Article recommendation method, device and equipment, computer readable medium and program product

Abstract

The embodiment of the invention discloses an article recommendation method, device and equipment, a computer readable medium and a program product. A specific embodiment of the method comprises the following steps: acquiring article search features and value behavior information corresponding to a target user; pre-generated article value division standard information corresponding to the article search feature is determined; generating a user value feature information set according to the article value division standard information and the value behavior information; inputting the user value feature information set into a pre-trained value gear preference information generation model to generate value gear preference information corresponding to the target user; an item recommendation is performed for the value level preference information. The implementation mode is related to artificial intelligence, the value gear preference information corresponding to the target user can be accurately and efficiently generated, and accurate recommendation for the target user is achieved.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the field of computer technology, and more specifically to methods, apparatus, devices, computer-readable media, and program products for recommending items. Background Technology

[0002] Currently, online item value processing (e.g., online shopping) has permeated all aspects of people's lives. How to recommend more refined items to users has become one of the main directions of development. For item recommendations, the common approach is as follows: First, by using item value equivalence partitioning, the value ranges corresponding to the item set under predetermined search features are equally divided to generate item value partitioning criteria. Then, based on the item value partitioning criteria, item recommendations at the corresponding value level are implemented for the target user.

[0003] However, the inventors discovered that when using the above method to recommend items, the following technical problems often arise:

[0004] The method of classifying items by their equivalent value cannot effectively reflect users' sensitivity to high-value and low-value items. Poor sensitivity often affects users' decisions on how to handle the value of items, thus failing to achieve accurate recommendations.

[0005] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0007] Some embodiments of this disclosure provide methods, apparatuses, electronic devices, computer-readable media, and program products for recommending items to address the technical problems mentioned in the background section above.

[0008] In a first aspect, some embodiments of this disclosure provide an item recommendation method, including: acquiring item search features and value behavior information corresponding to a target user; determining pre-generated item value classification standard information corresponding to the item search features, wherein the item value classification standard information is generated based on a clustering algorithm; generating a user value feature information set based on the item value classification standard information and the value behavior information; inputting the user value feature information set into a pre-trained value tier preference information generation model to generate value tier preference information corresponding to the target user; and performing item recommendation based on the value tier preference information.

[0009] Optionally, the above-mentioned item value classification standard information is generated through the following steps: obtaining the full set of item value information corresponding to the above-mentioned item search features; preprocessing the item value information of each item in the above-mentioned full set of item value information to generate a processed set of item value information; performing clustering processing on the above-mentioned processed set of item value information to generate a cluster of item value information; and generating the above-mentioned item value classification standard information based on the above-mentioned cluster of item value information.

[0010] Optionally, the aforementioned item value classification standard information includes: multiple value tier information; the aforementioned value behavior information includes: order information and value-adding behavior information; and the aforementioned generation of a user value feature information set based on the aforementioned item value classification standard information and the aforementioned value behavior information includes: generating order feature information based on the aforementioned multiple value tier information and the aforementioned order information; generating value-adding behavior feature information based on the aforementioned multiple value tier information and the aforementioned value-adding behavior information; and generating a user value feature information set based on the aforementioned order feature information and the aforementioned value-adding behavior feature information.

[0011] Optionally, the aforementioned value tier preference information includes: tier preference information corresponding to multiple value tier information; and the aforementioned execution of item recommendation based on the aforementioned value tier preference information includes: obtaining a real-time full set of item information corresponding to the aforementioned item search features; filtering out value tier information whose corresponding tier preference information meets preset information conditions from the obtained value tier information set as target value tier information; determining the item information in the aforementioned real-time full set of item information that is in the aforementioned target value tier information as target item information, thereby obtaining a target item information set; and executing item recommendation for the item set corresponding to the aforementioned target item information set.

[0012] Optionally, the above-mentioned clustering process for the processed item value information set to generate item value information clusters includes: obtaining a predetermined number of clusters; determining item value information groups in the processed item value information set whose value magnitudes satisfy a preset value condition; determining an initial cluster center value information set based on the predetermined number of clusters and the item value information groups; and performing clustering process for the processed item value information set based on the initial cluster center value information set to generate item value information clusters.

[0013] Optionally, generating a user value feature information set based on the order feature information and the value-adding behavior feature information includes: obtaining short-term value feature information for the target user; and generating a user value feature information set based on the order feature information, the value-adding behavior feature information, and the short-term value feature information.

[0014] Optionally, the value tier preference information generation model described above is trained through the following steps: obtaining an order dataset, a value tier label set corresponding to each order data, and a value tier preference label; generating an initial sample set corresponding to the order dataset based on the value tier label set corresponding to each order data; dividing each initial sample in the initial sample set into positive and negative samples based on the value tier preference labels to generate a positive sample set and a negative sample set; and training the initial value tier preference information generation model based on the positive and negative sample sets to generate a value tier preference information generation model.

