Product feature recall method and apparatus, device, medium

Abstract

This application relates to a product feature recall method, apparatus, device, and medium. The method includes: acquiring coarse product recall information pushed by a product search service, acquiring corresponding merchant identifiers and multiple product identifiers; detecting whether a merchant-specific product feature cluster exists for the merchant identifier; if it exists, acquiring product feature information for each product identifier from the merchant-specific product feature cluster; if it does not exist, detecting the range of product identifiers matched by each product identifier in a product feature node pool, and determining the target product feature nodes for each product identifier range; acquiring corresponding product feature information if each target product feature node exists; if no corresponding product feature information exists, acquiring the corresponding product feature information from a low-to-medium popularity partitioned storage cluster; sorting products according to the product feature information of each product identifier, generating a product sorting list, and pushing it to the user terminal. This application uses layered storage of product features to improve the product feature recall rate.

Description

Technical Field

[0001] This application relates to the field of e-commerce search technology, and in particular to a product feature recall method and corresponding apparatus, computer equipment, and computer-readable storage medium. Background Technology

[0002] Online e-commerce platforms typically offer product search services. Buyers can use these services to search for products on the platform or in individual stores, while merchants can use them to analyze sales data. Existing product search services, after retrieving products from the database based on user-selected search terms, generally require further sorting based on product characteristics before pushing the sorted list to the user. In other words, besides matching search terms, the service also needs to retrieve product characteristics for sorting. Current platforms often use the same resources for this task, but the access volume of different products varies significantly, and the percentage of actively accessed, popular products is generally small. Using the same resources for retrieving product characteristics from products with varying levels of popularity leads to unnecessary resource waste.

[0003] Given the shortcomings of traditional technologies, the applicant has long been engaged in research in related fields and has therefore taken a different approach to solve the industry problems in the field of e-commerce search technology. Summary of the Invention

[0004] The primary objective of this application is to address at least one of the aforementioned problems by providing a product feature recall method and corresponding apparatus, computer equipment, and computer-readable storage medium.

[0005] To achieve the various objectives of this application, the following technical solution is adopted:

[0006] A product feature recall method provided for one of the purposes of this application includes the following steps:

[0007] Obtain coarse recall information of search products pushed by the product search service, and obtain the merchant identifier and multiple product identifiers corresponding to the coarse recall information of search products;

[0008] Detect whether a merchant-specific product feature cluster corresponding to the merchant identifier exists. If it exists, obtain the product feature information corresponding to each product identifier from the product feature nodes in the merchant-specific product feature cluster.

[0009] When the merchant-specific product feature cluster does not exist, the range of product identifiers that each product identifier hits in the product feature node pool is detected, and the target product feature node corresponding to each product identifier range is determined.

[0010] Detect whether there is product feature information corresponding to the product identifier in each of the target product feature nodes. If there is, obtain the product feature information corresponding to the product identifier from the target product feature node. The product feature node stores the product feature information corresponding to the high-popularity product.

[0011] If the target product feature node does not contain the product feature information corresponding to the product identifier, then the product feature information corresponding to the product identifier is obtained from the medium-hot, medium-performance library or the low-hot, low-performance library in the medium-low-hot partition storage cluster.

[0012] The products are sorted according to the product feature information corresponding to each product identifier, a corresponding product sorting list is generated, and the product sorting list is pushed to the corresponding user terminal.

[0013] In a further embodiment, responding to the coarse product recall information pushed by the product search service, and obtaining the merchant identifier and multiple product identifiers corresponding to the coarse product recall information, includes the following steps:

[0014] Respond to product search requests pushed by the user and obtain the product search terms and merchant identifiers corresponding to the product search requests;

[0015] Obtain the product database corresponding to the merchant identifier, and query one or more product objects corresponding to the product search term from the product database;

[0016] Obtain the product identifier corresponding to each of the product objects, and generate coarse search product recall information with each of the product identifiers and each of the merchant identifiers.

[0017] In a further embodiment, before detecting whether a merchant-specific product feature cluster corresponding to the merchant identifier exists, and if so, before obtaining the product feature information corresponding to each product identifier from the product feature nodes in the merchant-specific product feature cluster, the following steps are included:

[0018] Obtain store popularity information from multiple merchant users on the platform. The store popularity information includes store product sales volume, store order volume, store visits, or store exposure.

[0019] The merchant user popularity analysis algorithm is invoked to determine the store popularity score of each merchant user based on the store popularity information of each merchant user;

[0020] Identify popular merchant users whose store popularity score exceeds the store popularity score threshold, generate a merchant-specific product feature cluster with multiple product feature nodes for the popular merchant users, and store the product feature information of the popular merchant users in each of the product feature nodes. The merchant-specific product feature cluster is marked with the merchant identifier of the popular merchant user.

[0021] In a further embodiment, the process includes detecting whether product feature information corresponding to the product identifier exists in each target product feature node. If it exists, the product feature information corresponding to the product identifier is obtained from the target product feature node. Alternatively, if the target product feature node does not contain product feature information corresponding to the product identifier, the process includes the following steps before retrieving the information from the medium- or low-heat database in the medium- or low-heat partition storage cluster:

[0022] Obtain product popularity information for multiple product objects on the platform, including product sales volume, product visits, or product click-through rate;

[0023] The popular product analysis algorithm is invoked to determine the popularity level of each product object based on the product popularity information of each product object. The popularity level of the product is divided into high popularity products, medium popularity products and low popularity products.

[0024] Identify high-popularity product objects with high popularity, obtain the product identifier and product feature information of each high-popularity product object, determine the target product feature node within the product identifier range hit in the product feature node pool for each product identifier, and store each product feature information into the target product feature node corresponding to its product identifier;

[0025] The popularity of the products is determined to be medium-popularity products and low-popularity products, respectively. The product identifiers and product feature information of each medium-popularity product and each low-popularity product are stored in the medium-popularity medium-performance database and the low-popularity low-performance database in the medium-low popularity partitioned storage cluster, respectively.

[0026] In a further embodiment, detecting whether the corresponding product feature information of the product identifier exists in each of the target product feature nodes includes the following steps:

[0027] The preset hash algorithm is invoked to calculate the hash value corresponding to the product identifier;

[0028] Obtain the product feature storage hash table of the target product feature node corresponding to the product identifier, and query the storage element corresponding to the hash value in the product feature storage hash table;

[0029] When the storage element is indicated to exist, it indicates that the target product feature node contains product feature information corresponding to the product identifier, and the product feature information corresponding to the storage element is obtained from the target product feature node.

[0030] In a further embodiment, when the target product feature node does not contain product feature information corresponding to the product identifier, after obtaining the product feature information corresponding to the product identifier from the medium-hot, medium-performance library or the low-hot, low-performance library in the medium-low-hot partitioned storage cluster, the following steps are included:

[0031] When there is no corresponding product feature information in the target product feature node corresponding to the product identifier, check whether the product identifier hits the product feature storage range of the medium-hot, medium-performance library in the medium-low hotness partition storage cluster.

[0032] When the product identifier matches the product feature storage range, the product feature information corresponding to the product identifier is obtained from the medium-hot, medium-performance library. If it does not match, the product feature information corresponding to the product identifier is obtained from the low-hot, low-performance library of the medium-low-hot partition storage cluster.

[0033] The acquired product feature information is stored in the target product feature node.

