, recommendation positions are reallocated to each target resource based on the first sequence, the first recommendation duration is recalculated, and the steps for determining target resources are repeated. This method can solve the problem of poor promotion effect when information is delivered based on user type. Since recommendation positions are allocated to target resources based on unit exposure revenue, the influence of the accuracy of user analysis data on recommendation position allocation can be avoided. Therefore, the accuracy of recommendation position allocation can be improved, and the promotion effect can be improved.

[0117] Based on the above technical solution, optionally, the resource recommendation method provided in this embodiment further includes: selecting a third number of test resources; allocating the test resources to the corresponding recommendation positions for pre-recommendation, and starting to calculate the pre-recommendation duration; if it is determined that the pre-recommendation duration has reached the preset recommendation duration, determining a first duration threshold based on the preset recommendation duration and the actual exposure count of each recommendation position, or, if it is determined that the pre-recommendation is completed, determining the pre-recommendation duration as the first duration threshold.

[0118] Optionally, the step of determining the first duration threshold based on test resources can be performed before step 101, or it can be performed after determining that the second recommended duration has reached the second duration threshold, or it can be performed periodically, such as once a month. This embodiment does not limit the execution time of the step of determining the first duration threshold based on test resources.

[0119] Optionally, selecting a third number of test resources includes: selecting a first number of resources from the test resource pool to obtain test resources. The resources in the test resource pool may include only one type, or they may include multiple different types; this embodiment does not limit this.

[0120] Optionally, the test resource pool can be the same as or different from the resource pool to be recommended. This embodiment does not limit this.

[0121] Optionally, a first duration threshold is determined based on a preset recommended duration and the actual number of exposures for each recommended position, including: multiplying the ratio of the preset recommended duration to the actual number of exposures for each recommended position by the number of exposures per unit to determine the expected first duration for each recommended position; and determining the first duration threshold based on the expected first duration for each recommended position.

[0122] In one example, determining the first duration threshold based on the expected first duration corresponding to each recommendation position includes: sorting each recommendation position based on the expected first duration, and determining the expected first duration corresponding to the recommendation position of the specified position as the first duration threshold.

[0123] For example: if the third quantity is 100, the expected first duration corresponding to the 80th recommended position will be determined as the first duration threshold.

[0124] In another example, the first duration threshold is determined based on the expected first duration corresponding to each recommendation position, including: sorting each recommendation position based on the expected first duration, and determining the average of the expected first durations corresponding to the recommendation positions of the specified multiple positions as the first duration threshold.

[0125] For example, if the first quantity is 100, the average of the expected first duration corresponding to the 90th recommended position of the 70th value is determined as the first duration threshold.

[0126] Since the first duration threshold is determined based on the preset recommended duration and the actual number of exposures for each recommended position, the determined first duration threshold can reflect the duration required for each recommended position to complete a unit number of exposures, thereby improving the accuracy of the determined first duration threshold.

[0127] In this embodiment, the method for determining the completion of pre-recommendation can be that the number of recommendation positions with actual exposure counts reaching the unit exposure count reaches a preset completion number, or the ratio of the number of recommendation positions with actual exposure counts reaching the unit exposure count to the third number reaches a preset completion rate. This embodiment does not limit the method for determining the completion of recommendation.

[0128] Since the first duration threshold is determined based on the actual time required for the recommended position to complete one unit of exposure, the determined first duration threshold can more accurately reflect the time required for the recommended position to complete one unit of exposure. Therefore, the determined first duration threshold can better reflect the actual situation of the recommended position.

[0129] Based on the above technical solution, the resource recommendation method provided in this embodiment further includes: when it is determined that the pre-recommendation duration has reached the preset recommendation duration, determining the order of each recommendation position based on the actual exposure count of each recommendation position.

[0130] In this embodiment, the example is to sort the recommended positions in descending order of actual exposure count. In actual implementation, the recommended positions can also be sorted in ascending order of actual exposure count. This embodiment does not limit this.

[0131] Optionally, for recommended positions with the same actual exposure count, they can be sorted according to other factors affecting the quality of the recommended positions, such as the size of the recommended area and the position of the recommended position. For example, recommended positions with the same actual exposure count can be sorted from largest to smallest recommended area, or from top to bottom based on their position. This embodiment does not limit the sorting method for recommended positions with the same actual exposure count.

[0132] Since the ranking of each recommendation slot is determined based on the actual exposure of each slot during the recommendation process, and the quality of the recommendation slots is relatively stable, the ranking of recommendation slots determined based on pre-recommendation can better reflect the quality between each recommendation slot. Therefore, the recommendation slots can be allocated to resources based on the ranking of recommendation slots, thus improving the accuracy of recommendation slot allocation.