[0015] Secondly, some embodiments of this disclosure provide an item recommendation apparatus, including: an acquisition unit configured to acquire item search features and value behavior information corresponding to a target user; a determination unit configured to determine pre-generated item value classification standard information corresponding to the item search features, wherein the item value classification standard information is generated based on a clustering algorithm; a generation unit configured to generate a user value feature information set based on the item value classification standard information and the value behavior information; an input unit configured to input the user value feature information set into a pre-trained value tier preference information generation model to generate value tier preference information corresponding to the target user; and an execution unit configured to perform item recommendation based on the value tier preference information.

[0016] Optionally, the above-mentioned item value classification standard information is generated through the following steps: obtaining the full set of item value information corresponding to the above-mentioned item search features; preprocessing the item value information of each item in the above-mentioned full set of item value information to generate a processed set of item value information; performing clustering processing on the above-mentioned processed set of item value information to generate a cluster of item value information; and generating the above-mentioned item value classification standard information based on the above-mentioned cluster of item value information.

[0017] Optionally, the aforementioned item value classification standard information includes: multiple value tier information; the aforementioned value behavior information includes: order information and value addition behavior information; and the generation unit can be configured to: generate order feature information based on the aforementioned multiple value tier information and the aforementioned order information; generate value addition behavior feature information based on the aforementioned multiple value tier information and the aforementioned value addition behavior information; and generate a user value feature information set based on the aforementioned order feature information and the aforementioned value addition behavior feature information.

[0018] Optionally, the aforementioned value tier preference information includes: tier preference information corresponding to multiple value tier information; and the execution unit can be configured to: obtain a real-time full set of item information corresponding to the aforementioned item search features; filter out value tier information whose corresponding tier preference information meets preset information conditions from the obtained value tier information set, as target value tier information; determine the item information in the aforementioned real-time full set of item information that is in the aforementioned target value tier information, as target item information, to obtain a target item information set; and perform item recommendation for the item set corresponding to the aforementioned target item information set.

[0019] Optionally, the value tier preference information generation model described above is trained through the following steps: obtaining an order dataset, a value tier label set corresponding to each order data, and a value tier preference label; generating an initial sample set corresponding to the order dataset based on the value tier label set corresponding to each order data; dividing each initial sample in the initial sample set into positive and negative samples based on the value tier preference labels to generate a positive sample set and a negative sample set; and training the initial value tier preference information generation model based on the positive and negative sample sets to generate a value tier preference information generation model.

[0020] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, such that when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any implementation of the first aspect.

[0021] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method as described in any implementation of the first aspect.

[0022] Fifthly, some embodiments of this disclosure provide a computer program product, including a computer program that, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0023] The above embodiments of this disclosure have the following beneficial effects: The item recommendation method of some embodiments of this disclosure can accurately and efficiently generate value level preference information corresponding to the target user, thereby achieving precise recommendations for the target user. Specifically, the reason for the lack of accuracy in related item recommendations is that the method of classifying item value equivalence cannot effectively reflect the user's sensitivity to high-value and low-value items. Poor sensitivity often affects the user's item value processing decisions, thus failing to achieve precise recommendations. Based on this, the item recommendation method of some embodiments of this disclosure first obtains the item search features and value behavior information corresponding to the target user. Here, by obtaining the item search features, the item range of each item under the item search features to be recommended to the target user is determined. In addition, the item search features are used to subsequently determine the corresponding item value classification standard information. The value behavior information is used to obtain the target user's value behavior habits and characteristics. Then, the pre-generated item value classification standard information corresponding to the above item search features is determined, wherein the above item value classification standard information is generated based on a clustering algorithm. Here, the item value classification criteria generated by the clustering algorithm ensures that the value distribution is consistent with the item supply distribution. It also balances the conflict between the value difference between different item tiers and the proportion of the item pool during price classification. Furthermore, it ensures that the value difference between each value tier gradually increases from low to high, while preventing items from becoming overly concentrated in a few value tiers. Next, based on the aforementioned item value classification criteria and value behavior information, a user value feature information set representing the user's value flow can be accurately generated. This user value feature information set is then input into a pre-trained value tier preference information generation model to accurately and quickly generate the value tier preference information corresponding to the target user. Finally, item recommendations based on the aforementioned value tier preference information are executed to achieve precise recommendations. In summary, the item value classification criteria and value tier preference information generation model generated by the clustering algorithm can subsequently accurately generate value tier preference information for the target user, thereby achieving precise item recommendations. Attached Figure Description

[0024] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0025] Figure 1 This is a schematic diagram illustrating an application scenario of an item recommendation method according to some embodiments of the present disclosure;

[0026] Figure 2 This is a flowchart of some embodiments of the item recommendation method according to this disclosure;

[0027] Figure 3 This is a schematic diagram illustrating the clustering effect of item value information clusters in some embodiments of the item recommendation method according to this disclosure;

[0028] Figure 4 These are flowcharts of other embodiments of the item recommendation method according to this disclosure;

[0029] Figure 5 These are schematic diagrams illustrating the structure of some embodiments of the item recommendation device according to this disclosure;

[0030] Figure 6 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0031] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0032] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0033] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0034] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0035] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0036] Before performing any of the operations involving the collection, storage, or use of user personal information (such as value behavior information) disclosed in this disclosure, the relevant organizations or individuals shall fulfill their obligations, including conducting personal information security impact assessments, informing personal information subjects, and obtaining prior authorization and consent from personal information subjects.