[0034] In a further embodiment, the products are sorted according to the product feature information corresponding to each product identifier to generate a corresponding product sorting list, and the product sorting list is pushed to the corresponding user terminal, including the following steps:

[0035] Obtain product feature information corresponding to multiple product identifiers. The product feature information includes basic product feature information and product business feature information. The basic product feature information includes product color, product size or product brand. The product business feature information includes the target sales audience, product sales volume, product click-through rate and product purchase rate.

[0036] Obtain the product object to which each product feature information belongs, call the preset product sorting algorithm, sort each product object according to the product feature information of each product object, and generate the corresponding product sorting list;

[0037] Once the user terminal corresponding to the coarse recall information of the searched products is identified, the product sorting list is pushed to the user terminal.

[0038] On the other hand, a product feature recall device provided to meet one of the purposes of this application includes:

[0039] The product coarse recall module is used to obtain the search product coarse recall information pushed by the product search service, and to obtain the merchant identifier and multiple product identifiers corresponding to the search product coarse recall information.

[0040] The dedicated cluster detection module is used to detect whether there is a merchant-specific product feature cluster corresponding to the merchant identifier. If it exists, the module obtains the product feature information corresponding to each product identifier from the product feature nodes in the merchant-specific product feature cluster.

[0041] The target node hitting module is used to detect the range of product identifiers that each product identifier hits in the product feature node pool when the merchant-specific product feature cluster does not exist, and to determine the target product feature node corresponding to each range of product identifiers.

[0042] The high-temperature feature acquisition module is used to detect whether there is product feature information of the corresponding product identifier in each of the target product feature nodes. If there is, the product feature information corresponding to the product identifier is obtained from the target product feature node. The product feature node stores the product feature information corresponding to the high-temperature product.

[0043] The cluster feature acquisition module is used to acquire the product feature information corresponding to the product identifier from the medium-hot, medium-performance library or the low-hot, low-performance library in the medium-low-hot partition storage cluster when the product feature information corresponding to the product identifier does not exist in the target product feature node.

[0044] The product feature sorting module is used to sort products according to the product feature information corresponding to each product identifier, generate a corresponding product sorting list, and push the product sorting list to the corresponding user terminal.

[0045] In a further embodiment, the product coarse recall module includes:

[0046] The product search request response submodule is used to respond to product search requests pushed by the user and obtain the product search terms and merchant identifiers corresponding to the product search request.

[0047] The product object search submodule is used to obtain the product database corresponding to the merchant identifier and query one or more product objects corresponding to the product search term from the product database.

[0048] The coarse recall information generation submodule is used to obtain the product identifier corresponding to each of the product objects and generate coarse recall information for search products with each of the product identifiers and each of the merchant identifiers.

[0049] In a further embodiment, the dedicated cluster detection module includes:

[0050] The store popularity information acquisition submodule is used to acquire store popularity information of multiple merchant users on the platform. The store popularity information includes store product sales volume, store order volume, store visit volume or store exposure volume.

[0051] The store popularity score determination submodule is used to call the merchant user popularity analysis algorithm to determine the store popularity score of each merchant user based on the store popularity information of each merchant user;

[0052] The dedicated cluster generation submodule is used to identify popular merchant users whose store popularity score exceeds the store popularity score threshold, generate a merchant-specific product feature cluster with multiple product feature nodes for the popular merchant users, and store the product feature information of the popular merchant users in each of the product feature nodes. The merchant-specific product feature cluster is marked with the merchant identifier of the popular merchant user.

[0053] In a further embodiment, the high-heat feature acquisition module includes:

[0054] The hash value calculation submodule is used to call a preset hash algorithm to calculate the hash value corresponding to the product identifier;

[0055] The storage element query submodule is used to obtain the product feature storage hash table of the target product feature node corresponding to the product identifier, and query the storage element corresponding to the hash value in the product feature storage hash table;

[0056] The product effect information acquisition submodule is used to acquire the product effect information corresponding to the product identifier from the target product feature node when the storage element is characterized as existing.

[0057] In a further embodiment, the product feature sorting module includes:

[0058] The product feature information acquisition submodule is used to acquire product feature information corresponding to multiple product identifiers. The product feature information includes basic product feature information and product business feature information. The basic product feature information includes product color, product size or product brand, and the product business feature information includes product target sales audience, product sales volume, product click rate and product purchase rate.

[0059] The product sorting list generation submodule is used to obtain the product objects to which each product feature information belongs, call the preset product sorting algorithm, sort each product object according to the product feature information of each product object, and generate the corresponding product sorting list;

[0060] The product sorting list push submodule is used to determine the user terminal corresponding to the coarse recall information of the searched products and push the product sorting list to the user terminal.

[0061] In another aspect, a computer device provided for one of the purposes of this application includes a central processing unit and a memory, the central processing unit being used to invoke and run a computer program stored in the memory to perform the steps of the product feature recall method described in this application.

[0062] In another aspect, a computer-readable storage medium is provided to suit another purpose of this application, which stores in the form of computer-readable instructions a computer program implemented according to the product feature recall method, which, when invoked by a computer, performs the steps included in the method.

[0063] The technical solution of this application has many advantages, including but not limited to the following aspects:

[0064] This application addresses the popularity of merchants and products on e-commerce platforms by implementing layered storage of product feature information. This approach aims to rationally allocate platform data storage and search resources, improve the retrieval efficiency of product feature information for popular merchants and products, and ensure the efficient use of data storage and search resources. For popular merchants on the e-commerce platform, a high-performance merchant-specific product feature cluster will be built to store their product feature information. This will improve the retrieval efficiency of product feature information for frequently visited popular merchants and meet the response efficiency requirements for a large number of product search requests from frequently visited stores. Furthermore, it will quickly generate corresponding product ranking queues for product search services in frequently visited stores, enhancing the product search experience for users. For product objects with different levels of popularity on the e-commerce platform, this application uses a high-performance product feature pool, a medium-performance medium-popularity library, and a low-popularity low-performance library to store high-popularity products respectively. This application categorizes product feature information into high-popularity, medium-popularity, and low-popularity product objects. Based on the popularity of each product object, the feature information is stored in storage spaces with different performance levels. High-performance storage resources from the e-commerce platform are used for the feature information of frequently accessed, high-popularity products, ensuring faster retrieval of this information. For medium- and low-popularity products, higher-performance storage spaces are prioritized for storing their feature information to improve retrieval efficiency. This demonstrates that the application rationally allocates data storage and search resources of different performance levels within the e-commerce platform, maximizing the utilization of search computing resources in the product search service. This effectively improves the service stability of the product search service, meets the high volume of product search requests from popular merchants or products, enhances product ranking processing efficiency, and ensures a superior product search experience for both merchants and buyers on the platform. Attached Figure Description

[0065] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0066] Figure 1 The network architecture of the e-commerce platform exemplified in this application;

[0067] Figure 2 This is a flowchart illustrating a typical embodiment of the product feature recall method of this application;

[0068] Figure 3 This is a flowchart illustrating a specific implementation method for generating coarse recall information of searched products in the product search service described in this application.

[0069] Figure 4This is a flowchart illustrating the specific implementation method for building a merchant-specific product feature cluster for popular merchant users in this application.

[0070] Figure 5 This is a flowchart illustrating a specific implementation method for determining the popularity of a product object in this application, and for performing data stratification processing of product feature information based on the popularity of the product.

