An object recommendation method, apparatus and device

Abstract

Embodiments of the present specification provide an object recommendation method, device and equipment, wherein the method comprises: obtaining a first candidate object and a second candidate object to be recommended, wherein the historical exposure times of the first candidate object are less than a preset number threshold, based on a preset quota and a traffic regulation target corresponding to the click rate requirement of the first candidate object, the first candidate object is subjected to screening processing to obtain a screened first candidate object, based on a preset display ratio and the number of the second candidate object, the screened first candidate object is subjected to screening processing to obtain a target display object, based on the target display object and the second candidate object, an object recommendation result is determined, and based on the object recommendation result, an object recommendation processing is performed on a user.

Description

Technical Field

[0001] This document relates to the field of computer technology, and in particular to an object recommendation method, apparatus, and device. Background Technology

[0002] With the rapid development of computer technology and financial services, financial leasing services such as credit leasing have become an important part of the modern service industry. To improve the success rate of financial leasing services, services such as product recommendation and product search can be provided to users.

[0003] However, in scenarios such as product recommendation and product retrieval, popular products continue to receive more exposure due to their historical data advantages, while mid-to-long-tail products with lower exposure struggle to gain traffic, resulting in uneven traffic distribution, poor product recommendation flexibility, and low product recommendation accuracy. Therefore, it is necessary to provide a more reliable object recommendation solution to improve the accuracy of object recommendations. Summary of the Invention

[0004] The purpose of the embodiments in this specification is to provide a more reliable object recommendation scheme to improve the accuracy of object recommendations.

[0005] To achieve the above technical solution, the embodiments in this specification are implemented as follows: This specification provides an object recommendation method, which includes: obtaining a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold; filtering the first candidate object based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object to obtain a filtered first candidate object; filtering the filtered first candidate object based on a preset display ratio and the number of the second candidate object to obtain a target display object; determining an object recommendation result based on the target display object and the second candidate object, and recommending the object to the user based on the object recommendation result.

[0006] This specification provides an object recommendation device, comprising: an object acquisition module for acquiring a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold; a first filtering module for filtering the first candidate object based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object, to obtain a filtered first candidate object; a second filtering module for filtering the filtered first candidate object based on a preset display ratio and the number of the second candidate object, to obtain a target display object; and a result determination module for determining an object recommendation result based on the target display object and the second candidate object, and performing object recommendation processing to the user based on the object recommendation result.

[0007] This specification provides an object recommendation device, comprising: a processor; and a memory configured to store computer-executable instructions, wherein, when executed, the processor: acquires a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold; filters the first candidate object based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object, to obtain a filtered first candidate object; filters the filtered first candidate object based on a preset display ratio and the number of the second candidate object, to obtain a target display object; determines an object recommendation result based on the target display object and the second candidate object, and performs object recommendation processing to the user based on the object recommendation result.

[0008] This specification also provides a storage medium for storing computer-executable instructions. When executed by a processor, the executable instructions implement the following process: obtaining a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold; filtering the first candidate object based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object to obtain a filtered first candidate object; filtering the filtered first candidate object based on a preset display ratio and the number of the second candidate objects to obtain a target display object; determining the object recommendation result based on the target display object and the second candidate object, and recommending the object to the user based on the object recommendation result.

[0009] This specification also provides a computer program product, including a computer program that, when executed by a processor, performs the following process: obtaining a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold; filtering the first candidate object based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object to obtain a filtered first candidate object; filtering the filtered first candidate object based on a preset display ratio and the number of the second candidate object to obtain a target display object; determining the object recommendation result based on the target display object and the second candidate object, and performing object recommendation processing to the user based on the object recommendation result. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Figure 1 This is a schematic diagram illustrating the implementation environment of one object recommendation method described in this specification; Figure 2 This is a flowchart illustrating the processing procedure of an object recommendation method described in this specification. Figure 3 This is a flowchart illustrating a method for determining a first candidate object according to this specification; Figure 4 This is a schematic diagram of a recommendation system described in this specification; Figure 5 This is a flowchart illustrating the generation process of a traffic control task as described in this specification. Figure 6 This is a flowchart illustrating another method for determining the first candidate object in this specification. Figure 7 This is a flowchart illustrating yet another method for determining the first candidate object in this specification; Figure 8 This is a flowchart illustrating the updating process of a traffic control label as described in this specification. Figure 9 This is a flowchart illustrating the process of determining a target object in this specification. Figure 10 This is a schematic diagram of an object recommendation device according to this specification; Figure 11 This is a schematic diagram of a recommended device for this application. Detailed Implementation

[0011] This specification provides an object recommendation method, apparatus, and device through its embodiments.

[0012] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0013] This specification provides a more reliable object recommendation scheme to improve the accuracy of object recommendations. With the rapid development of computer technology and financial services, financial leasing businesses such as credit leasing have become an important part of the modern service industry, resulting in explosive growth on the supply side. However, product recommendation systems suffer from a severe Matthew effect, where top-selling products occupy the majority of traffic (e.g., the top 200 rental items account for 78% of total exposure). This prevents a large number of new products and potentially valuable mid-to-long-tail products from being effectively distributed. This long-term traffic monopoly discourages merchants from submitting new products, hindering the prosperity of the leasing ecosystem and damaging its health. Furthermore, because users repeatedly see similar popular products, their opportunities to explore novel and cost-effective products are reduced, impacting the diversity of the user experience. Meanwhile, the business operations side also has significant traffic control needs, including routine cold starts for new products (i.e., newly listed products or products with low historical exposure and lack of user behavior data, requiring traffic support to initiate distribution and data accumulation), boosting potential products, and providing traffic support to partner merchants and brands during pre-planned marketing periods. Existing product recommendation systems lack a highly deterministic, near real-time, and scalable traffic control framework to support these needs. In practical applications, traffic control scenarios, tagging rules, and traffic control thresholds can be configured to obtain product metrics in real time and assign tags. When real-time product metrics exceed or fall below the threshold, weighting, deweighting, and filtering operations are performed to accurately target products requiring adjustment through the tagging system. However, because the calculation of traffic control thresholds relies on accurate initial data, high data quality is required. Furthermore, when a large number of products require simultaneous traffic control, the data processing volume becomes excessive, leading to inaccurate recommendation results. Alternatively, the cold start problem of recommendation systems can be alleviated through content feature extraction. Specifically, during content feature extraction, dependency parsing from natural language processing can be used to extract descriptive information features of products, converting them into word vectors. Then, a weighted word distance algorithm optimized based on TF-IDF is used to calculate the word distance between the content feature vectors. Finally, the similarity calculated by word distance can be combined with traditional similarity calculation methods for recommendation. However, this approach relies on a high-quality corpus and a well-trained word vector matrix, resulting in high computational resource requirements. Furthermore, dependency parsing is not mature enough for processing Chinese text and requires further optimization. It also does not consider issues such as potential product boosting. Therefore, this specification provides a more reliable object recommendation scheme. Firstly, by setting a quota, it ensures that mid-to-long-tail products with low exposure are not completely filtered out in subsequent object selection. Secondly, by setting a display ratio and the number of second candidate objects, the first candidate objects are filtered. This ensures that the control objectives are smoothly achieved while maximizing the preservation of personalized recommendation capabilities, balancing operational goals and user experience, and improving the accuracy of object recommendations.For specific processing details, please refer to the following embodiments.

