Resource processing method and device, electronic equipment and storage medium

Abstract

The invention relates to a resource processing method and device, electronic equipment and a storage medium. The method comprises the steps that prior information is acquired; the prior information represents that a target object is associated with a first article, and the target object executes resource operation on historical delivery resources including a second article; performing resource recall processing on the target object based on a recall model to obtain a first recall resource; performing resource recall processing on the target object based on the prior information to obtain a second recall resource; performing resource release based on the first recalled resource and the second recalled resource to obtain a resource release result index; and determining the credibility of the prior information based on the resource delivery result index. According to the invention, the accuracy of resource delivery can be improved, and the resource delivery effect is improved.

Description

Technical Field

[0001] This application relates to the field of information processing technology, and in particular to a resource processing method, apparatus, electronic device and storage medium. Background Technology

[0002] In existing technologies, to achieve targeted resource delivery, resource package binding capabilities are provided during resource delivery. Specifically, the resource delivery provider can upload audience packages and restrict users within the audience package to be the target audience for resource delivery. Since audience package discovery mainly relies on the subjective experience of the resource delivery provider, if errors occur in audience package discovery, or if the currently delivered resource does not match the bound audience package, it will lead to a mismatch between the delivered resource and the target audience, thereby reducing the accuracy of resource delivery and affecting the effectiveness of resource delivery. Summary of the Invention

[0003] The technical problem to be solved by this application is to provide a resource processing method, apparatus, electronic device and storage medium, which uses prior information to recall and deploy resources, and determines the credibility of prior information through resource deployment result indicators. This enables the evaluation and measurement of prior information, ensures the long-term sustainable use of prior information, facilitates resource deployment based on highly credible prior information, improves the accuracy of resource deployment, and enhances the effectiveness of resource deployment.

[0004] To address the aforementioned technical problems, this application provides a resource processing method, comprising:

[0005] Obtain prior information; the prior information indicates that the target object is associated with the first item, and the target object has performed resource operations on historical resources containing the second item.

[0006] Based on the recall model, resource recall processing is performed on the target object to obtain the first recalled resource;

[0007] Based on the prior information, resource recall processing is performed on the target object to obtain a second recalled resource;

[0008] Based on the first and second recalled resources, resources are deployed to obtain resource deployment result indicators;

[0009] The credibility of the prior information is determined based on the resource allocation result indicators.

[0010] On the other hand, this application provides a resource processing apparatus, including:

[0011] The prior information acquisition module is used to acquire prior information; the prior information indicates that the target object is associated with the first item, and the target object has performed resource operations on historical resources containing the second item;

[0012] The first recall module is used to perform resource recall processing for the target object based on the recall model to obtain the first recalled resource.

[0013] The second recall module is used to perform resource recall processing for the target object based on the prior information to obtain the second recalled resources.

[0014] The resource deployment module is used to deploy resources based on the first recalled resources and the second recalled resources, and obtain resource deployment result indicators.

[0015] The credibility determination module is used to determine the credibility of the prior information based on the resource deployment result indicators.

[0016] On the other hand, this application provides an electronic device including a processor and a memory, wherein the memory stores at least one instruction or at least one program, the at least one instruction or the at least one program being loaded and executed by the processor to implement the resource processing method as described above.

[0017] On the other hand, this application provides a computer storage medium storing at least one instruction or at least one program, wherein the at least one instruction or the at least one program is loaded by a processor and executed as described above in the resource processing method.

[0018] On the other hand, this application provides a computer program product including computer instructions stored in a computer-readable storage medium; a processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions to cause the electronic device to perform the resource processing method described above.

[0019] Implementing the embodiments of this application has the following beneficial effects:

[0020] This application, upon obtaining prior information, can, on the one hand, recall first recall resource information for the target object based on a recall model, and on the other hand, recall second recall resource information for the target object based on prior information. Specifically, the second recall resource can be a delivery resource containing the first item. Therefore, when delivery is based on both the first and second recall resources, the delivery resources delivered to the target object can include those from both the first and second recall resources. Furthermore, based on the target object's resource operations on the delivered resources, corresponding resource delivery result indicators can be obtained. Determining the credibility of prior information through these resource delivery result indicators enables the evaluation and measurement of prior information, ensuring its long-term sustainable use. This facilitates resource delivery based on highly credible prior information, improving the accuracy and effectiveness of resource delivery. Moreover, by jointly delivering the first and second recall resources, the application avoids the problem of resource delivery results being skewed due to delivery solely based on the second recall resource, thus affecting the accuracy of the resource delivery result indicators and improving their overall accuracy. Attached Figure Description

[0021] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the implementation environment provided in the embodiments of this application;

[0023] Figure 2 This is a flowchart of a resource processing method provided in an embodiment of this application;

[0024] Figure 3 This is a flowchart of the prior information generation method provided in the embodiments of this application;

[0025] Figure 4 This is a schematic diagram of the prior information configuration provided in an embodiment of this application;

[0026] Figure 5 This is a schematic diagram of the resource deployment process provided in the embodiments of this application;

[0027] Figure 6 This is a schematic diagram of the exploration process for prior information provided in an embodiment of this application;

[0028] Figure 7 This is a flowchart of the training method for the recall model provided in the embodiments of this application;

[0029] Figure 8 This is a flowchart of the method for determining the matching between deployed resources and the deployed objects provided in the embodiments of this application;

[0030] Figure 9 This is a schematic diagram of a resource processing device provided in an embodiment of this application;

[0031] Figure 10 This is a schematic diagram of an electronic device structure provided in an embodiment of this application. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0033] 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 server 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 devices.