[0071] Figure 6 This is a flowchart illustrating the specific implementation method for detecting whether product feature information corresponding to a product identifier exists in the product feature nodes in this application;

[0072] Figure 7 This is a flowchart illustrating the specific implementation method of storing product feature information recalled from the low-to-medium heat partition storage cluster to the corresponding product feature node in this application.

[0073] Figure 8 This is a flowchart illustrating the specific implementation method of sorting goods based on the product feature information of the product object, generating a corresponding product sorting list and pushing it out in this application.

[0074] Figure 9 This is a schematic diagram of the product feature recall device of this application;

[0075] Figure 10 This is a schematic diagram of the structure of a computer device used in this application. Detailed Implementation

[0076] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0077] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this application means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0078] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0079] like Figure 1 In the network architecture shown, the e-commerce platform 82 is deployed on the Internet to provide corresponding services to its users. Similarly, the devices 80 of the merchant users and the devices 81 of the consumer users of the e-commerce platform 82 are also connected to the Internet to use the services provided by the e-commerce platform.

[0080] An exemplary e-commerce platform 82 provides supply and demand matching of products and / or services to the general public through the Internet infrastructure. In e-commerce platform 82, products and / or services are provided as commodity information. For the sake of simplicity, the concepts of commodity and product are used in this application to refer to the products and / or services in e-commerce platform 82. Specifically, these may be physical products, digital products, tickets, service subscriptions, other offline services, etc.

[0081] In reality, various entities can access e-commerce platform 82 as users and utilize its online services to participate in the business activities facilitated by the platform. These entities can be natural persons, legal persons, or social organizations. Corresponding to the two types of entities in business activities—merchants and consumers—e-commerce platform 82 has two corresponding categories of users: merchant users and consumer users. Entities involved in the product distribution chain in business activities, including manufacturers, sellers, retailers, and logistics providers, can all use online services on e-commerce platform 82 as merchant users. Similarly, consumers in business activities, including actual or potential consumers, can use online services on e-commerce platform 82 as consumer users. In actual business activities, the same entity can operate as both a merchant user and a consumer user; this should be interpreted flexibly.

[0082] The infrastructure used to deploy the e-commerce platform 82 mainly includes the backend architecture and frontend devices. The backend architecture runs various online services through a service cluster, including middleware or frontend services for the platform, services for consumers, and services for merchants, to enrich and improve its service functions. The frontend devices mainly cover the terminal devices used by users as clients to access the e-commerce platform 82, including but not limited to various mobile terminals, personal computers, and point-of-sale devices. For example, merchant users can use their terminal device 80 to enter product information for their online stores or use the interfaces opened by the e-commerce platform to generate their product information; consumer users can use their terminal device 81 to access the webpage of the online store implemented by the e-commerce platform 82, trigger the shopping process by clicking the shopping button provided on the webpage, and call various online services provided by the e-commerce platform 82 during the shopping process to achieve the purpose of placing an order.

[0083] In some embodiments, the e-commerce platform 82 may be implemented via a processing facility including a processor and memory, which stores a set of instructions that, when executed, cause the e-commerce platform 82 to perform the e-commerce and support functions as described in this application. The processing facility may be part of a server, client, network infrastructure, mobile computing platform, cloud computing platform, fixed computing platform, or other computing platform, and may provide electronic components, merchant devices, payment gateways, application developers, marketing channels, transportation providers, customer devices, point-of-sale devices, etc., for the e-commerce platform 82.

[0084] E-commerce platform 82 can provide online services such as cloud computing services, Software as a Service (SaaS), Infrastructure as a Service (IaaS), Platform as a Service (PaaS), Desktop as a Service (DaaS), Hosted Software as a Service, Mobile Backend as a Service (MBaaS), and Information Technology Management as a Service (ITMaaS). In some embodiments, the various functional components of e-commerce platform 82 can be implemented to operate on various platforms and operating systems. For example, for an online store, its administrator user enjoys the same or similar functions regardless of whether it is on iOS, Android, HomonyOS, or a web page.

[0085] E-commerce platform 82 enables merchants to create their own independent websites to run their online stores. It provides merchants with corresponding business management engine instances, allowing them to establish, maintain, and operate one or more online stores across these independent websites. The business management engine instance can be used for content management, task automation, and data management for one or more online stores. It can be configured through interfaces or built-in components to support various specific business processes in the online store, supporting business activities. Independent websites are the infrastructure of e-commerce platform 82, which offers cross-border services. Merchants can maintain their online stores relatively independently and centrally based on these independent websites. Independent websites typically have dedicated domain names and storage space, and different independent websites are relatively independent. E-commerce platform 82 can provide standardized or customized technical support for a large number of independent websites, allowing merchants to customize a business management engine instance that suits their needs and use it to maintain one or more online stores.

[0086] Online stores can be configured and maintained in the backend by merchant users logging into their Business Management Engine instance as administrators. Supported by the various online services provided by the e-commerce platform 82's infrastructure, merchant users can configure various functions within their online stores and view various data as administrators. For example, merchant users can manage various aspects of their online stores, such as viewing recent online store activities, updating the online store's product catalog, managing orders, recent visit activity, and total order activity. Merchant users can also view more detailed information about their business and visitors to their online store by obtaining reports or metrics, such as displaying a sales summary of the merchant's overall business, specific sales and engagement data from promotional sales and marketing channels, etc.

[0087] E-commerce platforms 82 can provide communication facilities and associated merchant interfaces for electronic communication and marketing. For example, they can utilize electronic messaging aggregation facilities to collect and analyze communication interactions between merchants, consumers, merchant devices, customer devices, point-of-sale devices, etc., aggregating and analyzing communications to increase the potential for product sales. For instance, a consumer may have product-related questions, which could lead to a dialogue between the consumer and the merchant (or an automated processor-based agent representing the merchant), where the communication facilities handle the interaction and provide the merchant with analysis on how to increase the probability of a sale.

[0088] In some embodiments, applications suitable for installation on terminal devices can be provided to serve the access needs of different users, enabling various users to access the e-commerce platform 82 by running the application on their terminal devices. Examples include the merchant backend module of online stores within the e-commerce platform 82. During the process of conducting business activities through these functions, the e-commerce platform 82 can implement various functions related to business activities as middleware or online services and expose corresponding interfaces. Then, toolkits corresponding to the interface access functions are embedded into the application to achieve functional expansion and task completion. The business management engine can include a series of basic functions and expose these functions to online services and / or applications via APIs. Online services and applications use the corresponding functions by remotely calling the corresponding APIs.

[0089] With the support of various components of the Business Management Engine instance, the e-commerce platform 82 can provide online shopping functionality, enabling merchants to connect with customers in a flexible and transparent manner. Consumers can select items online, create orders, provide delivery addresses in the orders, and complete payment confirmation. Merchants can then review and fulfill or cancel orders. The review component included with the Business Management Engine instance ensures compliant use of business processes, guaranteeing that orders are suitable for fulfillment before actual execution. Orders may sometimes be fraudulent and require verification (e.g., ID checks). Payment methods that require merchants to wait to ensure receipt of funds can mitigate this risk, and so on. Order risks may arise from fraud detection tools submitted by third parties through order risk APIs, etc. Before fulfillment, merchants may need to obtain or wait to receive payment information to mark the order as paid before preparing to deliver the product. Such situations can all be subject to appropriate review. The review process can be implemented by the fulfillment component. Merchants can leverage fulfillment components to review and adjust operations, and trigger related fulfillment services. These include: manual fulfillment services, used when merchants select and pack products into boxes, purchase shipping labels and enter tracking numbers, or simply mark items as fulfilled; custom fulfillment services, which can define email notifications; API fulfillment services, which can trigger third-party applications to create fulfillment records; legacy fulfillment services, which can trigger custom API calls from the Commerce Management Engine to third parties; and gift card fulfillment services, which can generate and activate gift cards. Merchants can use an order printer application to print shipping documents. The fulfillment process can be executed once items are packed and ready for shipment, tracked, delivered, and verified by the consumer.