[0014] The object recommendation method described in one or more embodiments of this specification is applicable to the object recommendation implementation environment, such as... Figure 1 As shown, the implementation environment includes at least: Client 100 and server 200. Furthermore, server 200 can be configured with various network models and algorithms, among which: Client 100 can run on terminal devices, which can be mobile phones, personal computers, tablets, e-book readers, wearable devices, devices that interact with information based on AR (Augmented Reality) and VR (Virtual Reality), and laptop computers, etc. Client 100 can be installed on terminal devices, and Client 100 can be an application, a browser, or a subroutine embedded in an application, etc.

[0015] Server 200 can run on a server, which can be one or more servers, a server cluster consisting of several servers, or a cloud server on a cloud computing platform. Server 200 can be installed on the server. Server 200 can be an application or a subroutine embedded in an application. Various network models and algorithms can be integrated into server 200, or server 200 can call one or more of various network models and algorithms to perform corresponding operations.

[0016] In addition, it may include a database 300, which may be set in the server on which the server 200 runs or outside the server on which the server 200 runs. The database 300 may store a first candidate object, a second candidate object, a target candidate object carrying a traffic control label, and the traffic control task corresponding to the target candidate object.

[0017] In this implementation environment, server 200 can obtain the first candidate object and the second candidate object to be recommended. The historical exposure count of the first candidate object can be less than a preset threshold. Based on the preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate object, the first candidate object is filtered to obtain the filtered first candidate object. Based on the preset display ratio and the number of the second candidate object, the filtered first candidate object is filtered to obtain the target display object. Based on the target display object and the second candidate object, the object recommendation result is determined, and server 200 can return the object recommendation result to client 100.

[0018] like Figure 2As shown in the embodiments of this specification, an object recommendation method is provided. The execution subject of this method can be a server, which can be a single independent server or a server cluster composed of multiple servers. The server can be a backend server for businesses such as credit leasing, commodity trading, and video viewing, or a backend server for a credit leasing application. Specifically, the method may include the following steps: In step S202, the first and second candidate objects to be recommended are obtained.

[0019] The first candidate object can have a historical exposure count less than a preset threshold, while the second candidate object can have a historical exposure count greater than the preset threshold. Both the first and second candidate objects can be any objects that can be recommended to the user. For example, the first candidate object can be rentable goods, tradable goods or services, watchable videos or text, etc.

[0020] In implementation, taking the first and second candidate objects as leasable goods as an example, the server can recall candidate objects corresponding to the user's product recommendation needs (or user's product search requests) based on a preset recall strategy (such as vector recall strategy, constant recall strategy, etc.). The recalled candidate objects are then subjected to coarse and fine ranking to determine the second candidate object. The coarse ranking stage can use a lightweight first screening model to perform preliminary screening of the recalled candidate objects based on the relevance between the candidate objects and the user. The fine ranking stage can use a second screening model with larger model parameters to further screen the preliminary candidate objects based on the relevance between the candidate objects and the user to obtain the processed candidate objects.

[0021] Meanwhile, the server can also add an independent recall channel to the recommendation system to recall first candidate objects related to the user's product recommendation needs, so as to realize separate recall processing for different types of candidate objects (i.e., second candidate objects with high historical exposure and first candidate objects with low historical report frequency), thereby improving the efficiency and accuracy of object recall.

[0022] The user's product recommendation needs can be determined based on their historical search information, rental information, and credit status. For example, assuming that the user's product recommendation needs, determined based on their historical search information, may include recommendations for rentable computer products, then the server can recall rentable computer products as second candidates using the aforementioned recall method. Simultaneously, it can also recall rentable computer products with a historical exposure count less than a preset threshold as first candidates using a separate recall channel.

[0023] Furthermore, the above-mentioned method for obtaining the first and second candidate objects is an optional and implementable method. In actual application scenarios, there can be a variety of different methods. Different methods can be selected according to different actual application scenarios. This specification does not specifically limit this embodiment.

[0024] In step S204, based on the preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate object, the first candidate object is filtered to obtain the filtered first candidate object.

[0025] The preset quota can be a fixed number of display slots determined according to the type of the first candidate. The type of the first candidate can include cold start products, booster products, etc. Cold start products can include newly listed products or products with little historical exposure and lack of user behavior data, which need traffic support to start their distribution and data accumulation. Booster products can include products with certain potential but have not yet become bestsellers, which need additional traffic support to accelerate their growth. The traffic control target can be the traffic required to achieve the click-through rate requirement.

[0026] In implementation, the server can determine the preset quota based on the object type of the first candidate object, the number of first candidate objects recalled, and the number of second candidate objects. For example, the preset quota for the first candidate object of the cold start product type can be 100, and the preset quota for the first candidate object of the booster product type can be 150.

[0027] The server can use a pre-trained third screening model to screen the first candidate object based on a preset quota, the click-through rate requirement of the first candidate object, and the traffic control target corresponding to the click-through rate requirement of the first candidate object, and obtain the screened first candidate object. The third screening model can be a model built based on a preset machine learning algorithm.