[0034] In this application embodiment, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0035] It is understood that in the specific embodiments of this application, data such as user information are involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0036] Please see Figure 1 The illustration shows an implementation environment provided in the embodiments of this application. The implementation environment may include at least one electronic terminal 110 and a resource delivery server 120, wherein the electronic terminal 110 and the resource delivery server 120 can communicate with each other via a network.

[0037] Specifically, the electronic terminal 110 can send prior information to the resource delivery server 120, and the resource delivery server 120 can perform resource recall and resource delivery based on the prior information. Furthermore, the resource delivery server 120 can also verify the credibility of the prior information. If the prior information passes the credibility verification, it can recall resources for the target object based on the prior information and deliver resources based on the recalled resources.

[0038] The electronic terminal 110 can communicate with the resource delivery server 120 based on a browser / server (B / S) or client / server (C / S) model. The electronic terminal 110 may include physical devices such as smartphones, tablets, laptops, digital assistants, smart wearable devices, in-vehicle terminals, and servers, and may also include software running on the physical device, such as applications. The operating system running on the electronic terminal 110 in this embodiment may include, but is not limited to, Android, iOS, Linux, and Windows.

[0039] The resource distribution server 120 and the electronic terminal 110 can establish a communication connection via wired or wireless means. The resource distribution server 120 may include a stand-alone server, a distributed server, or a server cluster consisting of multiple servers, wherein the server may be a cloud server.

[0040] To address the technical problem in existing technologies where the deployed resources do not match the deployed objects, thereby reducing the accuracy of resource deployment and affecting its effectiveness, this application provides a resource processing method. The executing entity of this method can be the aforementioned resource deployment server. (See also...) Figure 2 The method may include:

[0041] S210. Obtain prior information; the prior information indicates that the target object is associated with the first item, and the target object has performed resource operations on historical resources containing the second item.

[0042] In this embodiment, prior information can be prior knowledge of resource allocation mined by the resource allocation provider or industry experts based on historical resource allocation data. This prior information is applied to actual resource allocation scenarios, enabling targeted resource allocation. Prior information can characterize the relationship between the allocated object and the target item that the allocated object is interested in. Therefore, resource allocation based on prior information can deliver resources containing the target item to the allocated object. In this embodiment, prior information can characterize the relationship between the target object and the first item. The target object can be an object that has performed resource operations on historical allocated resources containing the second item. These resource operations can include resource transformation, etc. There may not be an explicit relationship between the first item and the second item. For example, if the target object has performed resource operations on historical allocated resources containing item 2, data mining of historical resource allocation data can reveal that the target object may be interested in item 1. This generates prior information indicating that an object that has performed resource operations on allocated resources containing item 2 has a high probability of being interested in item 1.

[0043] Furthermore, the resources deployed in this embodiment can be in the form of images and text, or in the form of video. The deployed resources may include item description information of the corresponding items, which may include information in the form of text, images, videos, audio, etc. The items may be real items or virtual items, and virtual items may include e-books, game equipment, online software, video memberships, music memberships, etc.

[0044] S220. Based on the recall model, perform resource recall processing for the target object to obtain the first recalled resource.

[0045] The recall model can filter out resources that match the target object's attribute information and resource operation information from a large pool of resources to be deployed, thus narrowing down the scope of resources to be deployed and providing more accurate input for the subsequent ranking model. Resource operation information can include information on resources for which the target object has already performed resource operations. In the absence of prior information, the first recalled resource can include resources related to the second item (i.e., the recall includes resources for the second item). The first recalled resource can also include resources related to the third item (i.e., the recall includes resources for the third item). The third item can be an item with an explicit relationship to the second item; for example, the third item can be the same as or similar to the second item, or it can be an item used in conjunction with the second item, or it can be an item belonging to the same major category as the second item. In the absence of prior information in the recall model, the first recall resources may not include deployment resources related to the first item; when the number of first recall resources is large, the first recall resources may also include a small number of deployment resources related to the first item. This embodiment does not make specific limitations.

[0046] S230. Based on the prior information, perform resource recall processing for the target object to obtain a second recalled resource.

[0047] Prior information can characterize the relationship between the target object and the target item that the target object is interested in. Therefore, resource delivery based on prior information can deliver delivery resources containing the target item to the target object.

[0048] Given that prior information indicates that the target object is associated with the first item, resources containing the first item can be selected from a large number of resources to be deployed, and then second recall resources can be obtained based on the selected resources containing the first item.

[0049] S240. Based on the first recalled resources and the second recalled resources, resources are deployed to obtain resource deployment result indicators.