[0090] It can be seen that the services provided by e-commerce platforms are based on products, and the corresponding product data is the foundational data of these platforms. Providing product information through this data and mining and utilizing it are fundamental to various technical services. This includes using user transaction data and product data from the e-commerce platform to provide basic services for the operation of a real-time streaming data processing system. Therefore, the real-time streaming data processing system of this application can run on any one or more servers within the e-commerce platform's cluster, enabling the use of various product data provided by the e-commerce platform to achieve various functions.

[0091] The product feature recall method of this application can be programmed into a computer program product and deployed on a client or server for execution. For example, in an exemplary application scenario of this application, it can be deployed on the server of an e-commerce customer service platform. In this way, the method can be executed by human-computer interaction with the process of the computer program product through a graphical user interface by accessing the interface opened after the computer program product is run.

[0092] Please see Figure 2 The product feature recall method of this application, in its typical embodiment, includes the following steps:

[0093] Step S11: Obtain the coarse recall information of search products pushed by the product search service, and obtain the merchant identifier and multiple product identifiers corresponding to the coarse recall information of search products:

[0094] E-commerce platforms typically provide product search services for their users. Merchants and buyers on the platform can use these services to edit search terms to find products available on the platform or in the stores of the corresponding merchants. This allows merchants to search for products in their stores and develop corresponding sales strategies, while buyers can browse the stores to find the products they wish to purchase.

[0095] After a user on the platform edits a corresponding product search term using the product search service, the user's client will generate a product search request containing the product search term and push the product search request to the business server responsible for the product search service. The business server responds to the product search request, obtains the product search term contained in the product search request, and determines the merchant identifier corresponding to the product search request, that is, the merchant identifier corresponding to the store where the user's client is using the product search service. First, it determines the product database corresponding to the merchant identifier. The product database stores the product objects in the merchant's store. The product objects have corresponding product titles, product images, product prices, product attributes, and product descriptions, etc. Then, it queries the product database to retrieve one or more product objects corresponding to the product search term. Generally, the product objects corresponding to the product search term can be matched by text matching, or by a convergent image-text semantic matching model to match product objects similar to the product search term. After retrieving each product object corresponding to the product search term from the product database, the product identifiers of these product objects are obtained to generate coarse recall information of search products containing each product identifier and the merchant identifier.

[0096] After generating coarse product recall information corresponding to a product search request, the product search service pushes this information to a server responsible for product feature acquisition and product object sorting. Upon receiving the coarse product recall information, the server acquires the merchant identifier and multiple product identifiers contained within it. Based on these identifiers, it obtains the product feature information of the product objects corresponding to each identifier. Then, based on the product feature information, it sorts the product objects to generate a corresponding product sorting list. This sorting list is then pushed to the user terminal using the product search service, allowing the user terminal to view the product objects sequentially in the graphical user interface.

[0097] Step S12: Detect whether a merchant-specific product feature cluster corresponding to the merchant identifier exists. If it exists, obtain the product feature information corresponding to each product identifier from the product feature nodes in the merchant-specific product feature cluster.

[0098] After obtaining the merchant identifier corresponding to the coarse recall information of the searched products, it will detect whether there is a merchant-specific product feature cluster corresponding to the merchant identifier. If it exists, it will obtain the product feature information corresponding to each product identifier corresponding to the coarse recall information of the searched products from the product feature nodes in the merchant-specific product feature cluster.

[0099] The aforementioned merchant-specific product feature clusters are generally built for popular merchant users on e-commerce platforms. By building merchant-specific product feature clusters for popular merchant users to store the product feature information of their product objects, e-commerce platforms can improve the speed of recalling the product feature information of popular merchant users.

[0100] Popular merchants on e-commerce platforms are generally those with high store popularity data such as page views, product sales, or exposure. Specifically, by using popular store information such as store product sales, order volume, page views, or exposure, a merchant popularity analysis algorithm is used to calculate the weight of each type of popular store information to determine the corresponding store popularity score for the merchant. When the store popularity score exceeds a preset popularity score threshold, the merchant will be identified as a popular merchant. At this time, a corresponding merchant-specific product feature cluster will be built for that merchant.

[0101] The aforementioned merchant-specific product feature cluster is generally built on a Redis database structure. This cluster contains multiple product feature nodes, each storing product feature information for each product object in the merchant's store corresponding to the cluster. Furthermore, each feature node stores product feature information within a specific product identifier range; that is, it only stores product feature information for product objects whose identifiers are accessed by that identifier. Compared to traditional database structures like MySQL, TiDB, and HBase, this Redis-based merchant-specific product feature cluster offers higher data processing efficiency, providing faster product feature information search for popular merchants.

[0102] When there is a merchant-specific product feature cluster corresponding to the merchant identifier of the coarse recall information of search products, the product feature nodes corresponding to the product identifier range hit by each product identifier of the coarse recall information of search products will be queried from the merchant-specific product feature cluster, and then the product feature information corresponding to each product identifier will be obtained from each product feature node.

[0103] Step S13: When the merchant-specific product feature cluster does not exist, detect the range of product identifiers that each product identifier matches in the product feature node pool, and determine the target product feature node corresponding to each product identifier range:

[0104] When there is no merchant-specific product feature cluster corresponding to the merchant identifier in the search for coarse product recall information, the product feature information corresponding to each product identifier in the search for coarse product recall information will be obtained from each product feature node in the product feature node pool.

[0105] The product feature node pool is used to store product feature information of popular products on the e-commerce platform. The e-commerce platform identifies frequently accessed or purchased products as popular products and stores their product feature information in each feature node of the pool to improve the efficiency of retrieving this information. Specifically, it obtains product popularity information for multiple products on the e-commerce platform, including metrics such as sales volume, page views, and click-through rate. Then, it calls a popular product analysis algorithm to determine the popularity level of each product based on this information. The popularity level is categorized as high or low. For high-popularity products (high-popularity products), the product identifier and product feature information of each high-popularity product object will be obtained. The target product feature node within the product identifier range matched by each product identifier in the product feature node pool will be determined, and the product feature information of each product will be stored in the target product feature node corresponding to its product identifier. For medium-popularity products and low-popularity products (medium-popularity products and low-popularity products), the product identifier and product feature information of each medium-popularity product and each low-popularity product object will be stored in the medium-popularity medium-performance library and the low-popularity low-performance library in the medium-low popularity partitioned storage cluster, respectively.