[0037] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0038] Figure 1 This is a schematic diagram illustrating an application scenario of an item recommendation method according to some embodiments of the present disclosure.

[0039] exist Figure 1 In this application scenario, firstly, the electronic device 101 can acquire the item search features 103 and value behavior information 104 corresponding to the target user 102. In this application scenario, the item search feature 103 can be "white top". The value behavior information 104 can be "order placement behavior information". Then, the electronic device 101 can determine the pre-generated item value classification standard information 105 corresponding to the above item search feature 103. The above item value classification standard information 105 is generated based on a clustering algorithm. In this application scenario, the item value classification standard information 105 can be "first tier: 30-80, second tier: 80-150, third tier: 150-300, fourth tier: 300-500, fifth tier: 500-1000, sixth tier: 1000-5000, seventh tier: above 5000". Next, the electronic device 101 can generate a user value feature information set 106 based on the above item value classification standard information 105 and the above value behavior information 104. Furthermore, the electronic device 101 can input the aforementioned user value feature information set 106 into a pre-trained value tier preference information generation model 107 to generate value tier preference information 108 corresponding to the target user 102. In this application scenario, the value tier preference information 108 can be "Tier 1: 0.1, Tier 2: 0.2, Tier 3: 0.4, Tier 4: 0.3, Tier 5: 0, Tier 6: 0, Tier 7: 0". Finally, the electronic device 101 can perform item recommendations based on the aforementioned value tier preference information 108.

[0040] It should be noted that the aforementioned electronic device 101 can be either hardware or software. When the electronic device is hardware, it can be implemented as a distributed cluster consisting of multiple servers or terminal devices, or as a single server or a single terminal device. When the electronic device is software, it can be installed in the hardware devices listed above. It can be implemented as, for example, multiple software programs or software modules used to provide distributed services, or as a single software program or software module. No specific limitations are made here.

[0041] It should be understood that Figure 1 The number of electronic devices shown is merely illustrative. Any number of electronic devices can be used depending on the implementation requirements.

[0042] Continue to refer to Figure 2 The diagram illustrates a flow 200 of some embodiments of an item recommendation method according to the present disclosure. This item recommendation method includes the following steps:

[0043] Step 201: Obtain the item search characteristics and value behavior information corresponding to the target user.

[0044] In some embodiments, the entity executing the above-described item recommendation method (e.g.) Figure 1 The electronic device 101 shown can acquire the item search features and value behavior information of the target user through a wired or wireless connection. The target user can be a user to whom item recommendations are to be made. In an e-commerce scenario, the target user can be a user to whom product recommendations are to be made. Item search features can be the characteristics of the items being searched. In practice, item search features can be features input by the target user for item search. For example, an item search feature could be "white top." That is, for an item feature like "white top," the e-commerce platform can search for a set of items that support value transfer. Another example is that item search features can be item category or item brand. In a specific scenario, the target user will input item search features in the corresponding e-commerce platform field to search for a set of items under the corresponding item features, thereby realizing value transfer processing. For example, value transfer processing can be item purchase processing. Value behavior information can be the behavioral records of the target user performing value actions. In practice, value behavior information can be behavioral records of the target user within a historical time period. In practice, for e-commerce scenarios, valuable behavioral information may include, but is not limited to, at least one of the following: purchase behavior information, adding to cart behavior information, and return behavior information.

[0045] Step 202: Determine the pre-generated item value classification standard information corresponding to the above item search features.

[0046] In some embodiments, the aforementioned executing entity can determine pre-generated item value classification standard information corresponding to the aforementioned item search features. Each item search feature has corresponding pre-generated item value classification standard information. The item value classification standard information can characterize the standard classification of item value. That is, the item value classification standard information can characterize how the item value is classified. In practice, for e-commerce scenarios, the item value classification standard information can be item price ranges. That is, the item value classification standard information characterizes how the item price is classified. For example, the item value classification standard information could be "First tier: 30-80, Second tier: 80-150, Third tier: 150-300, Fourth tier: 300-500, Fifth tier: 500-1000, Sixth tier: 1000-5000, Seventh tier: 5000 and above". The aforementioned item value classification standard information is generated based on a clustering algorithm. In practice, clustering algorithms can be, but are not limited to, at least one of the following: K-Means algorithm (k-means clustering algorithm) and DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm.

[0047] As an example, the above-mentioned information on the criteria for classifying the value of items is generated through the following steps:

[0048] The first step is to obtain the set of items corresponding to the item search features. This set of items can be any item on the e-commerce platform that supports the value chain and possesses the item search features.

[0049] The second step is to determine the value and inventory level of each item in the item set, resulting in a set of item values ​​and a set of inventory levels. In practice, the item value can be the item price. The inventory level can be the amount of item stockpiled in the warehouse.