[0050] Having obtained the first and second recalled resources, these resources can be comprehensively ranked to obtain a resource ranking result. Then, based on the resource ranking result, resources can be deployed to the target objects to obtain resource deployment result indicators. The resource deployment result indicators can be obtained by statistically analyzing the resource operation data of the target objects after resource deployment. In other words, the resource deployment result indicators can reflect the target objects' interest in the deployed resources, thereby reflecting the deployment effect of the resources.

[0051] S250. Determine the credibility of the prior information based on the resource allocation result indicators.

[0052] The higher the target audience's interest in the allocated resources, or the better the allocation effect of the allocated resources, the higher the credibility of the prior information; conversely, the lower the credibility of the prior information, the lower the credibility.

[0053] Specifically, the resource delivery result indicators include the actual click information of the target delivery resource and the actual conversion information of the target delivery resource; the target delivery resource is the delivery resource that includes the first item;

[0054] Determining the credibility of the prior information based on the resource allocation result indicators includes:

[0055] The credibility of the prior information is determined based on at least one of the actual click information of the target delivery resource and the actual conversion information of the target delivery resource.

[0056] The actual click information of the target resource can be the actual click-through rate of the target object to the target resource, and the actual conversion information of the target resource can be the actual conversion rate of the target object to the target resource. Therefore, the credibility of the prior information can be determined based on at least one of the following: the actual click-through rate of the target object to the target resource and the actual conversion rate.

[0057] In one example, the credibility of prior information can be determined based on the actual click-through rate (CTR) of the target object on the target resource. Alternatively, the actual CTR can be directly used as the credibility of prior information, or a pre-defined correspondence between CTR and credibility level can be obtained. This pre-defined correspondence can represent the correspondence between multiple credibility levels and multiple CTR intervals; that is, each credibility level corresponds to a CTR interval. For example, the higher the credibility level, the higher the CTR within the corresponding CTR interval, and vice versa. Therefore, the credibility level corresponding to the actual CTR can be determined based on the pre-defined correspondence.

[0058] In another example, the credibility of prior information can be determined based on the actual conversion rate of the target object to the target resources. The actual conversion rate of the target object to the target resources can be directly determined as the credibility of prior information, or a preset correspondence between conversion rate and credibility level can be obtained. Each credibility level can correspond to a conversion rate range, and then the credibility level corresponding to the actual conversion rate can be determined according to the preset correspondence.

[0059] In another example, the credibility of prior information can be determined based on the actual click-through rate (CTR) and conversion rate of the target audience to the target resources. The actual CTR and conversion rate can be fused to obtain target fused data. This target fused data can be directly used to determine the credibility of prior information, or a preset correspondence between the fused data and credibility levels can be obtained, where each credibility level corresponds to a fused data range. The credibility level corresponding to the target fused data can then be determined based on this preset correspondence. Target fused data can be obtained by summing the actual CTR and conversion rate, or by weighting the actual CTR and conversion rate.

[0060] In this embodiment, since the actual click-through rate and actual conversion rate can characterize the actual effect of the deployed resources and reflect the target audience's interest in the deployed resources, the credibility of prior information can be determined based on the actual click-through rate and actual conversion rate, thereby improving the rationality and accuracy of the credibility determination.

[0061] This application, upon obtaining prior information, can, on the one hand, recall first recall resource information for the target object based on a recall model, and on the other hand, recall second recall resource information for the target object based on prior information. Specifically, the second recall resource can be a delivery resource containing the first item. Therefore, when delivery is based on both the first and second recall resources, the delivery resources delivered to the target object can include those from both the first and second recall resources. Furthermore, based on the target object's resource operations on the delivered resources, corresponding resource delivery result indicators can be obtained. Determining the credibility of prior information through these resource delivery result indicators enables the evaluation and measurement of prior information, ensuring its long-term sustainable use. This facilitates resource delivery based on highly credible prior information, improving the accuracy and effectiveness of resource delivery. Moreover, by jointly delivering the first and second recall resources, the application avoids the problem of resource delivery results being skewed due to delivery solely based on the second recall resource, thus affecting the accuracy of the resource delivery result indicators and improving their overall accuracy.

[0062] In this embodiment, the prior information can be standardized and formatted so that different prior information can be applied to the same resource deployment system. For details, please refer to [link / reference needed]. Figure 3 It illustrates a method for generating prior information, which may include:

[0063] S310. Based on the data source of the resource operation, the operation type of the resource operation, the item tag of the second item, and the item identifier of the second item, generate the object information of the target object.

[0064] In a resource delivery scenario, the data source for the target object's resource operations on historically delivered resources can be resource conversion data; the operation type of the target object's resource operations on historically delivered resources can be viewing or converting delivered resources containing a second item, i.e., purchasing the second item; the item tag of the second item can be the item category level to which the second item belongs; the item identifier of the second item can be the name of the second item or the item category name of the second item; thus, the object information of the target object can be generated. The object information of any object can be represented as <data source of resource operation, operation type of resource operation, item tag of second item, item identifier of second item>. Specifically, the object information of the target object can be represented as <resource conversion data, item purchase, second-level product category, item 2>.