[0028] In addition, when there are multiple types of first candidate objects, the server can filter the first candidate objects for each type based on the preset quota and traffic control target for each type, and obtain the filtered first candidate objects.

[0029] In step S206, based on the preset display ratio and the number of second candidate objects, the first candidate objects after filtering are filtered to obtain the target display object.

[0030] In implementation, to balance the control objectives and user experience, the server can perform density control based on a preset display ratio. That is, the server can control the number of first candidate objects (i.e., target display objects) with fewer historical exposures in the object recommendation results based on the preset display ratio. In other words, the server can determine the number of target display objects based on the preset display ratio and the number of second candidate objects, and then filter the first candidate objects based on this number to obtain the target display objects. For example, assuming the preset display ratio is 1:4 and the number of second candidate objects is 8, then the number of target display objects can be 2, that is, the number of candidate objects included in the final object recommendation results can be 10.

[0031] In step S208, based on the target display object and the second candidate object, the object recommendation result is determined, and the object recommendation process is performed to the user based on the object recommendation result.

[0032] In practice, the server can perform deduplication on the target display object and the second candidate object, and determine the deduplicated candidate object as the object recommendation result.

[0033] In addition, the server can use a pre-trained relevance determination model to determine the relevance between each object in the object recommendation results and the user, and rank the objects according to the relevance to recommend the ranked object recommendations to the user.

[0034] This specification provides an object recommendation method. It involves acquiring a first candidate object and a second candidate object, where the historical exposure count of the first candidate object may be less than a preset threshold. Based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object, the first candidate object is filtered to obtain a filtered first candidate object. Then, based on a preset display ratio and the number of second candidate objects, the filtered first candidate object is further filtered to obtain a target display object. Based on the target display object and the second candidate object, the object recommendation result is determined, and the object recommendation is then sent to the user based on this result. This approach firstly ensures that low-exposure, mid-to-long-tail products are not completely filtered out in subsequent object selection by using a preset quota. Secondly, by using a preset display ratio and the number of second candidate objects to filter the filtered first candidate object, it ensures that the control target is smoothly achieved while maximizing the preservation of personalized recommendation capabilities. This balances operational goals and user experience, improving the accuracy of object recommendations.

[0035] In practical applications, the specific processing method for obtaining the first candidate object to be recommended in step S202 above can vary. The following provides one optional processing method, such as... Figure 3 As shown, the specific process may include the following steps S2022 to S2024.

[0036] In step S2022, the target candidate object carrying the traffic control label and the traffic control task corresponding to the target candidate object are obtained.

[0037] Among them, the traffic control task can include the traffic control target.

[0038] In practice, the server can pre-determine the target candidates that need traffic control based on each candidate's historical exposure count, online time, transaction count (such as the number of rentals of rentable goods, the number of times videos are watched), user feedback information (such as evaluation information, like information, and other interaction information), and then label the target candidates with traffic control tags.

[0039] In this way, when recommending objects, the server can quickly obtain target candidate objects that need to be subject to traffic regulation based on traffic regulation tags.

[0040] In addition, when the server filters out target candidates, it can also generate traffic control tasks corresponding to each target candidate based on the click-through rate requirements of that target candidate.

[0041] In step S2024, the target candidate objects are screened based on the task release time of the traffic control task to obtain the first candidate object to be recommended.

[0042] In implementation, in such Figure 4 The recommendation system shown may include a task generation module, an object recall module, a ranking and flow control module, a re-ranking and data tracking module, and a real-time data synchronization module. The task generation module can be used to filter out target candidate objects and generate flow control tasks corresponding to the subsequent objects of the target.

[0043] The object recall module can be used to filter out the first candidate object from the target candidate objects. That is, the recommendation system can use a newly added independent recall channel to recall objects associated with the current recommendation task (i.e., candidate objects used to determine the second candidate object, and recall the first candidate object from the target candidate objects). The recall methods for the first candidate object can include inverted index based on the task release time of the traffic control task, inverted index based on the quantity value, and personalized recall that matches the user's search needs.

[0044] In practical applications, the specific processing method for filtering target candidate objects based on the task release time of the traffic control task in step S2024 above to obtain the first candidate object to be recommended can be varied. The following provides an optional processing method, which may include the following steps A1 to A2.

[0045] In step A1, the target candidate objects are filtered based on the task release time of the traffic control task to obtain the first object.

[0046] In practice, the server can use the task release time of the traffic control task to perform a reverse index to obtain the first object.

[0047] In step A2, a pre-trained relevance determination model is used to determine the matching degree between the first object and the user, and the first object is filtered based on the matching degree to obtain the first candidate object to be recommended.

[0048] The relevance determination model can be a model built based on a preset machine learning algorithm. For example, the relevance determination model can include a feature extraction module and a relevance calculation model. That is, the feature extraction module can be used to extract feature vectors, and the relevance calculation module can be used to calculate the relevance based on the extracted feature vectors.

[0049] In implementation, taking leasable goods as a candidate, the server can obtain user information such as historical search information, historical rental information, and credit status information. It then uses the feature extraction module in the relevance determination model to extract feature vectors from the user information, obtaining a first vector. Simultaneously, the server can also obtain product information such as the product type, online time, and historical rental information of the first object, and use the feature extraction module in the relevance determination model to extract feature vectors from the product information, obtaining a second vector. Finally, the server can use the relevance calculation module in the relevance determination model, based on a preset similarity determination algorithm (such as cosine similarity, Euclidean distance, etc.), to determine the relevance between the first vector and the second vector. Based on this relevance, the matching degree between the first object and the user is determined.

[0050] In this way, the server can perform object filtering based on the matching degree to obtain the first candidate object to be recommended. For example, it can sort the objects based on the matching degree and determine the first candidate object based on the sorted first object. Alternatively, the server can also determine the first object with a matching degree greater than a preset matching degree threshold as the first candidate object.

[0051] In practical applications, traffic control tasks corresponding to target candidate objects can also be generated. The specific processing methods for generating traffic control tasks can vary. The following provides one optional processing method, such as... Figure 5 As shown, the specific process may include the following steps S502 to S506.