[0106] It can be seen that, in addition to building a merchant-specific product feature cluster for popular merchants to store product feature information of their product objects, it also has a product feature node pool for storing product feature information of popular product objects, a medium-popularity medium-performance library for storing product feature information of medium-popularity products, and a low-popularity low-performance library for storing product feature information of low-popularity products. The product feature nodes in the product feature node pool are similar to those in the merchant-specific product feature cluster, both built on a Redis database structure. They have higher performance than the medium-popularity medium-performance library and the low-popularity low-performance library, enabling high-speed retrieval of product feature information for popular product objects. The medium-popularity medium-performance library is generally constructed with storage media hardware superior to the low-popularity low-performance library. For example, the medium-popularity medium-performance library... The server clusters of the medium-popularity, medium-performance database typically use solid-state drives (SSDs) as the primary storage medium for storing product feature information, while the low-popularity, low-performance database typically uses hard disk drives (HDDs) as the primary storage medium. This results in a faster product feature information retrieval rate for the medium-popularity, medium-performance database compared to the low-popularity, low-performance database. It can be seen that this method, in addition to stratifying product feature information storage based on the popularity of merchant users, also stratifies it based on the popularity of product objects. Product feature information for product objects with different popularity levels is stored in storage clusters with different performance levels, thus rationally allocating the platform's data storage and search resources, improving the retrieval efficiency of product feature information for popular products, and ensuring the rational utilization of data storage and search resources.

[0107] Each product feature node in the product feature node pool has its own corresponding product identifier range. Each product feature node stores product feature information of product objects whose product identifiers are within its product identifier range. For example, when the product identifier range of a product feature node is 0 to 1000, the node only stores product feature information of product objects whose product identifiers are within the range of 0 to 1000.

[0108] After determining the range of product identifiers corresponding to the product identifiers in the product feature node pool, the product feature nodes corresponding to the product identifier range are used as the target product feature nodes corresponding to the product identifiers, so as to detect whether the target product feature nodes store the product feature information corresponding to the product identifiers.

[0109] Step S14: Detect whether the product feature information corresponding to the product identifier exists in each of the target product feature nodes. If it exists, obtain the product feature information corresponding to the product identifier from the target product feature node. The product feature node stores the product feature information corresponding to high-popularity products.

[0110] After determining the target product feature node corresponding to the product identifier in the product feature node pool, it will query whether the target product feature node stores the product feature information corresponding to the product identifier. As mentioned above, the product feature node generally only stores the product feature information of popular product objects.

[0111] Product feature nodes typically have corresponding product feature storage hash tables to construct Bloom filters for querying whether product feature information corresponding to a product identifier exists in a product feature node. This improves the efficiency of querying whether product feature information exists in a product feature node. The product feature storage hash table consists of a binary bit array and a hash function. Each element in the binary bit array is represented by 0 or 1 to indicate whether data exists, where 0 indicates no data and 1 indicates data exists. Specifically, after determining that a product identifier is in the corresponding target product feature node, a preset hash algorithm is called to calculate multiple hash values ​​corresponding to the product identifier and obtain the quotient of the target product feature node. The product feature storage hash table is used to query the storage element corresponding to the hash value in the product feature storage hash table. That is, the element position corresponding to each hash value in the product feature storage hash table is queried, and then it is checked whether each storage element represents the existence of data, that is, whether each storage element is 1. If all storage elements represent the existence of data, that is, all storage elements are 1, it means that the target product feature node stores the product feature information corresponding to the product identifier. At this time, the product feature information corresponding to each storage element is obtained from the target product feature node, and the product feature information is used as the product feature information corresponding to the product identifier.

[0112] Step S15: If the target product feature node does not contain product feature information corresponding to the product identifier, then obtain the product feature information corresponding to the product identifier from the medium-hot, medium-performance library or the low-hot, low-performance library in the medium-low-hot partitioned storage cluster.

[0113] When there is no product feature information corresponding to any product identifier in the target product feature node, the product feature information corresponding to the product identifier will be obtained from the low-to-medium popularity partition storage cluster that stores the product feature information of low-to-medium popularity product objects.

[0114] The aforementioned medium-to-low popularity partitioned storage cluster consists of a medium-popularity medium-performance library and a low-popularity low-performance library. The medium-popularity medium-performance library stores product feature information of medium-popularity product objects, while the low-popularity low-performance library stores product feature information of low-popularity product objects. As mentioned earlier, the performance of the medium-popularity medium-performance library is generally higher than that of the low-popularity low-performance library. This hierarchical storage of product feature information based on the popularity of product objects allows for a higher rate of recalling product feature information of medium-popularity product objects compared to recalling product feature information of low-popularity product objects, thus making reasonable use of the platform's data storage resources.

[0115] Regarding the order in which product feature information is stored in the medium-popularity, medium-performance library and the low-popularity, low-performance library, it will be stored in ascending order according to the product identifier of the product object. The product identifier is generally a string. Therefore, the product identifiers can be sorted according to the size represented by the string of the product identifier. Then, the product feature information corresponding to each product identifier is stored in the medium-popularity, medium-performance library and the low-popularity, low-performance library in sequence, thereby improving the speed of obtaining the corresponding product feature information from the medium-popularity, medium-performance library and the low-popularity, low-performance library.

[0116] Similarly, in order to improve the efficiency of querying product feature information from the medium-popularity performance database, the medium-popularity performance database generally also has a corresponding product feature storage hash table. This allows for the use of a Bloom filter query method, similar to querying whether product feature information corresponding to a product identifier exists from the product feature nodes. By calculating the hash value corresponding to the product identifier, the database queries the storage element corresponding to the hash value in the product feature storage hash table to detect whether product feature information corresponding to the product identifier exists in the medium-popularity performance database.

[0117] When retrieving product feature information corresponding to a product identifier in the medium-low heat partition storage cluster, the system will first query whether the medium heat performance database contains product feature information corresponding to the product identifier. If the medium heat performance database does not contain product feature information corresponding to the product identifier, the system will then query the low heat low performance database to retrieve the product feature information corresponding to the product identifier. The low heat low performance database can store all product feature information in the platform to back up the product feature information of each product object for the platform.

[0118] After obtaining the product feature information corresponding to the product identifier from the low-to-medium popularity partition storage cluster, the product feature information can be stored in the target product feature node corresponding to the product identifier to improve the efficiency of the next retrieval of the product feature information. Of course, a corresponding retrieval threshold can also be set. When the number of times the product feature information of any product identifier is obtained from the low-to-medium popularity partition storage cluster exceeds the retrieval threshold, the product feature information will be stored in the target product feature node of the product identifier. While ensuring the storage space of each product feature node in the product feature node pool, the product feature information that is frequently recalled can also be stored in the product feature node to improve its retrieval efficiency.

[0119] Step S16: Sort the products according to the product feature information corresponding to each product identifier, generate a corresponding product sorting list, and push the product sorting list to the corresponding user terminal:

[0120] After retrieving the product feature information corresponding to each product identifier in the current push of the product search service, the product objects corresponding to each product identifier will be sorted according to the product feature information to generate a sorted product list corresponding to each product object and then push it.

[0121] The product feature information includes basic product feature information and business product feature information. The basic product feature information includes information such as product color, product size, product brand, product price or product name. The business product feature information includes information such as the target sales audience, product sales volume, product click rate, product purchase rate, product inventory, and product exposure rate, which are obtained through statistics or experiments and other algorithms.

[0122] After retrieving the product feature information corresponding to each product identifier in the coarse product recall information, if the coarse product recall information is generated when a buyer user on an e-commerce platform searches for products, then the buyer user's consumption preference information will be obtained. This consumption preference information includes user consumption profile information, user historical consumption information, and user personal characteristic information. Based on this consumption preference information and the product feature information, the product objects with each product feature information will be sorted, prioritizing those that match the buyer user's consumption preferences on the e-commerce platform. This allows the buyer user to quickly browse products that match their preferences after using the product search service, improving their online shopping experience. Specifically, a product sorting model trained to convergence can be used to calculate the similarity between the buyer user's consumption preference information and the product feature information of each searched product object. Products with higher similarity will be sorted higher, and a corresponding product sorting list will be generated and pushed to the buyer user's client.