[0065] S320. Generate item information for the first item based on the item tag and item identifier of the first item.

[0066] The item tag of the first item can be the item category level to which the first item belongs; the item identifier of the first item can be the name of the first item or the item category name of the first item, etc. The item information of any item can be represented as <item tag of the first item, item identifier of the first item>, and the item information of the first item can be represented as <second-level category of goods, item 1>.

[0067] S330. Associate the object information of the target object with the item information of the first item to obtain the prior information.

[0068] The aforementioned prior information can be obtained by establishing the association between the object information of the target object and the item information of the first item.

[0069] Please see Figure 4 The diagram illustrates the prior information configuration, which includes an object information area 410, an item information area 420, and a confirmation touch information 430. In the object information area, the data source of the resource operation, the operation type of the resource operation, the item label of the second item, and the item identifier of the second item can be selected and configured. In the item information area, the item label of the first item and the item identifier of the first item can be selected and configured. When the configuration is complete, in response to the trigger operation of the confirmation touch information, the association between the object information and the item information can be established, and the corresponding prior information can be generated.

[0070] By standardizing and formatting prior information, it can be understood by the resource deployment system, so that different prior information can be applied to the same resource deployment system. Furthermore, prior information can be configured according to different business scenarios, which improves the efficiency and convenience of resource deployment based on prior information.

[0071] As can be seen from the above, given the first and second recalled resources, they can be comprehensively ranked, and then resources can be deployed based on the ranking results. When ranking resources, they can be ranked based on the resource value information corresponding to each recalled resource. The first recalled resource corresponds to first resource value information; the second recalled resource corresponds to second resource value information; and the second resource value information is determined based on basic resource value information and additional resource value information.

[0072] The resource allocation based on the first and second recalled resources, and the resulting resource allocation result indicators, include:

[0073] Based on the first resource value information corresponding to the first recalled resource and the second resource value information corresponding to the second recalled resource, the first recalled resource and the second recalled resource are sorted to obtain the resource sorting result;

[0074] Based on the resource ranking results, resources are allocated to obtain the resource allocation result indicators.

[0075] In this embodiment, each resource to be deployed has corresponding resource value information. Therefore, when ranking the recalled resources, the ranking can be based on the resource value information corresponding to each recalled resource. For the first recalled resource corresponding to the recall model, it corresponds to first resource value information. For the second recalled resource corresponding to the prior information, it corresponds to basic resource value information and additional resource value information; that is, the second recalled resource also has corresponding additional resource value information, so as to support the second recalled resource during the resource ranking process. Given the resource ranking results, the resources ranked higher in the resource ranking results can be deployed.

[0076] Please see Figure 5 The diagram illustrates the resource deployment process, where a first batch of resources is recalled based on a recall model, and a second batch of resources is recalled based on prior information. These resources undergo targeted filtering to obtain filtered resources. Finally, the filtered resources are sorted to obtain sorted resources. At the resource sorting node, the resource value information of the first batch of resources is the basic resource value information, and the resource value information of the second batch of resources is the basic resource value information plus additional resource value information.

[0077] When ranking the first and second recalled resources, additional resource value information can be added to the second recalled resource to improve its overall resource value. This can increase the ranking of the second recalled resource, increase its probability of being deployed, and thus increase its exposure rate. This facilitates the analysis of the deployment results after the second recalled resource is exposed and the determination of the credibility of prior information. In other words, it provides more reliable resource deployment result data for credibility analysis and improves the accuracy of credibility analysis.

[0078] As can be seen from the above, the resource delivery result indicators include the actual click information of the target delivery resource and the actual conversion information of the target delivery resource; the target delivery resource is the delivery resource that includes the first item;

[0079] The method further includes:

[0080] Within a preset delivery time period, if at least one of the following conditions is met: the actual click information of the target delivery resource meets a preset click condition, and the actual conversion information of the target delivery resource meets a preset conversion condition, training samples are generated based on at least one of the following: the actual click information of the target delivery resource and the actual conversion information of the target delivery resource.

[0081] The recall model is trained based on the training samples to obtain the trained recall model; the trained recall model has the ability to recall resources based on the prior information.

[0082] During the resource deployment process based on the first and second recall resources, a preset deployment time period can also be set. That is, deployment is carried out based on the first and second recall resources within the preset deployment time period. The preset deployment time period can be 7 days, 30 days, etc. In this embodiment, the actual click information can be the actual click-through rate, and the actual conversion information can be the actual conversion rate. Correspondingly, the preset click condition can be the preset click-through rate, and the preset conversion condition can be the preset conversion rate. For the deployed resources, the longer the deployment duration, the more likely the corresponding actual click-through rate and actual conversion rate will increase, or they may remain unchanged. This allows for the analysis of campaign data within a set, unified, preset campaign period. Within that timeframe, if the actual clicks on the target campaign resource meet at least one of the preset click conditions and the actual conversion information meets at least one of the preset conversion conditions, training samples can be generated based on the current actual clicks and conversion information. Alternatively, at the end of the preset campaign period, if the actual clicks on the target campaign resource meet at least one of the preset click conditions and the actual conversion information meets at least one of the preset conversion conditions, training samples can be generated based on the current actual clicks and current actual conversion information at the end of the preset campaign period. For a target delivery resource that meets at least one of the following conditions within a preset delivery time period: actual click information meets preset click conditions, and actual conversion information meets preset conversion conditions, the prior information corresponding to the target delivery resource can be determined to meet the validity conditions. For a target delivery resource that does not meet the preset click conditions or the preset conversion conditions within a preset delivery time period, the prior information corresponding to the target delivery resource can be determined to not meet the validity conditions. For prior information that meets the validity conditions, corresponding training samples can be generated and the credibility can be determined based on the resource delivery result indicators of the target delivery resource corresponding to the prior information.