[0050] The third step is to perform distance processing on the above-mentioned item value set and the above-mentioned backlog value set to obtain a cluster of candidate item value information.

[0051] The fourth step is to determine the value range corresponding to each candidate item value information cluster in the above candidate item value information cluster set, so as to generate a value range set as the standard information for classifying item value.

[0052] In some optional implementations of certain embodiments, the above-mentioned item value classification criteria information is generated through the following steps:

[0053] The first step is to obtain the full set of item value information corresponding to the aforementioned item search features. This full set of item value information is the set of item value information corresponding to the entire target item set. Target items can be any item that possesses the corresponding item search features. Specifically, the full set of target items can be all items on an e-commerce platform that support value transfer and possess the corresponding item search features. For example, if the item search feature is "white top," the full set of item value information could be the price information set of all white clothing items on the target e-commerce platform.

[0054] The second step involves preprocessing the value information of each item in the aforementioned full set of item value information to generate a processed set of item value information. This preprocessing may include, but is not limited to, at least one of the following: outlier removal. Outlier removal may involve removing 5% of the highest and lowest item value information.

[0055] The third step involves performing clustering processing on the processed item value information set to generate item value information clusters. Within each item value information cluster, the values ​​of the items are similar.

[0056] As an example, using the value difference between the processed item value information as the standard for measuring distance, the K-Means algorithm is used to preprocess the value information of each item in the above full set of item value information to generate a processed set of item value information.

[0057] As an example, see Figure 3 This diagram illustrates the clustering effect of clusters of item value information.

[0058] The fourth step is to generate the value classification standard information for the above-mentioned items based on the above-mentioned item value information cluster.

[0059] As an example, firstly, the cluster center corresponding to each item value information cluster in the item value information cluster set is determined, resulting in a cluster center set. Then, the cluster centers in the cluster center set are sorted in ascending order to generate a cluster center sequence. Next, the average value between every two adjacent cluster centers in the cluster center sequence is determined, resulting in an average value sequence. Finally, every two adjacent average values ​​in the average value sequence are determined as value tiers, resulting in a value tier sequence, which serves as the standard information for classifying the aforementioned item values.

[0060] As another example, the value range corresponding to each item value information cluster in the item value information cluster set is determined to obtain the value range set, which serves as the standard information for classifying the value of the items mentioned above.

[0061] In some optional implementations of certain embodiments, the above-described clustering process for the processed item value information set to generate item value information clusters may include the following steps:

[0062] The first step is to obtain the predetermined number of clusters. The number of clusters represents the number of item value information clusters included in the item value information cluster set. For example, the number of clusters could be 9.

[0063] The second step is to identify the item value information groups in the processed item value information set that satisfy preset value conditions. These preset value conditions can be pre-defined value-related criteria. For example, preset value conditions could be filtering out the item value information with the highest and lowest values. That is, the item value information group could include: the item value information with the highest value and the item value information with the lowest value. As another example, preset value conditions could also be filtering out multiple item information items with the largest backlog.

[0064] The third step is to determine the initial cluster center value information set based on the number of clusters and the item value information set mentioned above. The initial cluster center value information can be the initial value corresponding to the cluster center.

[0065] As an example, first, determine the item value information interval corresponding to the item value information group. Then, based on the number of clusters, divide the item value information interval proportionally to determine the division values ​​between each interval as the initial cluster center value information set. For example, the item value information group includes: the item value information with the largest value and the item value information with the smallest value. The item value information with the largest value is 100. The item value information with the smallest value is 10. The number of clusters is 9. Divide the interval [10,100] proportionally to obtain "[10,20], (20,30], (30,40], (40,50], (50,60], (60,70], (70,80], (90,100]". Then, the initial cluster center value information set can be "10, 20, 30, 40, 50, 60, 70, 80, 90".

[0066] As another example, firstly, the item value information group corresponds to the item value information range. The item value information group may include: the item value information with the highest value and the item value information with the lowest value. Then, based on this, the item value information with the highest value and the item value information with the lowest value are used as the initial cluster center value information. Then, a predetermined number of initial cluster center value information are randomly selected from the item value information range to generate the initial cluster center information set.

[0067] The fourth step involves performing clustering processing on the processed item value information set based on the initial cluster center value information set to generate item value information clusters.

[0068] As an example, the aforementioned execution entity can use the initial cluster center value information set as the cluster center value of each cluster center, and perform clustering processing on the processed item value information set to generate item value information clusters.

[0069] Step 203: Generate a user value feature information set based on the above-mentioned item value classification standard information and the above-mentioned value behavior information.

[0070] In some embodiments, the aforementioned executing entity may generate a user value characteristic information set based on the aforementioned item value classification standard information and the aforementioned value behavior information. This user value characteristic information set may be a set of information related to the value characteristics set and targeting specific users. In practice, the value characteristic set may include: product positioning characteristics of the items purchased by the user, category characteristics of the items purchased by the user, price characteristics of the items purchased by the user, and the value tier of the items purchased by the user.