[0052] In step S502, the publication time and historical exposure count of each candidate object are obtained.

[0053] Among them, the release time can be the online time of the candidate, the historical exposure count can be the number of exposures from the release time to the present, or the historical exposure count can be the number of exposures within a preset detection period.

[0054] In step S504, candidate objects whose publication time is within a preset time range and / or whose historical exposure count is less than a preset threshold are identified as target candidate objects carrying traffic control tags.

[0055] In implementation, in such Figure 4 In the recommendation system shown, the task generation module can be used to filter out target candidate objects that need to be subject to traffic regulation based on the release time and historical exposure count of each candidate object, according to preset rules or algorithm models.

[0056] In step S506, a pre-trained task generation model is used to determine the traffic control target corresponding to the target candidate object based on the click-through rate requirement corresponding to the target candidate object, and a traffic control task corresponding to the target candidate object is generated based on the traffic control target.

[0057] The task generation model can be a model built based on a preset machine learning algorithm. For example, the task generation model can be a model built based on a Bayesian algorithm.

[0058] In implementation, the task generation module can use a pre-trained task generation model to determine the traffic control target corresponding to the target candidate object based on the click-through rate requirement. Then, based on the traffic control target, a traffic control task can be generated and written into the database and search engine.

[0059] Among them, the traffic control task can be the smallest atomic intervention strategy for traffic control. Its core contains a four-tuple of information, namely: object, target, location, and validity period. "Object" can refer to a specific supply instance (such as rentable goods), "target" can refer to the traffic indicator to be achieved (such as 1,000 exposures), that is, the traffic control target, "location" can refer to the specific scenario of traffic placement (such as the homepage recommendation flow of the rental platform), and "validity period" can refer to the time window for the execution of the traffic control task.

[0060] In practical applications, step S204 above filters the first candidate objects based on a preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate object. The specific processing methods for the filtered first candidate objects can vary. The following provides one optional processing method, such as... Figure 6 As shown, the specific process may include the following steps S2042 to S2046.

[0061] In step S2042, a first quantity corresponding to the preset quota is determined.

[0062] In practice, when the first candidate object contains a large number of object types, the first quantity corresponding to each object type can be obtained.

[0063] In step S2044, the first candidate object is filtered based on the matching degree between the first candidate object and the user, and the traffic control target corresponding to the click-through rate requirement of the first candidate object, to obtain the first filtering result.

[0064] In implementation, the server can construct the first screening result based on the first candidate object whose matching degree is greater than the preset matching degree threshold and / or whose traffic control target is greater than the preset traffic threshold.

[0065] In step S2046, the first candidate object after screening is determined based on the first quantity and the first screening result.

[0066] In implementation, in such Figure 4 In the recommendation system shown, the ranking and flow control module can first perform coarse and fine ranking on the first candidate objects. After coarse and fine ranking, the model scores are used to truncate the candidate set. For the first candidate objects with flow control tags, an independent quota mechanism can be applied to ensure that they have a fixed number of slots to enter the re-ranking stage. That is, the server can select the first candidate objects corresponding to the first quantity from the first screening results to obtain the filtered first candidate objects.

[0067] In practical applications, step S204 above filters the first candidate objects based on a preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate object. The specific processing methods for the filtered first candidate objects can vary. The following provides one optional processing method, such as... Figure 7 As shown, the specific process may include the following steps S2048 to S20410.

[0068] In step S2048, based on the preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate object, the first candidate object is filtered to obtain the second filtering result.

[0069] In practice, the method for determining the second screening result can refer to the method for determining the first screening result described above, and will not be repeated here.

[0070] In step S20410, a preset exposure count constraint is obtained, and based on the preset exposure count constraint and the exposure count of the first candidate object corresponding to the second screening result, the first candidate object corresponding to the second screening result is screened to obtain the screened first candidate object.

[0071] Among them, the preset exposure count constraint can be used to constrain the upper and / or lower limits of the exposure count, and the exposure count of the first candidate object corresponding to the second screening result can be the traffic control target.

[0072] In implementation, in such Figure 4In the recommendation system shown, the ranking and flow control module can also request the flow control center to make real-time flow control decisions based on the constraint configuration (such as preset exposure constraints for upper and lower limits of exposure times) and the exposure times of the first candidate objects corresponding to the second screening results. This allows for the application of uniform constraints to all tasks within a set of multiple flow control tasks (e.g., the sum of the goals of all tasks in the plan does not exceed the total goal of the plan), minimizing overall losses such as click-through rate and conversion rate while smoothly completing tasks that maintain volume. For example, assuming the preset exposure constraints include an upper limit constraint, i.e., the total number of exposures cannot exceed 1,000, the server can remove first candidate objects with a larger exposure time (e.g., the flow control target is greater than the preset flow threshold) from the first candidate objects corresponding to the second screening results to balance the exposure times of the first and second candidate objects.

[0073] In practical applications, the traffic control labels of target candidates can also be updated. There are various ways to update traffic control labels; one optional method is provided below. Figure 8 As shown, the specific process may include the following steps S802 to S806.

[0074] In step S802, the number of times the target candidate object is exposed within a preset display period is obtained.

[0075] The preset display period can be the last 3 days, the last week, etc.

[0076] In step S804, based on the number of exposures and the traffic control target corresponding to the target candidate object, it is determined whether the target candidate object has reached the preset exposure target.

[0077] In implementation, in such Figure 4 In the recommendation system shown, the real-time data synchronization module can utilize distributed big data processing (such as Flink) to acquire behavior logs in real time. Based on these logs, it can statistically analyze data such as the number of exposures and page views (PV) of the target candidate object under each control task within a preset display period, and send the aggregated behavior data to a message queue. Upon receiving this message, the traffic control management platform can use real-time and offline statistical data to determine whether the traffic control task for that target candidate object has "crossed the line" (i.e., the number of exposures has reached the traffic control target).

[0078] In step S806, if it is determined that the target candidate object has reached the preset exposure target, the traffic control label of the target candidate object is updated.