[0133] Based on the above technical solutions, alternatives include, for example... Figure 2 As shown, in step 102, after assigning the resources to be recommended to the corresponding recommendation slots, the following steps are also included:

[0134] Step 201: Begin calculating the second recommended duration.

[0135] In one example, step 201 and the start of calculating the first recommended duration in step 102 are performed simultaneously.

[0136] The calculation method for the second recommended duration is the same as that for the first recommended duration, and this embodiment does not limit it.

[0137] Step 202: If the second recommendation duration is determined to have reached the second duration threshold, determine the unit exposure revenue value corresponding to each resource that has not been recommended.

[0138] "Not recommended" means that the exposure of a resource within the continuous recommendation period of the current recommendation position has not reached the exposure threshold.

[0139] In this embodiment, the second duration threshold is pre-stored in the electronic device and can be adjusted automatically or manually.

[0140] Optionally, the second duration threshold is greater than the first duration threshold.

[0141] In one example, the second duration threshold is an integer multiple of the first duration threshold. In one instance, the second duration threshold is 72 hours.

[0142] Optionally, the unit exposure revenue value corresponding to each resource that has not been fully recommended is determined, including: the product of the ratio of the total revenue value of the resource that has not been fully recommended within the continuous recommendation period of the current recommendation position to the total number of exposures of the resource within the continuous recommendation period of the current recommendation position and the unit number of exposures is determined as the unit exposure value corresponding to the resource.

[0143] Optionally, if it is determined that the second recommended duration has reached the second duration threshold and the first recommended duration has reached the first duration threshold at the same time, step 203 is executed, but step 106 is not executed.

[0144] Step 203: Based on the unit exposure revenue value corresponding to each resource, the recommendation positions are redistributed for each resource, the first recommendation duration and the second recommendation duration are recalculated, and the steps of determining the target resource and determining the second recommendation duration are repeated, namely steps 202 and 103.

[0145] In this way, recommendation slots can be redistributed to all resources to be recommended based on the second duration threshold. Therefore, the recommendation slots corresponding to all resources to be recommended can be updated periodically based on the revenue per unit exposure, thereby improving the accuracy of recommendation slot allocation.

[0146] In this embodiment, as Figure 2 As shown, steps 202 and 203 are executed simultaneously with steps 103, 104, 105 and 106.

[0147] Optionally, the recommendation slots for each resource can be reallocated based on the unit exposure value corresponding to each resource, including: sorting each resource according to the unit exposure value corresponding to each resource to obtain a third sequence; and reallocating recommendation slots for each resource based on the third sequence.

[0148] The method of sorting each resource based on the unit exposure value corresponding to each resource to obtain the third sequence is the same as the method of sorting each target resource based on the unit exposure revenue value corresponding to the target resource to obtain the first sequence in step 105 above. The method of reallocating recommendation positions to each resource based on the third sequence is the same as the method of reallocating recommendation positions to each target resource based on the second sequence in step 106 above. This embodiment will not repeat the details.

[0149] In the above technical solution, when the second recommendation duration reaches the second duration threshold, resources that have not completed the recommendation also participate in the reallocation of recommendation positions. This can help solve the problem that resources in recommendation positions with low exposure cannot be reallocated because they have never met the recommendation completion conditions. Therefore, it can improve the accuracy of recommendation position allocation and thus improve the promotion effect.

[0150] Optionally, if it is determined that the second recommendation duration has reached the second duration threshold, the method further includes: selecting a second number of new resources to be recommended; allocating the new resources to be recommended to the corresponding reserved recommendation slots for recommendation; and / or, allocating the new resources to be recommended to the corresponding recommendation slots for recommendation through manual intervention.

[0151] In one example, selecting a second number of new resources to be recommended includes: selecting a second number of resources from the pool of resources to be selected, thus obtaining new resources to be recommended. The implementation of the pool of resources to be selected is the same as that of the pool of resources to be recommended in step 101, and will not be described again in this embodiment.

[0152] In one example, the number of recommendation slots is unlimited. For instance, if the promotion page can be continuously scrolled down and / or flipped through, then each time a new batch of resources to be recommended is added, the corresponding number of recommendation slots are activated to allocate resources.

[0153] Optionally, newly enabled recommendation slots may be of lower quality than existing recommendation slots.

[0154] In actual implementation, the number of recommendation slots can also be limited. In this case, if the number of unallocated recommendation slots is less than the number of new resources to be recommended before adding new resources to be recommended, then remove the resources ranked last among all resources or the resources ranked last among all recommendation slots from the recommendation slots and end the recommendation of these resources. Then, allocate recommendation slots for the new resources to be recommended.