[0071] As an example, the aforementioned executing entity can determine multiple value tiers corresponding to the standard information for classifying the value of an item. Then, it extracts a set of feature information related to these multiple value tiers from the aforementioned value behavior information, as the first feature information set. Next, it extracts a set of feature information corresponding to the item's general feature set from the value behavior information, as the second feature information set. Finally, it merges the first and second feature information sets to generate a user value feature information set.

[0072] In some optional implementations of certain embodiments, the aforementioned item value classification standard information includes: multiple value tier information. The aforementioned value behavior information includes: order information and value-adding behavior information. The multiple value tier information can be multiple value ranges involved in the item value classification standard information. Order information can be the order information of each historically issued order for the target user. Value-adding behavior information can be behavior information related to value-adding behavior. In practice, for e-commerce scenarios, value-adding behavior can be adding items to the cart.

[0073] Optionally, generating a user value feature information set based on the aforementioned item value classification criteria and value behavior information may include the following steps:

[0074] The first step involves the execution entity generating order feature information based on the multiple value tiers and order information. This order feature information can be related to various order characteristics. In practice, each order characteristic includes features related to multiple value tiers. These order characteristics can be categorized into two types: user-based subsets and item-search-based subsets. For example, the user-based subset includes the total number of orders corresponding to each value tier, the percentage of orders corresponding to each value tier, the maximum value tier for each order, and the minimum value tier for each order. The item-search-based subset includes the number of orders under the item-search-feature and the percentage of orders under the item-search-feature. As an example, the execution entity can generate a first subset of order characteristics for multiple value tiers and a second subset of order characteristics for item-search-features. Then, it extracts a first subset of order feature information corresponding to the first subset of order characteristics from the order information, and extracts a second subset of order feature information corresponding to the second subset of order characteristics from the order information. Finally, the subset of the first order feature information and the subset of the second order feature information are determined as the order feature information.

[0075] The second step involves the aforementioned executing entity generating value-adding behavior feature information based on the multiple value tiers and the value-adding behavior information. This value-adding behavior feature information can be feature information related to various features corresponding to the value-adding behavior. In practice, these features include features related to multiple value tiers. These features can be divided into two categories: user-based feature subsets and item search feature subsets. For example, the user-based feature subset includes: the total number of items added to cart for each value tier, the percentage of items added to cart for each value tier, the maximum value tier for items added to cart, and the minimum value tier for items added to cart. The item search feature subset includes: the number of items added to cart under the item search feature and the percentage of items added to cart under the item search feature.

[0076] Third, the aforementioned executing entity can generate a user value feature information set based on the aforementioned order feature information and the aforementioned value-added behavioral feature information.

[0077] As an example, the aforementioned executing entity can combine the aforementioned order feature information and the aforementioned value-adding behavior feature information to generate a user value feature information set.

[0078] Optionally, generating a user value feature information set based on the aforementioned order feature information and the aforementioned value-added behavioral feature information may include the following steps:

[0079] The first step involves the aforementioned implementing entity acquiring short-term value characteristic information for the target users. This short-term value characteristic information can be feature information corresponding to a short-term value characteristic set. Short-term value characteristics can be value characteristics within a short time period. In practice, the short-term value characteristic set may include, but is not limited to, at least one of the following: user click characteristics, user characteristics, user short-term order characteristics, and user short-term add-to-cart characteristics.

[0080] The second step involves generating a user value feature information set based on the aforementioned order feature information, value-adding behavior feature information, and short-term value feature information. The fused feature information of the aforementioned order feature information and value-adding behavior feature information can be long-term value feature information specific to the target user. Long-term value feature information can be value feature information over a long period of time.

[0081] As an example, the aforementioned executing entity can combine the aforementioned order characteristic information, the aforementioned value-adding behavior characteristic information, and the aforementioned short-term value characteristic information to generate a user value characteristic information set.

[0082] Step 204: Input the above user value feature information set into the pre-trained value tier preference information generation model to generate the value tier preference information corresponding to the above target user.

[0083] In some embodiments, the executing entity can input the user value feature information set into a pre-trained value tier preference information generation model to generate value tier preference information corresponding to the target user. The value tier preference information generation model can be a neural network model that generates value tier preference information. In practice, value tier preference information can characterize the target user's preference for each price tier over a future time period. The preference level can be a value between 0 and 1. The larger the value, the stronger the target user's preference for the corresponding value tier. In practice, the value tier preference information generation model can include a deep neural network (DNN). The value tier preference information generation model can also include a residual network (ResNet). For example, the value tier preference information could be "Tier 1: 0.1, Tier 2: 0.2, Tier 3: 0.4, Tier 4: 0.3, Tier 5: 0, Tier 6: 0, Tier 7: 0". In practice, the loss function corresponding to the value tier preference information generation model can be a cross-entropy loss function.