[0079] In practice, if the target candidate object is determined to have reached the preset exposure target, the flow control task of the target candidate object can be taken offline and the flow control tag of the target candidate object can be removed, thus updating the target candidate object to a candidate object.

[0080] In practical applications, the specific processing method for filtering the first candidate objects based on the preset display ratio and the number of second candidate objects in step S206 above to obtain the target display object can be varied. The following provides one optional processing method, such as... Figure 9 As shown, the specific process may include the following steps S2062 to S2064.

[0081] In step S2062, a second quantity is determined based on a preset display ratio and the number of second candidate objects.

[0082] In step S2064, the first candidate objects after filtering are sorted based on the matching degree between the first candidate objects and the user, and the target display object is determined based on the second quantity and the sorting result.

[0083] In implementation, in such Figure 4 In the recommendation system shown, the re-ranking and event tracking modules can be used to control the density of first-choice objects based on a preset display ratio (e.g., a maximum of 2 first-choice objects are displayed for every 10 objects) to balance the control objectives and user experience. Simultaneously, the server can inject traffic control tags into the object event tracking for subsequent experimental effect analysis and report cleaning.

[0084] In this way, the recommendation system can connect the three major data links of offline, real-time, and online, forming a closed-loop, data-driven control system. The data processing of the recommendation system can include an offline data processing module, a real-time data processing module, an online recommendation service module, and a traffic control management platform.

[0085] Specifically, the offline data processing module can be used to clean the historical behavioral data of candidate objects, such as exposure and click data, and to perform data aggregation from different dimensions.

[0086] The real-time data processing module can be used to perform real-time statistics on the exposure, clicks, and other data of candidate objects within a preset display period, and to perform data aggregation from different dimensions.

[0087] The online recommendation service module may include new recall channels for cold start products and booster products, quota cutoff for coarse and fine sorting, flow control center and rearrangement concentration control processing and other components.

[0088] As a central hub, the traffic control management platform can manage the lifecycle of control tasks (such as creation, approval, launch, and decommissioning). It can also synchronize traffic control tasks to search engines and various databases, and use offline and real-time data to control the progress of traffic control tasks.

[0089] A unified traffic control system, encompassing offline data processing, real-time data synchronization, online recall, ranking, and online resource allocation, can address the Matthew effect (where popular products in a recommendation system consistently receive more exposure due to their historical data advantage, while mid-to-long-tail products struggle to gain traffic, resulting in uneven traffic distribution) that exists in service platforms such as product rental. This system displays rentable products to users through personalized recommendations and uses a traffic allocation mechanism to support precise traffic control for specific products (such as new products, potential products, and partner products), meeting operational needs and improving the overall ecosystem health and user experience.

[0090] Specifically, through the real-time data production and consumption chain of the data synchronization module, the system can perceive the progress of traffic control tasks within seconds and immediately remove the target after it crosses the line, ensuring high certainty and timeliness of traffic control, avoiding traffic waste, solving the problem of "lack of real-time feedback," and achieving precise release of control resources. Since each traffic control task is based on a single product and managed through an independent recall channel and preset quotas, it can achieve refined and independent control of a large number of single products, solving the problem of "coarse control granularity." While ensuring the smooth completion of control tasks, it can maximize overall efficiency. Furthermore, through multi-channel recall fusion, coarse-rank quota protection, and multi-layered collaboration between flow control algorithms and re-ranking concentration control, it can ensure the smooth achievement of control goals while maximizing the preservation of personalized recommendation capabilities, balancing operational goals and user experience, and solving the problem of "damaging the recommendation experience." Finally, by using a unified traffic control task as a carrier, a complete closed loop of offline generation, online execution, and real-time feedback is connected, forming a stable and scalable unified control framework. This solves the problem of "offline and online separation" and improves the certainty and timeliness of control.

[0091] This specification provides an object recommendation method. It involves acquiring a first candidate object and a second candidate object, where the historical exposure count of the first candidate object may be less than a preset threshold. Based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object, the first candidate object is filtered to obtain a filtered first candidate object. Then, based on a preset display ratio and the number of second candidate objects, the filtered first candidate object is further filtered to obtain a target display object. Based on the target display object and the second candidate object, the object recommendation result is determined, and the object recommendation is then sent to the user based on this result. This approach firstly ensures that low-exposure, mid-to-long-tail products are not completely filtered out in subsequent object selection by using a preset quota. Secondly, by using a preset display ratio and the number of second candidate objects to filter the filtered first candidate object, it ensures that the control target is smoothly achieved while maximizing the preservation of personalized recommendation capabilities. This balances operational goals and user experience, improving the accuracy of object recommendations.

[0092] The above describes the object recommendation method provided in the embodiments of this specification. Based on the same idea, the embodiments of this specification also provide an object recommendation device, such as... Figure 10 As shown.

[0093] The object recommendation device includes: an object acquisition module 1001, a first filtering module 1002, a second filtering module 1003, and a result determination module 1004, wherein: The object acquisition module 1001 acquires a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold. The first filtering module 1002 filters the first candidate object based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object, and obtains the filtered first candidate object. The second filtering module 1003 filters the first candidate object after filtering based on the preset display ratio and the number of the second candidate objects to obtain the target display object. The result determination module 1004 determines the object recommendation result based on the target display object and the second candidate object, and performs object recommendation processing to the user based on the object recommendation result.

[0094] In this embodiment of the specification, the object acquisition module 1001 is used for: Obtain target candidate objects carrying traffic control tags, and traffic control tasks corresponding to the target candidate objects, wherein the traffic control tasks include the traffic control targets; Based on the task release time of the traffic control task, the target candidate objects are filtered to obtain the first candidate object to be recommended.

[0095] In this embodiment of the specification, the object acquisition module 1001 is used for: Based on the task release time of the traffic control task, the target candidate objects are filtered to obtain the first object; A pre-trained relevance determination model is used to determine the matching degree between the first object and the user, and based on the matching degree, the first object is filtered to obtain the first candidate object to be recommended.

[0096] In the embodiments described in this specification, the device further includes: The data acquisition module retrieves the publication time and historical exposure count for each candidate. The object determination module determines candidate objects whose publication time is within a preset time range and / or whose historical exposure count is less than the preset number threshold as the target candidate objects carrying the traffic control tag. The task generation module uses a pre-trained task generation model to determine the traffic control target corresponding to the target candidate object based on the click-through rate requirement corresponding to the target candidate object, and generates a traffic control task corresponding to the target candidate object based on the traffic control target.