[0155] For example: If there are 500 recommendation slots in total, and all of them have been allocated resources, and there are 50 new resources to be recommended, then the last 50 resources in the list of all resources or the last 50 recommendation slots in the list of all recommendation slots are removed from the recommendation slots and the recommendation of these 50 resources ends. Then, recommendation slots are allocated to the new resources to be recommended.

[0156] In one example, reserved recommendation slots are set by defining the expected recommendation criteria corresponding to the recommendation slot. For instance, by setting the expected recommendation criteria for a recommendation slot to the recommendation slot for a new resource, that recommendation slot will prioritize recommending the new resource.

[0157] "New resources" refers to resources whose cumulative recommendation time is less than the second duration threshold. Cumulative recommendation time is the sum of the recommendation times of the resource across all recommendation slots. For example, if a resource is recommended for 2 hours in recommendation slot A and then reassigned to recommendation slot B, the resource's cumulative recommendation time is the sum of 2 hours and the time it was recommended in recommendation slot B.

[0158] In this case, only new resources selected when the second recommendation duration reaches the second duration threshold can be recommended in the new resource recommendation slot.

[0159] In practice, the reserved recommendation slots can include other types of recommendation slots. Different types of reserved recommendation slots have different expected recommendation conditions, which can be set according to actual needs.

[0160] Optionally, the expected recommendation criteria can be set based on factors such as the type, level, whether it is a new resource, and the recommendation round of the target resource. Alternatively, they can be set to be unrestricted, meaning that any resource can be recommended. This embodiment does not limit the type of expected recommendation criteria.

[0161] Optionally, the new resources to be recommended are allocated to the corresponding reserved recommendation slots for recommendation, including: allocating the new resources to the corresponding reserved recommendation slots for recommendation based on the expected recommendation conditions of the reserved recommendation slots.

[0162] In one instance, new resources to be recommended are assigned to reserved new resource recommendation slots for recommendation.

[0163] Optionally, new resources to be recommended can be assigned to corresponding recommendation slots through manual intervention, including: assigning new resources to manually designated recommendation slots for recommendation.

[0164] Optionally, the reserved recommendation slots and manually designated recommendation slots may include the recommendation slots corresponding to the original resources to be recommended, or they may include newly enabled recommendation slots after acquiring the second number of second resources. This embodiment does not limit this.

[0165] Optionally, if the second number is greater than the number of reserved recommendation slots and / or manually intervened recommendation slots, after the reserved recommendation slots are allocated and / or the manual intervention is completed, the remaining new resources to be recommended are allocated to the newly enabled recommendation slots.

[0166] In one example, the second quantity is 200, and the original recommendation slots are 100, including 10 new book recommendation slots. In this case, the 10 new resources to be recommended will be allocated to the 10 new book recommendation slots. Correspondingly, the resources in the 10 original recommendation slots will be allocated to the newly enabled recommendation slots ranked 101 to 110. The remaining 190 new resources to be recommended will be allocated to the newly enabled recommendation slots ranked 111 and thereafter.

[0167] In the above technical solution, since allocating new resources to be recommended to reserved recommendation slots or assigning recommendation slots to new resources to be recommended through manual intervention can improve the quality of recommendations assigned to new resources to be recommended, it helps to solve the problem that the calculated revenue per unit exposure is inaccurate due to new resources to be recommended being assigned to poor-quality recommendation slots, thus making it impossible to assign suitable recommendation slots to new resources to be recommended. Therefore, it can improve the accuracy of recommendation slot allocation.

[0168] To adapt to the needs of different recommendation scenarios, further, in step 106, recommendation positions are reallocated for each target resource based on the first sequence, including: determining the expected recommendation position corresponding to each target resource based on the first and second sequences; determining whether each target resource meets the expected recommendation conditions for the corresponding expected recommendation position in the order indicated by the first sequence; if a target resource does not meet the expected recommendation conditions, determining whether the resources following the target resource meet the expected recommendation conditions in the order indicated by the first sequence; if a resource meets the expected recommendation conditions, removing the resource from the first sequence and assigning it to the expected recommendation position, shifting the expected recommendation positions of the target resource and the resources in the first sequence that are ordered after the expected recommendation position by one position, and determining again whether the target resource meets the expected recommendation conditions for the corresponding expected recommendation position; if the target resource meets the expected recommendation conditions, removing the target resource from the first sequence and assigning it to the expected recommendation position.

[0169] Optionally, determining the expected recommendation position for each target resource based on the first sequence includes: determining the expected recommendation position for each target resource based on the first sequence and the second sequence.