[0083] The preset delivery time period can be seen as an exploration period for prior information, which can be understood as a period of exploring the delivery effect of the delivery resources corresponding to the prior information. For details, please refer to [link / reference]. Figure 6 It illustrates a flowchart of the exploration process for prior information. Taking the exploration of prior information based on actual click information and actual conversion information as an example, the method specifically includes:

[0084] S610. Obtain the current actual click information and the current actual conversion information of the target delivery resources.

[0085] S620. Determine whether the current actual click information meets the preset click conditions and whether the current actual conversion information meets the preset conversion conditions; if not, proceed to step S630; if yes, proceed to step S650.

[0086] S630. Determine whether the preset delivery time period has been reached; if yes, proceed to step S640; if no, proceed to step S610.

[0087] S640. It is determined that the exploration of prior information has failed.

[0088] S650. Confirmation that the prior information exploration was successful.

[0089] In this embodiment, since resource deployment consumes system resources, the deployment resources corresponding to prior information will not be explored indefinitely. The length of the exploration time window is limited by setting a preset deployment time period. If at least one of the following conditions is met: the actual click information of the target deployment resource meets the preset click condition, and the actual conversion information of the target deployment resource meets the preset conversion condition, it can be considered that enough samples have been accumulated; otherwise, the exploration will continue until the preset deployment time period ends. This allows the resource deployment model to be trained based on the samples corresponding to the successfully explored prior information, resulting in a resource deployment model with the ability to recall resources based on prior information.

[0090] When training samples are generated based on at least one piece of information from the actual conversion information of the target deployment resources, the recall model can be trained based on the training samples; please refer to [link to details]. Figure 7 It illustrates the training method for the recall model, including:

[0091] S710. Determine click tags based on the actual click information of the target delivery resources, and determine conversion tags based on the actual conversion information of the target delivery resources.

[0092] Based on the above, within the preset delivery time period, when at least one of the following conditions is met: the actual click information of the target delivery resource meets the preset click condition, and the actual conversion information of the target delivery resource meets the preset conversion condition, training samples are generated based on at least one of the actual click information and the actual conversion information of the target delivery resource. It can be seen that the actual click information of the target delivery resource in the training samples all meet the preset click condition, and the actual conversion information of the target delivery resource in the training samples also meets the preset conversion condition. Therefore, the click tag corresponding to the target delivery resource can be determined as 1 or true, and the conversion tag corresponding to the target delivery resource can be determined as 1 or true.

[0093] S720. Generate the training sample based on at least one of the click data sample and the conversion data sample; the click data sample includes the actual click information of the target delivery resource and the click tag, and the conversion data sample includes the actual conversion information of the target delivery resource and the conversion tag.

[0094] In the training samples, for each target resource, there can be corresponding actual click information and click tags, or actual conversion information and actual conversion tags, or actual click information, click tags, actual conversion information, and actual conversion tags.

[0095] S730. Determine the actual click information of the target delivery resource and the first loss information of the click tag.

[0096] The actual click information of the target resource can be the actual click-through rate of the target resource. Therefore, the first loss information can be the difference between the actual click information of the target resource and the click tag. That is, the first loss information can be the difference between the actual click-through rate of the target resource and the click tag. Taking the click tag as 1 as an example, the first loss information = 1 - actual click-through rate.

[0097] S740. Determine the actual conversion information of the target deployment resources and the second loss information of the conversion tag.

[0098] The actual conversion information of the target resources can be the actual conversion rate of the target resources. Therefore, the first loss information can be the difference between the actual conversion information of the target resources and the conversion tag. That is, the first loss information can be the difference between the actual conversion rate of the target resources and the conversion tag. Taking the conversion tag as 1 as an example, the first loss information = 1 - actual conversion rate.

[0099] S750. The recall model is trained based on at least one of the first loss information and the second loss information to obtain the trained recall model.

[0100] The recall model can be trained solely based on the first loss information to obtain the trained recall model; it can also be trained solely based on the second loss information to obtain the trained recall model; or it can be trained based on both the first and second loss information to obtain the trained recall model. When training the recall model based on the first and second loss information, it can be based on the sum of the first and second loss information, or it can be based on the weighted sum of the first and second loss information. The weight values ​​can be determined based on the specific application scenario.