[0123] After retrieving the product feature information corresponding to each product identifier in the coarse product recall information, when the coarse product recall information is generated by merchants on e-commerce platforms when searching for products, the product objects corresponding to each product identifier can be sorted based on the product sorting rules set by the merchants when using the product search service, according to the product feature information corresponding to each product identifier. For example, the product sorting rules may include sorting by price, product sales volume, product click-through rate, product purchase rate, product inventory, or product exposure rate, or sorting rules formulated by combining multiple product feature information for weight calculation, etc. Then, according to the product sorting rules set by the merchants, the product objects are sorted based on the product feature information of each product object, and a corresponding product sorting list is generated and pushed to the user terminal where the merchants are located.

[0124] The above typical embodiments and their variations fully disclose the implementation scheme of the product feature recall method of this application. However, various variations of this method can still be derived by changing and expanding some technical means. Other embodiments are briefly described below:

[0125] Please refer to Figure 3 In one embodiment, in response to the coarse recall information of search products pushed by the product search service, the following steps are taken to obtain the merchant identifier and multiple product identifiers corresponding to the coarse recall information of search products:

[0126] Step S111: Respond to the product search request pushed by the user client and obtain the product search terms and merchant identifiers corresponding to the product search request.

[0127] When a user on an e-commerce platform uses the product search service to search for products in a merchant's shop, they typically edit the corresponding product search terms to describe the product they are searching for. After the user completes the product search term editing, the user's client will generate a product search request with the search terms. The product search request can carry the merchant identifier of the merchant user to whom the product search is conducted. The generated product search request will be pushed to the product search service.

[0128] Step S112: Obtain the product database corresponding to the merchant identifier, and query one or more product objects corresponding to the product search term from the product database.

[0129] In response to a product search request pushed by the user, and to obtain the product search terms and merchant identifiers corresponding to the product search request, the system first determines the product database corresponding to the merchant identifier. The product database generally stores the product objects in the store of the merchant identifier. Then, it queries the product database to retrieve one or more product objects corresponding to the product search terms.

[0130] Step S113: Obtain the product identifier corresponding to each of the product objects, and generate coarse search product recall information containing each of the product identifiers and each of the merchant identifiers:

[0131] After retrieving one or more product objects corresponding to the product search term, the product identifier of each product object is obtained, and then coarse search product recall information with each product identifier and the merchant identifier is generated. Then, based on the merchant identifier and each product identifier, the product feature information of the product object corresponding to each product identifier is obtained, and according to the product feature information of each product object, the product objects are sorted to generate a corresponding product sorting list.

[0132] In this embodiment, while providing product search services to platform users, the product search service also retrieves product feature information to sort product objects. The product retrieval is carried out in stages. First, the basic information of the searched products is retrieved, and then the feature information of the products is retrieved. Then, the products are sorted according to the feature information of the products, ensuring the independence between the basic product search business and the product feature retrieval business, which facilitates business maintenance.

[0133] Please refer to Figure 4 In one embodiment, before detecting whether a merchant-specific product feature cluster corresponding to the merchant identifier exists, and if so, before obtaining the product feature information corresponding to each product identifier from the product feature nodes in the merchant-specific product feature cluster, the following steps are included:

[0134] Step S121: Obtain store popularity information from multiple merchant users on the platform. This store popularity information includes store product sales volume, store order volume, store pageviews, or store exposure.

[0135] To differentiate the popularity of different merchants on e-commerce platforms, a merchant-specific product feature cluster is constructed for merchants with higher popularity to store product feature information of each product in their store. Generally, a store popularity score is determined based on the store popularity information of each merchant. The store popularity information includes store product sales volume obtained by counting the sales volume of each product in the store, store order volume obtained by counting the order volume in the store, store visit volume obtained by counting the number of times the store is visited within a certain period of time, or store exposure volume obtained by counting the exposure of the store on the e-commerce platform or other third-party platforms.

[0136] Step S122: Invoke the merchant user popularity analysis algorithm to determine the store popularity score of each merchant user based on their store popularity information.

[0137] The merchant user popularity analysis algorithm is configured with weights corresponding to different store popularity information, so as to determine the store popularity score of the merchant user by calculating the weights of store popularity information such as store product sales volume, store order volume, store visits or store exposure.

[0138] Step S123: Identify popular merchant users whose store popularity score exceeds the store popularity score threshold, generate a merchant-specific product feature cluster with multiple product feature nodes for the popular merchant users, and store the product feature information of the popular merchant users into each of the product feature nodes. The merchant-specific product feature cluster is marked with the merchant identifier of the popular merchant user.

[0139] After determining the popularity score of a merchant user's store, it will be judged whether the popularity score exceeds the popularity score threshold. If it does, the merchant user will be identified as a popular merchant user. The popularity score threshold is set based on the maximum popularity score of the store. For example, if the maximum popularity score of the store is 100, the popularity score threshold can be set in the range of 70 to 90. Of course, those skilled in the art can design it flexibly, which will not be elaborated here.

[0140] Once a merchant user is identified as a popular merchant user, a corresponding merchant-specific product feature cluster will be built for them to store the product feature information of each product object in the popular merchant user's store. The merchant-specific product feature cluster also has its corresponding merchant user's merchant identifier, so as to distinguish different merchant-specific product feature clusters.

[0141] In this embodiment, a merchant-specific product feature cluster is constructed for popular merchants on the platform to store product feature information of their products. This improves the recall rate of product feature information of frequently visited popular merchants' stores on the platform, meets the response requirements of high product search request volume in popular merchants' stores, and ensures the product search experience for both merchants and buyers on the platform.

[0142] Please refer to Figure 5 In one embodiment, the process includes detecting whether product feature information corresponding to the product identifier exists in each of the target product feature nodes. If it exists, the product feature information corresponding to the product identifier is obtained from the target product feature node. Alternatively, if the product feature information corresponding to the product identifier does not exist in the target product feature node, the process includes the following steps before retrieving the information from the medium-hot database or low-hot database in the medium-low-hot partition storage cluster:

[0143] Step S131: Obtain product popularity information for multiple product objects on the platform. This product popularity information includes product sales volume, product pageviews, or product click-through rate.

[0144] To differentiate the popularity of different product objects on an e-commerce platform, and to store their product feature information in a product feature node pool, a medium-popularity medium-performance library, or a low-popularity low-performance library based on their popularity, it is necessary to obtain the product popularity information of each product object on the e-commerce platform. This product popularity information includes the product sales volume, the number of times the product object is accessed on the e-commerce platform, or the product click-through rate, etc.

[0145] Step S132: Invoke the popular product analysis algorithm to determine the popularity level of each product based on its popularity information. The popularity level is categorized into high-popularity products, medium-popularity products, and low-popularity products.

[0146] After obtaining the product popularity information of each product object, the popular product analysis algorithm will be called to determine the popularity of each product object. The popular product analysis algorithm generally has weights corresponding to various types of product popularity information, and then the product popularity information of the product object will be weighted to determine the popularity of the product object. The popularity of the product includes high-popularity products, medium-popularity products and low-popularity products.