[0170] The method for determining the expected recommendation position corresponding to each target resource based on the first sequence and the second sequence is the same as the method for determining the recommendation position corresponding to each target resource based on the first sequence and the second sequence described above, and will not be repeated here.

[0171] In the above technical solution, by setting the expected recommendation conditions corresponding to the target resources to restrict the resources in the recommendation positions, it can help solve the problem that the recommendation positions allocated based on the revenue per unit exposure do not meet the recommendation requirements. Therefore, the allocation result of the recommendation positions can meet the recommendation requirements.

[0172] In addition, since resources that meet the expected recommendation conditions but are ranked after the target resource are allocated to the expected recommendation position when the target resource does not meet the expected recommendation conditions, resources that meet the expected recommendation conditions can be allocated to the recommendation positions ranked first. This can improve the matching degree between the recommendation position allocation results and the expectations, thereby improving the promotion effect.

[0173] Optionally, reallocating recommendation positions for each target resource based on the first sequence and the second sequence further includes: if no resource that meets the desired recommendation conditions is determined, removing the target resource from the first sequence and assigning it to the desired recommendation position.

[0174] Since resources ranked after the target resource do not meet the desired recommendation criteria, resources matching the desired recommendation position will be allocated to the recommendation position. This avoids the problem of wasted exposure in the desired recommendation position due to unallocated resources. Therefore, the exposure of the desired recommendation position can be fully utilized, thus improving the promotion efficiency.

[0175] Based on the above technical solution, the resource recommendation method provided in this embodiment further includes: in response to a manual intervention command, based on the sorting corresponding to each recommendation position, sequentially shifting the target recommendation position indicated by the manual intervention command and the resource allocation recommendation positions in the recommendation positions sorted after the target recommendation position by one position; and allocating the resource indicated by the manual intervention command to the target recommendation position.

[0176] Optionally, the execution time of the manual intervention instruction can be before or after the allocation of the recommended bit for the resource; this embodiment does not limit this.

[0177] Optionally, after a manual intervention instruction allocates a recommendation slot to a resource, since the first duration has already begun to be calculated before the manual intervention, the first duration will not be recalculated after the recommendation slot corresponding to the resource is adjusted based on the manual intervention instruction, but will continue to be calculated. In this case, the number of exposures and revenue values ​​of the resource in different recommendation slots within the first duration can be calculated cumulatively.

[0178] In one example, the manual intervention command has the highest priority. When adjusting the recommended bits for bit resource allocation based on the manual intervention command, it is not necessary to consider factors such as the expected recommendation conditions corresponding to the recommended bits.

[0179] In other examples, when adjusting the recommended bits for bit resource allocation based on manual intervention instructions, it is also necessary to consider factors such as the expected recommendation conditions corresponding to the recommended bits.

[0180] Optionally, based on the sorting corresponding to each recommendation position, the recommendation position for resource allocation in the recommendation positions after the target recommendation position indicated by the manual intervention instruction and the recommendation positions after the target recommendation position are sequentially shifted one position before the target recommendation position, and also includes: determining whether there are resources in the target recommendation position indicated by the manual intervention instruction;

[0181] If there are resources in the target recommendation position indicated by the manual intervention command, the recommended position for allocating resources in the target recommendation position indicated by the manual intervention command and the recommended positions after the target recommendation position will be sequentially shifted by one position based on the sorting corresponding to each recommendation position, and the resources indicated by the manual intervention command will be allocated to the target recommendation position.

[0182] If there are no resources in the target recommendation slot indicated by the manual intervention instruction, directly execute the step of allocating the resources indicated by the manual intervention instruction to the target recommendation slot.

[0183] In the above technical solution, since the recommended bits of resource allocation can be adjusted based on manual intervention instructions, it can better adapt to the needs of different recommendation scenarios and broaden the applicable scenarios of resource recommendation methods.

[0184] This application also provides an electronic device, such as... Figure 3 As shown, Figure 3 The illustrated electronic device 300 includes a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, for example, via a bus 302. Optionally, the electronic device 300 may also include a transceiver 304. It should be noted that in practical applications, the transceiver 304 is not limited to one type, and the structure of this electronic device 300 does not constitute a limitation on the embodiments of this application.

[0185] Processor 301 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 301 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0186] Bus 302 may include a pathway for transmitting information between the aforementioned components. Bus 302 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 302 can be divided into address bus, data bus, control bus, etc. For ease of representation... Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0187] The memory 303 may be a ROM (Read Only Memory) or other type of static storage device capable of storing static information and instructions, RAM (Random Access Memory) or other type of dynamic storage device capable of storing information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0188] The memory 303 is used to store application code that executes the solution of this application, and its execution is controlled by the processor 301. The processor 301 is used to execute the application code stored in the memory 303 to implement the content shown in the foregoing method embodiments.