[0101] Furthermore, the resource delivery system may include a recall model, a coarse-ranking model, and a fine-ranking model, etc., so that the recall model, coarse-ranking model, and fine-ranking model can be trained simultaneously based on the first loss information and the second loss information, thereby obtaining the trained recall model, the trained coarse-ranking model, and the trained fine-ranking model, etc.

[0102] In this embodiment, for a target delivery resource that meets at least one of the following conditions within a preset delivery time period: actual click information meets preset click conditions and actual conversion information meets preset conversion conditions, the prior information corresponding to the target delivery resource can be determined to meet the validity conditions. For prior information that meets the validity conditions, corresponding training samples can be generated based on the resource delivery result indicators of the target delivery resource corresponding to the prior information, so as to train models such as the recall model. This enables the trained recall model to have the ability to recall resources based on prior information. Moreover, the training samples are obtained from the actual resource delivery process without the need for other means, thereby achieving targeted training of the recall model and improving the efficiency of recall model training.

[0103] Furthermore, this application also provides a method for determining the matching between deployed resources and the targeted audience; please refer to [link to relevant documentation]. Figure 8 The method may include:

[0104] S810. Obtain the first click information and the first conversion information of the target delivery resource in the target object; the target delivery resource is a delivery resource that includes the first item.

[0105] By delivering targeted resources to the target audience, the system can obtain the target audience's first click and first conversion information regarding those resources. The targeted resources are determined based on prior information and include the first item.

[0106] S820. Obtain the second click information and second conversion information of the target resource in the full object.

[0107] By delivering the target resources to all objects, we can obtain the second click information and the second conversion information of all objects to the target resources.

[0108] It should be noted that, in order to ensure data synchronization, the first click information, the first conversion information, the second click information, and the second conversion information can all be information obtained within the same time period.

[0109] S830. If the first fusion information is greater than the second fusion information, determine that the target delivery resource matches the target object; the first fusion information is obtained by fusing the first click information and the first conversion information; the second fusion information is obtained by fusing the second click information and the second conversion information.

[0110] There can be multiple target delivery resources, so when performing information fusion, the first fusion information can be calculated based on the first click information and the first conversion information corresponding to each of the multiple target delivery resources, and the second fusion information can be calculated based on the second click information and the second conversion information corresponding to each of the multiple target delivery resources.

[0111] In one example, when calculating the first fusion information, the first click information and the first conversion information of each target delivery resource can be summed to obtain the first sum information. Then, the first sum information corresponding to each of the multiple target delivery resources can be accumulated to obtain the first fusion information. When calculating the second fusion information, the second click information and the second conversion information of each target delivery resource can be summed to obtain the second sum information. Then, the second sum information corresponding to each of the multiple target delivery resources can be accumulated to obtain the second fusion information.

[0112] In another example, when calculating the first fusion information, the first click information and the first conversion information of each target resource can be multiplied to obtain the first product information. Then, the first product information corresponding to each of the multiple target resources is accumulated to obtain the first fusion information. When calculating the second fusion information, the second click information and the second conversion information of each target resource can be multiplied to obtain the second product information. Then, the second product information corresponding to each of the multiple target resources is accumulated to obtain the second fusion information.

[0113] Taking the product of click information and conversion information as an example, the formula is as follows:

[0114] ∑CTR1×CVR1>∑CTR2×CVR2,whereclick≥n (1)

[0115] Wherein, CTR1 is the first click information corresponding to the target object, CVR1 is the first conversion information corresponding to the target object, CTR2 is the second click information corresponding to the target object, CVR2 is the second conversion information corresponding to the target object, and n is the click count threshold. That is, when the number of clicks on the target resource is greater than or equal to n, it is reasonable to determine that the first click information, the first conversion information, the second click information, and the second conversion information are reasonable. Therefore, it is possible to determine whether the first fused information is greater than the second fused information based on the first click information, the first conversion information, the second click information, and the second conversion information.

[0116] In this embodiment, after the target resources are deployed, the effect of the prior information corresponding to the target resources can be measured. For a certain prior information of "object-item" matching, for the deployment resources involved in the item lock under the prior information, the click-through rate and conversion rate of the target object specified by the prior information can be compared with those of all exposed objects. When the combined information of the click-through rate and conversion rate of the target object is greater than that of the combined information of the click-through rate and conversion rate of the overall objects, it can be determined that the prior information is a reasonable prior information. It can be understood that for a certain item, the prior information has found a matching target for it, thereby improving the click-through rate and conversion rate after the resources are deployed.

[0117] This application provides a general framework and process for industry marketers to input expert prior information in e-commerce scenarios, explore prior information within the campaign system, and measure and evaluate the effectiveness of prior information. Specifically, it employs structured "object-item" matching expert prior information and guides industry operators to rationally allocate resources and input prior information by applying and evaluating the input prior information in a controllable and responsive manner within the campaign system. Within this framework, specific details, such as the definition of "object-item" matching prior information, the time window for exploration within the campaign system, and the threshold parameters for the completion and effectiveness of prior information exploration, can be differentiated according to actual business scenarios.