[0147] Step S133: Identify high-popularity product objects with high popularity, obtain the product identifier and product feature information of each high-popularity product object, determine the target product feature node within the product identifier range matched by each product identifier in the product feature node pool, and store each product feature information in the target product feature node corresponding to its product identifier.

[0148] For highly popular products, their product feature information will be stored in a product feature node pool. In the product feature node pool, the product feature information will be stored in the corresponding product feature node according to the product identifier of the product object. First, the range of product identifiers that the product object's product identifier matches in the product feature node pool is determined. Then, the target product feature node corresponding to the range of product identifiers is determined so that the product feature information of the product object can be stored in the target product feature node.

[0149] Step S134: Determine the medium-popularity product objects and low-popularity product objects, respectively, based on their popularity levels. Store the product identifiers and product feature information of each medium-popularity product object and each low-popularity product object into the medium-popularity medium-performance database and the low-popularity low-performance database in the medium-low popularity partitioned storage cluster, respectively.

[0150] For medium-popularity products, their product feature information will be stored in the medium-popularity, medium-performance database of the medium-low popularity partitioned storage cluster. For low-popularity products, their product feature information will be stored in the low-popularity, low-performance database of the medium-low popularity partitioned storage cluster.

[0151] In this embodiment, the product feature information is stored in a hierarchical manner based on the popularity of the product object. The product feature information of product objects with different popularity is stored in storage clusters with different performance levels to reasonably allocate the platform's data storage and search resources, improve the recall efficiency of product feature information of popular products, and ensure the rational use of data storage and search resources.

[0152] Please refer to Figure 6In one embodiment, detecting whether the corresponding product feature information of the product identifier exists in each of the target product feature nodes includes the following steps:

[0153] Step S141: Invoke the preset hash algorithm to calculate the hash value corresponding to the product identifier:

[0154] After obtaining the product identifier in the coarse recall information of the searched products and determining the target product feature node corresponding to the product identifier in the product feature node pool, the hash algorithm preset in the target product feature node will be called to determine the hash value corresponding to the product identifier. Then, the target product feature node will be queried to see if the product feature information corresponding to the product identifier is present through a Bloom filter.

[0155] Step S142: Obtain the product feature storage hash table of the target product feature node corresponding to the product identifier, and query the storage element corresponding to the hash value in the product feature storage hash table:

[0156] The product feature storage hash table consists of a binary bit array and a hash function. Each element in the binary bit array is represented by a 0 or a 1 to indicate whether data exists. 0 indicates that no data exists, and 1 indicates that data exists.

[0157] After obtaining the product feature storage hash table of the target product feature node, the hash value corresponding to the product identifier will be queried, and the corresponding element bit in the product feature storage hash table will be retrieved to obtain the storage element in that element bit.

[0158] Step S143: When the storage element is indicated to exist, it indicates that the target product feature node contains product feature information corresponding to the product identifier, and the product feature information corresponding to the storage element is obtained from the target product feature node.

[0159] When the storage element indicates the presence of data, i.e., when the storage element is 1, it indicates that the target product feature node contains product feature information corresponding to the product identifier, and the storage element will be retrieved from the target product feature node. If the storage element represents no data, i.e., the storage element is 0, then the product feature information corresponding to the product identifier will be obtained from the low-to-medium heat partition storage cluster.

[0160] In this embodiment, by setting a corresponding product feature storage hash table for each product feature node in the product feature node pool, a Bloom filter is constructed to query whether the product feature node has product feature information corresponding to the product identifier, thereby improving the query efficiency of querying whether product feature information exists in the product feature node.

[0161] Please refer to Figure 7 In one embodiment, when the target product feature node does not contain product feature information corresponding to the product identifier, after obtaining the product feature information corresponding to the product identifier from the medium-hot, medium-performance library or the low-hot, low-performance library in the medium-low-hot partitioned storage cluster, the following steps are included:

[0162] Step S151: When the target product feature node corresponding to the product identifier does not contain the corresponding product feature information, check whether the product identifier hits the product feature storage range of the medium-performance library in the medium-low heat partition storage cluster.

[0163] The medium-high performance library in the medium-low heat partitioned storage cluster has its corresponding product feature storage range, which represents the range of product identifiers of the product feature information stored therein. When a product identifier matches the product feature storage range of the medium-high performance library, it indicates that the medium-high performance library has product feature information corresponding to the product identifier.

[0164] Step S152: When the product identifier matches the product feature storage range, retrieve the product feature information corresponding to the product identifier from the medium-hot, medium-performance database; if not, retrieve the product feature information corresponding to the product identifier from the low-hot, low-performance database of the medium-low-hot partitioned storage cluster.

[0165] When a product identifier matches the product feature storage range of the medium-popularity, medium-performance library, the product feature information corresponding to the product identifier will be obtained from the medium-popularity, medium-performance library. If it does not match, the product feature information will be obtained from the low-popularity, low-performance library.

[0166] Step S153: Store the acquired product feature information into the target product feature node:

[0167] After obtaining the product feature information of the product identifier from the medium-popularity medium-performance library or the low-popularity low-performance library, the product feature information can be stored in the target product feature node corresponding to the product identifier.

[0168] In this embodiment, after obtaining the product feature information corresponding to the product identifier from the low-to-medium heat partition storage cluster, the product feature information can be stored in the target product feature node corresponding to the product identifier to improve the efficiency of recalling the product feature information in the next time.

[0169] Please refer to Figure 8 In one embodiment, the products are sorted according to the product feature information corresponding to each product identifier to generate a corresponding product sorting list, and the product sorting list is pushed to the corresponding user terminal, including the following steps:

[0170] Step S161: Obtain product feature information corresponding to multiple product identifiers. The product feature information includes basic product feature information and business product feature information. The basic product feature information includes product color, product size, or product brand. The business product feature information includes the target sales audience, sales volume, click-through rate, and purchase rate.

[0171] After retrieving the product feature information corresponding to each product identifier in the coarse recall information of the search products, the product objects corresponding to each product identifier will be sorted according to the product feature information. The product feature information includes the basic product feature information and the product business feature information of the product object. The basic product feature information includes information such as product color, product size, product brand, product price or product name, etc. The product business feature information includes information such as the target sales audience, product sales volume, product click rate, product purchase rate, product inventory and product exposure rate, etc., obtained through statistics or other algorithms.

[0172] Step S162: Obtain the product object to which each product feature information belongs, call the preset product sorting algorithm, sort the product objects according to the product feature information of each product object, and generate the corresponding product sorting list:

[0173] Obtain the product object corresponding to each product identifier, that is, obtain the product object to which the product feature information corresponding to the product identifier belongs, so as to call the preset product sorting algorithm, sort the product objects according to the product feature information of each product object, and generate the corresponding product sorting list.

[0174] When the coarse product recall information is generated when buyers search for products on e-commerce platforms, the user consumption preference information of the searching buyers will be obtained. Based on the user consumption preference information and the product feature information, the product objects with each product feature information will be sorted. Product objects that match the buyer's consumption preferences on the e-commerce platform will be sorted first, so that the buyer can browse product objects that match their consumption preferences more quickly after using the product search service, thus improving the buyer's online shopping experience on the e-commerce platform. Specifically, a product sorting model trained to convergence can be used to calculate the similarity between each product object and the user's consumption preference information based on the buyer's user consumption preference information and the product feature information of each searched product object. Product objects with higher similarity will be sorted first, thereby generating a corresponding product sorting list.