[0189] Electronic devices include, but are not limited to: mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), and in-vehicle terminals (such as in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Servers can also be included. Figure 3 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0190] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed in a computer, causes the computer to perform the resource recommendation method provided in the above embodiments.

[0191] It should be understood that although the steps in the flowcharts in the accompanying drawings are shown sequentially as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise expressly stated herein, there is no strict order in which these steps are performed, and they may be performed in other orders.

[0192] The above are only some embodiments of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A resource recommendation method, characterized by, The method includes: Select the first number of resources to be recommended; The resources to be recommended are assigned to the corresponding recommendation slots for recommendation, and the first recommendation duration is calculated. The resource to be recommended that has been recommended is identified as the target resource. Determine the revenue per unit exposure for each of the target resources, where the unit exposure is 1000 times; Based on the unit exposure revenue value corresponding to the target resource, each target resource is sorted to obtain a first sequence; If it is determined that the first recommendation duration has reached the first duration threshold, recommendation positions are reallocated for each of the target resources based on the first sequence, the first recommendation duration is recalculated, and the step of determining the target resources is repeated. The step of reallocating recommendation positions for each of the target resources based on the first sequence includes: determining the order of the recommendation positions of each target resource to obtain a second sequence; and reallocating the recommendation positions for each target resource based on the first sequence and the second sequence. The step of reallocating the recommendation positions for each of the target resources based on the first sequence includes: determining the expected recommendation position corresponding to each of the target resources based on the first sequence; sequentially determining whether each of the target resources meets the expected recommendation conditions for the corresponding expected recommendation position according to the order indicated by the first sequence; if a target resource does not meet the expected recommendation conditions, sequentially determining whether the resources following the target resource meet the expected recommendation conditions according to the order indicated by the first sequence; if a resource that meets the expected recommendation conditions is determined, removing the resource that meets the expected recommendation conditions from the first sequence and allocating it to the expected recommendation position, shifting the expected recommendation positions corresponding to the target resource and the resources in the first sequence that are ordered after the expected recommendation position by one position, and determining again whether the target resource meets the expected recommendation conditions for the corresponding expected recommendation position; if the target resource meets the expected recommendation conditions, removing the target resource from the first sequence and allocating it to the expected recommendation position. After assigning the resources to be recommended to the corresponding recommendation slots, the method further includes: starting to calculate the second recommendation duration; if it is determined that the second recommendation duration has reached the second duration threshold, determining the unit exposure revenue value corresponding to each resource that has not been recommended; based on the unit exposure revenue value corresponding to each resource, reassigning recommendation slots to each resource for recommendation, recalculating the first recommendation duration and the second recommendation duration, and repeating the steps of determining the target resource and determining the second recommendation duration. If the second recommendation duration is determined to have reached the second duration threshold, the method further includes: selecting a second number of new resources to be recommended, and assigning the new resources to the corresponding recommendation slots for recommendation through manual intervention.

2. The method of claim 1, wherein, Before recalculating the first recommendation duration and repeating the step of determining the target resource when the first recommendation duration is determined to have reached a first duration threshold, the method further includes: Resources that have not been fully recommended will be retained in their original recommendation slots.

3. The method of claim 1, wherein, The method further includes: In response to a manual intervention command, the target recommendation position indicated by the manual intervention command and the resource allocation recommendation positions in the recommendation positions after the target recommendation position are sequentially shifted by one position based on the order corresponding to each recommendation position. The resources indicated by the manual intervention instruction are allocated to the target recommendation slot.

4. The method of claim 1, wherein, The method further includes: Select the third number of test resources; The test resources are assigned to the corresponding recommendation slots for pre-recommendation, and the pre-recommendation duration is calculated. If it is determined that the pre-recommendation duration has reached the preset recommendation duration, the first duration threshold is determined based on the preset recommendation duration and the actual exposure count of each recommendation position; or, if it is determined that the pre-recommendation is completed, the pre-recommendation duration is determined as the first duration threshold.

5. An electronic device, comprising: The electronic device includes: At least one processor; Memory; At least one application, wherein the at least one application is stored in memory and configured to be executed by at least one processor, said at least one application being configured to: perform the resource recommendation method according to any one of claims 1 to 4.

6. A computer-readable storage medium having stored thereon a computer program, characterized in that, When the computer program is executed in a computer, the computer is instructed to perform the resource recommendation method according to any one of claims 1 to 4.

Images