[0097] In this embodiment of the specification, the first screening module 1002 is used for: Determine the first quantity corresponding to the preset quota; Based on the matching degree between the first candidate and the user, and the traffic control target corresponding to the click-through rate requirement of the first candidate, the first candidate is filtered to obtain the first filtering result; Based on the first quantity and the first screening result, the first candidate object after screening is determined.

[0098] In this embodiment of the specification, the first screening module 1002 is used for: Based on the preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate, the first candidate is filtered to obtain the second filtering result. Obtain a preset exposure count constraint, and based on the preset exposure count constraint and the exposure count of the first candidate object corresponding to the second screening result, perform screening processing on the first candidate object corresponding to the second screening result to obtain the filtered first candidate object.

[0099] In the embodiments described in this specification, the device further includes: The acquisition module acquires the number of times the target candidate object is exposed within a preset display period; The judgment module determines whether the target candidate has reached the preset exposure target based on the number of exposures and the traffic control target corresponding to the target candidate; The update module updates the traffic control label of the target candidate object when it is determined that the target candidate object has reached the preset exposure target.

[0100] In this embodiment of the specification, the second screening module 1003 is used for: The second quantity is determined based on the preset display ratio and the number of the second candidate objects; Based on the matching degree between the filtered first candidate object and the user, the filtered first candidate object is sorted, and based on the second quantity and sorting result, the target display object is determined.

[0101] This specification provides an object recommendation device. It acquires a first candidate object and a second candidate object, wherein the historical exposure count of the first candidate object may be less than a preset threshold. Based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object, the first candidate object is filtered to obtain a filtered first candidate object. Based on a preset display ratio and the number of the second candidate object, the filtered first candidate object is further filtered to obtain a target display object. Based on the target display object and the second candidate object, the object recommendation result is determined, and the object recommendation is then performed on the user based on the object recommendation result. In this way, firstly, the preset quota ensures that mid-to-long-tail products with low exposure are not completely filtered out in subsequent object filtering processes. Secondly, by filtering the first candidate object using a preset display ratio and the number of the second candidate object, the control target can be smoothly achieved while maximizing the preservation of personalized recommendation capabilities, balancing operational goals and user experience, and improving the accuracy of object recommendations.

[0102] The above describes the object recommendation device provided in the embodiments of this specification. Based on the same idea, the embodiments of this specification also provide an object recommendation device, such as... Figure 11 As shown.

[0103] The object recommendation device can be a terminal device or server, etc., provided in the above embodiments.

[0104] like Figure 11As shown, device 1100 mainly consists of a communication interface 1102, a user interface 1104, a processor 1106, and a data storage 1108. These components are interconnected and communicate with each other via a system bus, network, or other connection mechanism 1110. The communication interface 1102 enables device 1100 to communicate with other devices, access networks, and transmission networks via analog or digital modulation. For example, the communication interface 1102 may include a chipset and antenna for wireless communication with a radio access network or access point. Furthermore, the communication interface 1102 can be a wired interface such as Ethernet, Token Ring, or a USB port, or a wireless interface such as Wi-Fi, Bluetooth, Global Positioning System (GPS), or a wide-area wireless interface (e.g., WiMAX or LTE). Of course, the communication interface 1102 can also support other forms of physical layer interfaces and standard or proprietary communication protocols. The communication interface 1102 may also include multiple physical communication interfaces, such as Wi-Fi, Bluetooth, and wide-area wireless interfaces.

[0105] User interface 1104 includes receiving user input and providing output to the user. Therefore, user interface 1104 may include input components such as a keypad, keyboard, touch-sensitive or presence-sensitive panel, computer mouse, trackball, joystick, microphone, still camera, and video camera, and output components such as a display screen (which may be combined with a touch-sensitive panel), CRT, LCD, LED, display using DLP technology, printer, and other similar devices known or developed in the future. User interface 1104 may also generate auditory output via speakers, speaker jacks, audio output ports, audio output devices, headphones, and other similar devices known or developed in the future. In some embodiments, user interface 1104 may include software, circuitry, or other forms of logic capable of transmitting and receiving data from external user input / output devices. Additionally or alternatively, device 1100 may support remote access from other devices via communication interface 1102 or another physical interface (not shown). User interface 1104 may be configured to receive user input, the position and movement of which may be indicated by indicators or cursors described herein. User interface 1104 can also be configured as a display device for rendering or displaying text fragments.

[0106] Processor 1106 may include one or more general-purpose processors and / or special-purpose processors.

[0107] Data storage 1108 may include one or more volatile storage components and may be integrated wholly or partially with processor 1106. Data storage 1108 may include removable and non-removable components.

[0108] Processor 1106 is capable of executing program instructions 1118 (e.g., compiled or uncompiled program logic and / or machine code) stored in data storage 1108 to perform the various functions described herein. Data storage 1108 may contain a non-transitory computer-readable medium on which program instructions are stored, which, when executed by device 1100, enable device 1100 to perform any methods, processes, or functions disclosed in this specification and / or the accompanying drawings. Execution of program instructions 1118 by processor 1106 may result in processor 1106 using data 1112.

[0109] For example, program instructions 1118 may include an operating system 1122 (e.g., an operating system kernel, device drivers, and / or other modules) installed on device 1100 and one or more applications 1120 (e.g., a browser, social application, or game application). Similarly, data 1112 may include operating system data 1116 and application data 1114. Operating system data 1116 is primarily accessible to the operating system 1122, while application data 1114 is primarily accessible to one or more applications 1120. Application data 1114 may reside in a file system visible or hidden from the user of device 1100.

[0110] Application 1120 can communicate with operating system 1112 through one or more application programming interfaces (APIs). These APIs help application 1120 read and / or write application data 1114, transmit or receive information via communication interface 1102, receive or display information on user interface 1104, etc.