[0084] In some optional implementations of certain embodiments, the above-described value tier preference information generation model is trained through the following steps:

[0085] The first step is to obtain the order dataset, the value tier label set for each order, and the value tier preference label. The order data can be historical order data. The value tier label can be the actual value tier corresponding to the items in the order. For example, the order data might include the first order item and the second order item. The first order item corresponds to the first value tier, and the second order item corresponds to the second value tier. Therefore, the value tier label set for the order data would include both the first and second value tiers. The value tier preference label can be the user's preferred value tier information for the order data.

[0086] The second step is to generate an initial sample set corresponding to the above order dataset based on the value tier label set corresponding to each order data.

[0087] As an example, the aforementioned executing entity can generate a set of user value feature information corresponding to the order data, which serves as the target user value feature information set. Then, the target user value feature information set and each value tier label in the value tier label set are combined to generate an initial sample, which serves as the initial sample set.

[0088] The third step is to divide each initial sample in the initial sample set into positive and negative samples based on the aforementioned value tier preference labels, so as to generate a positive sample set and a negative sample set.

[0089] As an example, for each initial sample in the initial sample set, first, it is determined whether the value tier label in the initial sample matches the value tier preference label. Then, in response to the determination of consistency, the initial sample is identified as a positive sample. In response to the determination of inconsistency, the initial sample is identified as a positive sample.

[0090] The fourth step is to train the initial value tier preference information generation model based on the positive and negative sample sets mentioned above, so as to generate a value tier preference information generation model.

[0091] As an example, the aforementioned execution entity can use the positive and negative sample sets as training sample sets, and use the gradient descent method to train the initial value tier preference information generation model to generate the value tier preference information generation model.

[0092] Step 205: Perform item recommendations based on the aforementioned value tier preference information.

[0093] In some embodiments, the aforementioned executing entity may perform item recommendations based on the aforementioned value tier preference information.

[0094] As an example, the aforementioned execution entity can filter out at least one value tier from the value tier preference information, where the corresponding preference level is greater than a preset value. Then, it recommends the item set corresponding to the item search features of at least one value tier to the target user.

[0095] The above embodiments of this disclosure have the following beneficial effects: The item recommendation method of some embodiments of this disclosure can accurately and efficiently generate value level preference information corresponding to the target user, thereby achieving precise recommendations for the target user. Specifically, the reason for the lack of accuracy in related item recommendations is that the method of classifying item value equivalence cannot effectively reflect the user's sensitivity to high-value and low-value items. Poor sensitivity often affects the user's item value processing decisions, thus failing to achieve precise recommendations. Based on this, the item recommendation method of some embodiments of this disclosure first obtains the item search features and value behavior information corresponding to the target user. Here, by obtaining the item search features, the item range of each item under the item search features to be recommended to the target user is determined. In addition, the item search features are used to subsequently determine the corresponding item value classification standard information. The value behavior information is used to obtain the target user's value behavior habits and characteristics. Then, the pre-generated item value classification standard information corresponding to the above item search features is determined, wherein the above item value classification standard information is generated based on a clustering algorithm. Here, the item value classification criteria generated by the clustering algorithm ensures that the value distribution is consistent with the item supply distribution. It also balances the conflict between the value difference between different item tiers and the proportion of the item pool during price classification. Furthermore, it ensures that the value difference between each value tier gradually increases from low to high, while preventing items from becoming overly concentrated in a few value tiers. Next, based on the aforementioned item value classification criteria and value behavior information, a user value feature information set representing the user's value flow can be accurately generated. This user value feature information set is then input into a pre-trained value tier preference information generation model to accurately and quickly generate the value tier preference information corresponding to the target user. Finally, item recommendations based on the aforementioned value tier preference information are executed to achieve precise recommendations. In summary, the item value classification criteria and value tier preference information generation model generated by the clustering algorithm can subsequently accurately generate value tier preference information for the target user, thereby achieving precise item recommendations.

[0096] Further reference Figure 4 The diagram illustrates a flow 400 of another embodiment of the item recommendation method according to this disclosure. This item recommendation method includes the following steps:

[0097] Step 401: Obtain the item search characteristics and value behavior information corresponding to the target user.

[0098] Step 402: Determine the pre-generated item value classification standard information corresponding to the above item search features.

[0099] Step 403: Generate a user value feature information set based on the above-mentioned item value classification standard information and the above-mentioned value behavior information.

[0100] Step 404: Input the above user value feature information set into the pre-trained value tier preference information generation model to generate the value tier preference information corresponding to the above target user.

[0101] In some embodiments, the specific implementation of steps 401-404 and the resulting technical effects can be found in [reference needed]. Figure 2 Steps 201-204 in the corresponding embodiments will not be repeated here.

[0102] Step 405: Obtain the real-time full set of item information corresponding to the above item search features.

[0103] In some embodiments, the executing entity can acquire the real-time full set of item information corresponding to the item search features via wired or wireless means. The real-time full set of item value information can be the full set of item information under the item search features acquired at the current moment.

[0104] Step 406: Select value tier information that meets the preset information conditions for corresponding tier preference information from the obtained value tier information set, and use it as the target value tier information.

[0105] In some embodiments, the executing entity can filter out value tier information whose corresponding tier preference information meets preset information conditions from the obtained value tier information set, and use this as target value tier information. The preset information conditions may be that the value tier information is the tier information with the highest degree of preference for the corresponding tier.