[0118] It should be noted that any of the methods described above in this embodiment can be combined based on the actual implementation situation and have corresponding beneficial effects, which will not be elaborated here.

[0119] Please see Figure 9 This embodiment also provides a resource processing device, including:

[0120] The prior information acquisition module 910 is used to acquire prior information; the prior information indicates that the target object is associated with the first item, and the target object has performed resource operations on historical resources containing the second item.

[0121] The first recall module 920 is used to perform resource recall processing for the target object based on the recall model to obtain the first recalled resource;

[0122] The second recall module 930 is used to perform resource recall processing for the target object based on the prior information to obtain the second recalled resources.

[0123] The resource deployment module 940 is used to deploy resources based on the first recalled resources and the second recalled resources, and obtain resource deployment result indicators.

[0124] The credibility determination module 950 is used to determine the credibility of the prior information based on the resource deployment result indicators.

[0125] Furthermore, the device also includes:

[0126] The object information generation module is used to generate object information of the target object based on the data source of the resource operation, the operation type of the resource operation, the item tag of the second item, and the item identifier of the second item;

[0127] The item information generation module is used to generate item information of the first item based on the item tag and the item identifier of the first item;

[0128] The prior information is obtained by associating the object information of the target object with the item information of the first item.

[0129] Furthermore, the first recalled resource corresponds to first resource value information; the second recalled resource corresponds to second resource value information; the second resource value information is determined based on basic resource value information and additional resource value information.

[0130] The resource delivery module includes:

[0131] The sorting module is used to sort the first recalled resource and the second recalled resource based on the first resource value information corresponding to the first recalled resource and the second resource value information corresponding to the second recalled resource, so as to obtain the resource sorting result;

[0132] The resource allocation result acquisition module is used to allocate resources based on the resource ranking results and obtain the resource allocation result indicators.

[0133] Furthermore, the resource delivery result indicators include the actual click information of the target delivery resource and the actual conversion information of the target delivery resource; the target delivery resource is the delivery resource containing the first item;

[0134] The credibility determination module includes:

[0135] The first determining module is used to determine the credibility of the prior information based on at least one of the actual click information of the target delivery resource and the actual conversion information of the target delivery resource.

[0136] Furthermore, the resource delivery result indicators include the actual click information of the target delivery resource and the actual conversion information of the target delivery resource; the target delivery resource is the delivery resource containing the first item;

[0137] The device further includes:

[0138] The training sample generation module is used to generate training samples based on at least one of the following conditions: the actual click information of the target delivery resource meets a preset click condition, and the actual conversion information of the target delivery resource meets a preset conversion condition, within a preset delivery time period.

[0139] The first training module trains the recall model based on the training samples to obtain the trained recall model; the trained recall model has the ability to recall resources based on the prior information.

[0140] Furthermore, the training sample generation module includes:

[0141] The tag determination module is used to determine click tags based on the actual click information of the target delivery resources and to determine conversion tags based on the actual conversion information of the target delivery resources.

[0142] The training sample determination module is used to generate the training sample based on at least one of click data samples and conversion data samples; the click data sample includes the actual click information of the target delivery resource and the click tag, and the conversion data sample includes the actual conversion information of the target delivery resource and the conversion tag.

[0143] The first training module includes:

[0144] The second determining module is used to determine the actual click information of the target delivery resource and the first loss information of the click tag;

[0145] The first determining module is used to determine the actual conversion information of the target deployment resources and the second loss information of the conversion tag;

[0146] The second training module is used to train the recall model based on at least one of the first loss information and the second loss information to obtain the trained recall model.

[0147] Furthermore, the device also includes:

[0148] The first information acquisition module is used to acquire the first click information and the first conversion information of the target delivery resource in the target object; the target delivery resource is a delivery resource that includes the first item;

[0149] The second information acquisition module is used to acquire the second click information and the second conversion information of the target delivery resource in the full object;

[0150] The fusion information comparison module is used to determine that the target delivery resource matches the target object when the first fusion information is greater than the second fusion information; the first fusion information is obtained by fusing the first click information and the first conversion information; the second fusion information is obtained by fusing the second click information and the second conversion information.

[0151] The apparatus provided in the above embodiments can execute the methods provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in the above embodiments can be found in the methods provided in any embodiment of this application.

[0152] This embodiment also provides a computer-readable storage medium storing at least one instruction or at least one program, which is loaded by a processor and executed as any of the methods described above in this embodiment.

[0153] According to one aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform any of the methods described above.

[0154] Figure 10 This is a block diagram illustrating an electronic device for resource processing according to an exemplary embodiment. The electronic device may be a server, and its internal structure diagram may be as follows: Figure 10As shown, this electronic device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a resource processing method.

[0155] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present disclosure and does not constitute a limitation on the electronic device to which the present disclosure is applied. A specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0156] This specification provides the operational steps of the methods described in the embodiments or flowcharts, but more or fewer operational steps may be included based on conventional or non-inventive labor. The steps and order listed in the embodiments are merely one possible execution order among many steps and do not represent the only execution order. In actual system or interrupt product execution, the methods shown in the embodiments or drawings can be executed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment).