[0111] In some terminology, application 1120 may be simply referred to as "app". Furthermore, application 1120 can be downloaded to device 1100 through one or more online app stores or app markets. However, applications can also be installed on device 1100 in other ways, such as through a web browser or a physical interface on device 1100 (e.g., a USB port).

[0112] Specifically, in this embodiment, the object recommendation device includes a memory and one or more programs, wherein one or more programs are stored in the memory, and one or more programs may include one or more modules, and each module may include a series of computer-executable instructions for the object recommendation device, and is configured to be executed by one or more processors. The one or more programs include computer-executable instructions for performing the following: Obtain a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold. Based on the preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate, the first candidate is filtered to obtain the filtered first candidate. Based on the preset display ratio and the number of the second candidate objects, the filtered first candidate objects are filtered to obtain the target display objects; Based on the target display object and the second candidate object, an object recommendation result is determined, and an object recommendation process is performed to the user based on the object recommendation result.

[0113] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the object recommendation device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0114] This specification provides an object recommendation device. It acquires a first candidate object and a second candidate object, wherein the historical exposure count of the first candidate object may be less than a preset threshold. Based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object, the first candidate object is filtered to obtain a filtered first candidate object. Based on a preset display ratio and the number of the second candidate object, the filtered first candidate object is further filtered to obtain a target display object. Based on the target display object and the second candidate object, the object recommendation result is determined, and the object recommendation is then performed on the user based on the object recommendation result. In this way, firstly, the preset quota ensures that mid-to-long-tail products with low exposure are not completely filtered out in subsequent object filtering processes. Secondly, by filtering the first candidate object using a preset display ratio and the number of the second candidate object, the device can ensure that the control target is smoothly achieved while maximizing the preservation of personalized recommendation capabilities, balancing operational goals and user experience, and improving the accuracy of object recommendations.

[0115] Furthermore, based on the above Figures 1 to 9 This specification also provides a storage medium for storing computer-executable instruction information in one or more embodiments. In one specific embodiment, the storage medium may be a USB flash drive, optical disc, hard disk, etc. When the computer-executable instruction information stored in the storage medium is executed by a processor, it can realize the following process: Obtain a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold. Based on the preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate, the first candidate is filtered to obtain the filtered first candidate. Based on the preset display ratio and the number of the second candidate objects, the filtered first candidate objects are filtered to obtain the target display objects; Based on the target display object and the second candidate object, an object recommendation result is determined, and an object recommendation process is performed to the user based on the object recommendation result.

[0116] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the above-described storage medium embodiment is basically similar to the method embodiment, so the description is relatively simple; relevant parts can be referred to the description of the method embodiment.

[0117] This specification provides a storage medium that acquires a first candidate object and a second candidate object to be recommended. The historical exposure count of the first candidate object may be less than a preset threshold. Based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object, the first candidate object is filtered to obtain a filtered first candidate object. Based on a preset display ratio and the number of the second candidate object, the filtered first candidate object is further filtered to obtain a target display object. Based on the target display object and the second candidate object, the object recommendation result is determined, and the object recommendation is performed to the user based on the object recommendation result. In this way, firstly, the preset quota ensures that mid-to-long-tail products with low exposure are not completely filtered out in subsequent object filtering processes. Secondly, by filtering the first candidate object using a preset display ratio and the number of the second candidate object, the control target can be smoothly achieved while maximizing the preservation of personalized recommendation capabilities, balancing operational goals and user experience, and improving the accuracy of object recommendations.

[0118] Furthermore, based on the above Figures 1 to 9 This specification also provides one or more embodiments of a computer program product, including a computer program, which, when executed by a processor, can perform the following processes: Obtain a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold. Based on the preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate, the first candidate is filtered to obtain the filtered first candidate. Based on the preset display ratio and the number of the second candidate objects, the filtered first candidate objects are filtered to obtain the target display objects; Based on the target display object and the second candidate object, an object recommendation result is determined, and an object recommendation process is performed to the user based on the object recommendation result.

[0119] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the above-described embodiment of a computer program product is relatively simple in description because it is fundamentally similar to the method embodiment; relevant parts can be referred to the description of the method embodiment.

[0120] This specification provides a computer program product that acquires a first candidate object and a second candidate object to be recommended. The historical exposure count of the first candidate object may be less than a preset threshold. Based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object, the first candidate object is filtered to obtain a filtered first candidate object. Based on a preset display ratio and the number of the second candidate object, the filtered first candidate object is further filtered to obtain a target display object. Based on the target display object and the second candidate object, the object recommendation result is determined, and the object recommendation is performed to the user based on the object recommendation result. In this way, firstly, the preset quota ensures that mid-to-long-tail products with low exposure are not completely filtered out in subsequent object filtering processes. Secondly, by filtering the first candidate object using a preset display ratio and the number of the second candidate object, the control target can be smoothly achieved while maximizing the preservation of personalized recommendation capabilities, balancing operational goals and user experience, and improving the accuracy of object recommendations.

[0121] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0122] In the 1990s, improvements to a technology could be clearly distinguished as either hardware improvements (e.g., improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the methodology). However, with technological advancements, many methodological improvements today can be considered direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it cannot be said that a methodological improvement cannot be implemented using a hardware physical module. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. Designers can program and "integrate" a digital system onto a PLD themselves, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software. Similar to the software compiler used in program development, the original code before compilation must also be written in a specific programming language, called a Hardware Description Language (HDL). There are many HDLs, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, and RHDL (Ruby Hardware Description Language). Currently, the most commonly used are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also understand that by simply performing some logic programming on the method flow using one of these hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logical method flow can be easily obtained.

[0123] The controller can be implemented in any suitable manner. For example, it can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Alternatively, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0124] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0125] For ease of description, the above apparatus is described by dividing it into various functional units. Of course, when implementing one or more embodiments of this specification, the functions of each unit can be implemented in one or more software and / or hardware.

[0126] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, one or more embodiments of this specification may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0127] The embodiments described herein are illustrated with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0128] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0129] These computer program instructions may also be loaded onto a computer or other programmable device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable device for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0130] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0131] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0132] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0133] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Without further limitations, the presence of additional identical elements in the process, method, article, or apparatus that includes said elements is not excluded.