[0106] Step 407: Determine the item information that is in the target value range from the above real-time full item information set, and use it as the target item information to obtain the target item information set.

[0107] In some embodiments, the execution entity may determine the item information in the real-time full item information set that is at the target value level as the target item information, and obtain the target item information set.

[0108] As an example, the aforementioned executing entity can determine the item information that falls within the aforementioned target value range in the aforementioned real-time full set of item information by matching item value, and use it as the target item information to obtain the target item information set.

[0109] Step 408: Perform item recommendation for the item set corresponding to the target item information set.

[0110] In some embodiments, the aforementioned execution entity may perform item recommendations for the item set corresponding to the aforementioned target item information set.

[0111] from Figure 4 It can be seen from this that, with Figure 2 Compared to the description of some corresponding embodiments, Figure 4 In some corresponding embodiments, the item recommendation method process 400 can filter out value tier information that meets preset information conditions, thereby filtering out value tiers that the target user is interested in, and accurately recommending the set of items corresponding to the value tier information to the target user.

[0112] Further reference Figure 5 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of an item recommendation device, which are similar to... Figure 2 Corresponding to the method embodiments shown, the recommended device can be specifically applied to various electronic devices.

[0113] like Figure 5 As shown, an item recommendation device 500 includes: an acquisition unit 501, a determination unit 502, a generation unit 503, an input unit 504, and an execution unit 505. The acquisition unit 501 is configured to acquire item search features and value behavior information corresponding to a target user; the determination unit 502 is configured to determine pre-generated item value classification standard information corresponding to the aforementioned item search features, wherein the aforementioned item value classification standard information is generated based on a clustering algorithm; the generation unit 503 is configured to generate a user value feature information set based on the aforementioned item value classification standard information and the aforementioned value behavior information; the input unit 504 is configured to input the aforementioned user value feature information set into a pre-trained value tier preference information generation model to generate value tier preference information corresponding to the target user; and the execution unit 505 is configured to perform item recommendations based on the aforementioned value tier preference information.

[0114] In some optional implementations of certain embodiments, the above-mentioned item value classification standard information is generated through the following steps: obtaining a full set of item value information corresponding to the above-mentioned item search features; preprocessing the item value information of each item in the above-mentioned full set of item value information to generate a processed set of item value information; performing clustering processing on the above-mentioned processed set of item value information to generate a cluster of item value information; and generating the above-mentioned item value classification standard information based on the above-mentioned cluster of item value information.

[0115] In some optional implementations of certain embodiments, the aforementioned item value classification standard information includes: multiple value tier information; the aforementioned value behavior information includes: order information and value addition behavior information; and the generation unit 503 can be further configured to: generate order feature information based on the aforementioned multiple value tier information and the aforementioned order information; generate value addition behavior feature information based on the aforementioned multiple value tier information and the aforementioned value addition behavior information; and generate a user value feature information set based on the aforementioned order feature information and the aforementioned value addition behavior feature information.

[0116] In some optional implementations of some embodiments, the aforementioned value tier preference information includes: tier preference information corresponding to multiple value tier information; and the execution unit 505 can be further configured to: obtain a real-time full set of item information corresponding to the aforementioned item search features; filter out value tier information whose corresponding tier preference information meets preset information conditions from the obtained value tier information set, as target value tier information; determine the item information in the aforementioned real-time full set of item information that is in the aforementioned target value tier information, as target item information, to obtain a target item information set; and perform item recommendation for the item set corresponding to the aforementioned target item information set.

[0117] In some optional implementations of certain embodiments, the value tier preference information generation model described above is trained through the following steps: obtaining an order dataset, a value tier label set corresponding to each order data, and a value tier preference label; generating an initial sample set corresponding to the order dataset based on the value tier label set corresponding to each order data; dividing each initial sample in the initial sample set into positive and negative samples based on the value tier preference labels to generate a positive sample set and a negative sample set; and training the initial value tier preference information generation model based on the positive and negative sample sets to generate a value tier preference information generation model.

[0118] It is understandable that the units described in the recommended device 500 are related to the reference. Figure 2 The steps in the described method correspond to each other. Therefore, the operations, features, and beneficial effects described above for the method also apply to the item recommendation device 500 and the units contained therein, and will not be repeated here.

[0119] The following is for reference. Figure 6 It illustrates electronic devices suitable for implementing some embodiments of this disclosure (e.g., Figure 1 A schematic diagram of the structure of electronic device 101)600 in the middle. Figure 6 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0120] like Figure 6 As shown, the electronic device 600 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory 602 or a program loaded from a storage device 608 into a random access memory 603. The random access memory 603 also stores various programs and data required for the operation of the electronic device 600. The processing unit 601, the read-only memory 602, and the random access memory 603 are interconnected via a bus 604. An input / output interface 605 is also connected to the bus 604.