[0157] The structure shown in this embodiment is only a partial structure related to the solution of this application and does not constitute a limitation on the device to which the solution of this application is applied. Specific devices may include more or fewer components than shown, or combinations of certain components, or arrangements of different components. It should be understood that the methods, apparatuses, etc., disclosed in this embodiment can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of modules 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 an indirect coupling or communication connection between devices or unit modules through some interfaces.

[0158] 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 described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0159] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this specification can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0160] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A resource processing method, characterized in that, include: Obtain prior information; the prior information indicates that the target object is associated with the first item, and the target object has performed resource operations on historical resources containing the second item. Based on the recall model, resource recall processing is performed on the target object to obtain the first recalled resource; Based on the prior information, resource recall processing is performed on the target object to obtain a second recalled resource; Based on the first and second recalled resources, resources are deployed to obtain resource deployment result indicators; The credibility of the prior information is determined based on the resource allocation result indicators.

2. The method according to claim 1, characterized in that, Before obtaining the prior information, the method further includes: Based on the data source of the resource operation, the operation type of the resource operation, the item tag of the second item, and the item identifier of the second item, the object information of the target object is generated; Based on the item tag and item identifier of the first item, generate the item information of the first item; The prior information is obtained by associating the object information of the target object with the item information of the first item.

3. The method according to claim 1, characterized in that, The first recalled resource corresponds to first resource value information; the second recalled resource corresponds to second resource value information; the second resource value information is determined based on basic resource value information and additional resource value information. The resource allocation based on the first and second recalled resources, and the resulting resource allocation result indicators, include: Based on the first resource value information corresponding to the first recalled resource and the second resource value information corresponding to the second recalled resource, the first recalled resource and the second recalled resource are sorted to obtain the resource sorting result; Based on the resource ranking results, resources are allocated to obtain the resource allocation result indicators.

4. The method according to claim 1, characterized in that, The resource delivery result indicators include the actual click information of the target delivery resource and the actual conversion information of the target delivery resource; the target delivery resource is the delivery resource that includes the first item; Determining the credibility of the prior information based on the resource allocation result indicators includes: The credibility of the prior information is determined based on at least one of the actual click information of the target delivery resource and the actual conversion information of the target delivery resource.

5. The method according to claim 1, characterized in that, The resource delivery result indicators include the actual click information of the target delivery resource and the actual conversion information of the target delivery resource; the target delivery resource is the delivery resource that includes the first item; The method further includes: Within a preset delivery time period, if at least one of the following conditions is met: the actual click information of the target delivery resource meets a preset click condition, and the actual conversion information of the target delivery resource meets a preset conversion condition, training samples are generated based on at least one of the following: the actual click information of the target delivery resource and the actual conversion information of the target delivery resource. The recall model is trained based on the training samples to obtain the trained recall model; the trained recall model has the ability to recall resources based on the prior information.

6. The method according to claim 5, characterized in that, The step of generating training samples based on at least one of the actual click information and the actual conversion information of the target delivery resource includes: Click tags are determined based on the actual click information of the target delivery resources, and conversion tags are determined based on the actual conversion information of the target delivery resources. The training sample is generated based on at least one of the click data sample and the conversion data sample; the click data sample includes the actual click information of the target delivery resource and the click tag, and the conversion data sample includes the actual conversion information of the target delivery resource and the conversion tag; The step of training the recall model based on the training samples to obtain the trained recall model includes: Determine the actual click information of the target resource and the first loss information of the click tag; Determine the actual conversion information of the target resources and the second loss information of the conversion tag; The recall model is trained based on at least one of the first loss information and the second loss information to obtain the trained recall model.

7. The method according to claim 1, characterized in that, The method further includes: Obtain the first click information and first conversion information of the target delivery resource in the target object; the target delivery resource is a delivery resource that includes the first item; Obtain the second click information and second conversion information of the target resource in the full object; If the first fused information is greater than the second fused information, it is determined that the target delivery resource matches the target object; the first fused information is obtained by fusing the first click information and the first conversion information; the second fused information is obtained by fusing the second click information and the second conversion information.

8. A resource processing device, characterized in that, include: The prior information acquisition module is used to acquire prior information; The prior information indicates that the target object is associated with the first item, and the target object has performed resource operations on historical resources containing the second item. The first recall module is used to perform resource recall processing for the target object based on the recall model to obtain the first recalled resource. The second recall module is used to perform resource recall processing for the target object based on the prior information to obtain the second recalled resources. The resource deployment module is used to deploy resources based on the first recalled resources and the second recalled resources, and obtain resource deployment result indicators. The credibility determination module is used to determine the credibility of the prior information based on the resource deployment result indicators.

9. An electronic device, characterized in that, The device includes a processor and a memory, the memory storing at least one instruction or at least one program, the at least one instruction or the at least one program being loaded and executed by the processor to implement the resource processing method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The storage medium stores at least one instruction or at least one program, which is loaded by a processor and executed by the resource processing method as described in any one of claims 1 to 7.

Images