[0134] It should also be noted that the terms "one," "an," and "the" do not specifically refer to the singular; they can also include the plural. Ordinal numbers such as "first," "second," etc., do not necessarily indicate order; often they are used to distinguish between objects. For example, "first server" and "second server" usually refer to two servers. To differentiate between these two servers, they are described as "first server" and "second server." Of course, sometimes these two servers may be the same server.

[0135] In this specification, unless explicitly stated otherwise, "receiving and sending data" does not necessarily mean direct receiving and sending; it can also mean indirect receiving and sending. For example, A receiving data sent by B can be understood as A directly receiving data sent by B, or it can be understood as A indirectly receiving data sent by B through other entities such as C. Similarly, B sending data to A can be understood as B sending data directly to A, or it can be understood as B indirectly sending data to A through other entities such as C. Here, C can be a single subject, or two or more subjects.

[0136] This specification uses specific terms to describe embodiments thereof. Terms such as "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "an embodiment," "one embodiment," or "an alternative embodiment" in different locations within this specification do not necessarily refer to the same embodiment. Furthermore, those skilled in the art can combine and integrate the different embodiments or examples described herein, as well as the features of those different embodiments or examples, without contradiction.

[0137] Although one or more embodiments of this specification provide method steps as described in the embodiments or flowcharts, it is understood that the order of steps listed in the embodiments or flowcharts is only one of many possible execution orders and does not represent the only execution order. Therefore, when the claims involve method steps, any changes or adjustments to the order of such steps, or the parallelism between steps, are also within the scope of protection of the claims.

[0138] It should be noted that the user data obtained in this manual is authorized by the user and does not involve user privacy.

[0139] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, one or more embodiments of this specification may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0140] One or more embodiments of this specification can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. One or more embodiments of this specification can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0141] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0142] The above description is merely an embodiment of this specification and is not intended to limit this document. Various modifications and variations can be made to this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of the claims of this specification.

Claims

1. An object recommendation method, comprising: Obtain a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold. Based on the preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate, the first candidate is filtered to obtain the filtered first candidate. Based on the preset display ratio and the number of the second candidate objects, the filtered first candidate objects are filtered to obtain the target display objects; Based on the target display object and the second candidate object, an object recommendation result is determined, and an object recommendation process is performed to the user based on the object recommendation result.

2. The method according to claim 1, wherein obtaining the first candidate object to be recommended includes: Obtain target candidate objects carrying traffic control tags, and traffic control tasks corresponding to the target candidate objects, wherein the traffic control tasks include the traffic control targets; Based on the task release time of the traffic control task, the target candidate objects are filtered to obtain the first candidate object to be recommended.

3. The method according to claim 2, wherein the step of filtering the target candidate objects based on the task release time of the traffic control task to obtain the first candidate object to be recommended includes: Based on the task release time of the traffic control task, the target candidate objects are filtered to obtain the first object; A pre-trained relevance determination model is used to determine the matching degree between the first object and the user, and based on the matching degree, the first object is filtered to obtain the first candidate object to be recommended.

4. The method according to claim 2, further comprising, before obtaining the target candidate object carrying the traffic control tag and the traffic control task corresponding to the target candidate object: Obtain the publication time and historical exposure count for each candidate; Candidates whose release time is within a preset time range and / or whose historical exposure count is less than the preset number of times are identified as the target candidate carrying the traffic control tag. Using a pre-trained task generation model, based on the click-through rate requirement corresponding to the target candidate object, the traffic control target corresponding to the target candidate object is determined, and based on the traffic control target, a traffic control task corresponding to the target candidate object is generated.

5. The method according to claim 1, wherein the step of filtering the first candidate object based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object to obtain the filtered first candidate object includes: Determine the first quantity corresponding to the preset quota; Based on the matching degree between the first candidate and the user, and the traffic control target corresponding to the click-through rate requirement of the first candidate, the first candidate is filtered to obtain the first filtering result; Based on the first quantity and the first screening result, the first candidate object after screening is determined.

6. The method according to claim 1, wherein the step of filtering the first candidate object based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object to obtain the filtered first candidate object includes: Based on the preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate, the first candidate is filtered to obtain the second filtering result. Obtain a preset exposure count constraint, and based on the preset exposure count constraint and the exposure count of the first candidate object corresponding to the second screening result, perform screening processing on the first candidate object corresponding to the second screening result to obtain the filtered first candidate object.

7. The method according to claim 2, further comprising: Obtain the number of times the target candidate object is exposed within a preset display period; Based on the number of exposures and the traffic control target corresponding to the target candidate, determine whether the target candidate has reached the preset exposure target; If it is determined that the target candidate object has reached the preset exposure target, the traffic control label of the target candidate object is updated.

8. The method according to claim 1, wherein the step of filtering the first candidate objects based on a preset display ratio and the number of the second candidate objects to obtain the target display object includes: The second quantity is determined based on the preset display ratio and the number of the second candidate objects; Based on the matching degree between the filtered first candidate object and the user, the filtered first candidate object is sorted, and based on the second quantity and sorting result, the target display object is determined.

9. An object recommendation device, comprising: The object acquisition module acquires a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold. The first filtering module filters the first candidate object based on a preset quota and a traffic control target corresponding to the click-through rate requirement of the first candidate object, and obtains the filtered first candidate object. The second filtering module filters the first candidate objects after filtering based on a preset display ratio and the number of the second candidate objects to obtain the target display objects. The result determination module determines the object recommendation result based on the target display object and the second candidate object, and performs object recommendation processing to the user based on the object recommendation result.

10. An object recommendation device, wherein the object recommendation includes: processor; as well as A memory configured to store computer-executable instructions, which, when executed, cause the processor to: Obtain a first candidate object and a second candidate object to be recommended, wherein the historical exposure count of the first candidate object is less than a preset threshold. Based on the preset quota and the traffic control target corresponding to the click-through rate requirement of the first candidate, the first candidate is filtered to obtain the filtered first candidate. Based on the preset display ratio and the number of the second candidate objects, the filtered first candidate objects are filtered to obtain the target display objects; Based on the target display object and the second candidate object, an object recommendation result is determined, and an object recommendation process is performed to the user based on the object recommendation result.

Images