[0121] Typically, the following devices can be connected to the input / output interface 605: input devices 606 including, for example, a touchscreen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 607 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 608 including, for example, magnetic tape, hard disk, etc.; and communication devices 609. Communication device 609 allows electronic device 600 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 An electronic device 600 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 6 Each box shown can represent a device or multiple devices as needed.

[0122] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 609, or installed from a storage device 608, or installed from a read-only memory 602. When the computer program is executed by the processing device 601, it performs the functions defined above in the methods of some embodiments of this disclosure.

[0123] It should be noted that, in some embodiments of this disclosure, the computer-readable medium described above may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0124] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0125] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire item search features and value behavior information corresponding to the target user; determine pre-generated item value classification standard information corresponding to the aforementioned item search features, wherein the aforementioned item value classification standard information is generated based on a clustering algorithm; generate a user value feature information set based on the aforementioned item value classification standard information and the aforementioned value behavior information; and input the aforementioned user value feature information set into a pre-trained value tier preference information generation model to generate the value tier preference information corresponding to the target user.

[0126] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0127] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0128] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including an acquisition unit, a determination unit, a generation unit, an input unit, and an execution unit. The names of these units do not necessarily limit the specific unit; for example, an acquisition unit may also be described as "a unit that acquires item search features and value behavior information corresponding to a target user."

[0129] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0130] Some embodiments of this disclosure also provide a computer program product, including a computer program that, when executed by a processor, implements any of the above-described item recommendation methods.

[0131] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A method for recommending items, including: Obtain item search characteristics and value behavior information corresponding to target users; Determine the pre-generated item value classification standard information corresponding to the item search features, wherein the item value classification standard information is generated based on a clustering algorithm; Based on the item value classification standard information and the value behavior information, a user value feature information set is generated; The user value feature information set is input into a pre-trained value tier preference information generation model to generate value tier preference information corresponding to the target user. Perform item recommendations based on the value tier preference information.

2. The method of claim 1, wherein, The information on the criteria for classifying the value of the goods is generated through the following steps: Obtain the full set of item value information corresponding to the item search features; The value information of each item in the full set of item value information is preprocessed to generate a processed set of item value information. Perform clustering processing on the processed set of item value information to generate a cluster set of item value information; Based on the cluster of item value information, the standard information for classifying item value is generated.

3. The method according to claim 1, wherein, The item value classification criteria information includes: multiple value tiers; the value behavior information includes: order information and value addition behavior information; and The step of generating a user value feature information set based on the item value classification standard information and the value behavior information includes: Based on the multiple value tiers and the order information, order feature information is generated; Based on the multiple value tier information and the value addition behavior information, value addition behavior feature information is generated; A user value feature information set is generated based on the order feature information and the value-adding behavior feature information.

4. The method according to claim 1, wherein, The value tier preference information includes: tier preference information corresponding to multiple value tiers; and The process of recommending items based on the value tier preference information includes: Obtain the real-time full set of item information corresponding to the item search features; The value tier information that meets the preset information conditions and is selected from the obtained value tier information set is used as the target value tier information. The item information that falls within the target value range in the real-time full item information set is determined as the target item information, thus obtaining the target item information set; Perform item recommendations for the item set corresponding to the target item information set.

5. The method according to claim 2, wherein, The step of performing clustering processing on the processed set of item value information to generate a cluster set of item value information includes: Obtain a predetermined number of clusters; Determine the value information groups of items in the processed item value information set that satisfy the preset value conditions; Based on the number of clusters and the item value information group, determine the initial cluster center value information set; Based on the initial cluster center value information set, clustering processing is performed on the processed item value information set to generate item value information clusters.

6. The method according to claim 3, wherein, The step of generating a user value feature information set based on the order feature information and the value-added behavioral feature information includes: Obtain short-term value characteristic information for the target user; A user value feature information set is generated based on the order feature information, the value-adding behavior feature information, and the short-term value feature information.

7. The method according to claim 1, wherein, The value tier preference information generation model is trained through the following steps: Obtain the order dataset, the value tier label set for each order, and the value tier preference label; Based on the value tier label set corresponding to each order data, an initial sample set corresponding to the order dataset is generated; Based on the value tier preference label, each initial sample in the initial sample set is divided into positive and negative samples to generate a positive sample set and a negative sample set; Based on the positive and negative sample sets, the initial value tier preference information generation model is trained to generate a value tier preference information generation model.

8. An item recommendation device, comprising: The acquisition unit is configured to acquire the item search characteristics and value behavior information of the target user. The determining unit is configured to determine the pre-generated item value classification standard information corresponding to the item search features, wherein the item value classification standard information is generated based on a clustering algorithm; The generation unit is configured to generate a set of user value feature information based on the item value classification standard information and the value behavior information. The input unit is configured to input the user value feature information set into a pre-trained value tier preference information generation model to generate value tier preference information corresponding to the target user. The execution unit is configured to perform item recommendations based on the value tier preference information.

9. An electronic device, comprising: One or more processors; Storage device, on which one or more programs are stored, When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-7.

10. A computer-readable medium having a computer program stored thereon, wherein, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-7.

11. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-7.

Images