[0175] When the coarse product recall information is generated by merchants on e-commerce platforms when they search for products, it can sort the product objects corresponding to each product identifier based on the product sorting rules set by the merchants when using the product search service, according to the product feature information corresponding to each product identifier, and generate a corresponding product sorting list.

[0176] Step S163: Determine the user terminal corresponding to the coarse recall information of the searched products, and push the product sorting list to the user terminal:

[0177] When the preliminary product search results are generated when a buyer searches for products on an e-commerce platform, the resulting product ranking list will be pushed to the buyer's client. When the preliminary product search results are generated when a merchant searches for products on an e-commerce platform, the ranking list of the merchant's products will be pushed to the merchant's client.

[0178] In this embodiment, based on the two product search scenarios of buyer users and merchant users in the e-commerce platform, and according to the product feature information recalled for the product objects, each product object can be sorted to generate a corresponding product sorting list. For buyer users, product objects that match their consumption preferences can be sorted first, improving the online shopping experience for buyer users. For merchant users, product objects can be sorted according to their defined product sorting rules to facilitate the analysis of the sales performance of each product object in the store.

[0179] Please see Figure 9This product feature recall device, provided to meet one of the purposes of this application, is a functional embodiment of the product feature recall method of this application. On another aspect, this product feature recall device, provided to meet one of the purposes of this application, includes: a coarse product recall module 11, used to obtain coarse product recall information pushed by a product search service, and to obtain a merchant identifier and multiple product identifiers corresponding to the coarse product recall information; a dedicated cluster detection module 12, used to detect whether a merchant-specific product feature cluster corresponding to the merchant identifier exists, and if so, to obtain product feature information corresponding to each product identifier from the product feature nodes in the merchant-specific product feature cluster; and a target node hitting module 13, used to detect the range of product identifiers that each product identifier hits in the product feature node pool when the merchant-specific product feature cluster does not exist, and to determine the target node hitting range of each product identifier. The system includes: a target product feature node corresponding to the product identifier range; a high-popularity feature acquisition module 14, used to detect whether there is product feature information corresponding to the product identifier in each of the target product feature nodes; if so, to acquire the product feature information corresponding to the product identifier from the target product feature node; the product feature node stores the product feature information corresponding to high-popularity products; a cluster feature acquisition module 15, used to acquire the product feature information corresponding to the product identifier from the medium-popularity medium-performance library or the low-popularity low-performance library in the medium-low popularity partition storage cluster when there is no product feature information corresponding to the product identifier in the target product feature node; and a product feature sorting module 16, used to sort the products according to the product feature information corresponding to each product identifier, generate a corresponding product sorting list, and push the product sorting list to the corresponding user terminal.

[0180] In one embodiment, the coarse product recall module 11 includes: a product search request response submodule, used to respond to a product search request pushed by a user terminal, and obtain the product search terms and merchant identifiers corresponding to the product search request; a product object search submodule, used to obtain the product database corresponding to the merchant identifiers, and query one or more product objects corresponding to the product search terms from the product database; and a coarse recall information generation submodule, used to obtain the product identifiers corresponding to each product object, and generate coarse product recall information with each product identifier and each merchant identifier.

[0181] In one embodiment, the dedicated cluster detection module 12 includes: a store popularity information acquisition submodule, used to acquire store popularity information of multiple merchant users in the platform, the store popularity information including store product sales volume, store order volume, store visits or store exposure volume; a store popularity score determination submodule, used to call a merchant user popularity analysis algorithm to determine the store popularity score of each merchant user based on the store popularity information of each merchant user; and a dedicated cluster generation submodule, used to identify popular merchant users whose store popularity scores exceed the store popularity score threshold, generate a merchant-specific product feature cluster with multiple product feature nodes for the popular merchant users, and store the product feature information of the popular merchant users in each of the product feature nodes, wherein the merchant-specific product feature cluster is marked with the merchant identifier of the popular merchant user.

[0182] In one embodiment, the high-heat feature acquisition module 14 includes: a hash value calculation submodule, used to call a preset hash algorithm to calculate the hash value corresponding to the product identifier; a storage element query submodule, used to obtain the product feature storage hash table of the target product feature node corresponding to the product identifier, and query the storage element corresponding to the hash value in the product feature storage hash table; and a product effect information acquisition submodule, used to indicate that the product feature information corresponding to the product identifier exists in the target product feature node when the storage element is indicated to exist, and to obtain the product effect information corresponding to the storage element from the target product feature node.

[0183] In one embodiment, the product feature sorting module 16 includes: a product feature information acquisition submodule, used to acquire product feature information corresponding to multiple product identifiers, wherein the product feature information includes basic product feature information and product business feature information, wherein the basic product feature information includes product color, product size or product brand, and the product business feature information includes product target sales audience, product sales volume, product click-through rate and product purchase rate; a product sorting list generation submodule, used to acquire the product object to which each of the product feature information belongs, call a preset product sorting algorithm, sort each product object according to the product feature information of each product object, and generate a corresponding product sorting list; and a product sorting list push submodule, used to determine the user terminal corresponding to the coarse recall information of the searched products, and push the product sorting list to the user terminal.

[0184] To address the aforementioned technical problems, embodiments of this application also provide computer equipment. For example... Figure 10The diagram shows the internal structure of a computer device. The computer device includes a processor, a computer-readable storage medium, a memory, and a network interface connected via a system bus. The computer-readable storage medium stores an operating system, a database, and computer-readable instructions. The database may store a sequence of control information. When the computer-readable instructions are executed by the processor, they enable the processor to implement a product feature recall method. The processor provides computing and control capabilities to support the operation of the entire computer device. The memory stores computer-readable instructions, which, when executed by the processor, enable the processor to execute the product feature recall method of this application. The network interface of the computer device is used for communication with a terminal. Those skilled in the art will understand that… Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0185] In this embodiment, the processor is used to execute... Figure 9 The system contains the specific functions of each module and its sub-modules, and the memory stores the program code and various data required to execute these modules or sub-modules. The network interface is used for data transmission between the user terminal and the server. In this embodiment, the memory stores the program code and data required to execute all modules / sub-modules in the product feature recall device of this application, and the server can call the server's program code and data to execute the functions of all sub-modules.

[0186] This application also provides a storage medium storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the product feature recall method of any embodiment of this application.

[0187] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. This computer program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The aforementioned storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0188] In summary, this application stores product features in a tiered manner based on two dimensions: merchant popularity and product popularity, in order to improve the product feature retrieval rate.

[0189] Those skilled in the art will understand that the steps, measures, and solutions in the various operations, methods, and processes discussed in this application can be alternated, modified, combined, or deleted. Furthermore, other steps, measures, and solutions in the various operations, methods, and processes discussed in this application can also be alternated, modified, rearranged, decomposed, combined, or deleted. Furthermore, steps, measures, and solutions in the prior art that are similar to those in the open-source operations, methods, and processes of this application can also be alternated, modified, rearranged, decomposed, combined, or deleted.

[0190] The above description is only a partial embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A product feature recall method, characterized in that, Includes the following steps: Obtain coarse recall information of search products pushed by the product search service, and obtain the merchant identifier and multiple product identifiers corresponding to the coarse recall information of search products; Detect whether a merchant-specific product feature cluster corresponding to the merchant identifier exists. If it exists, obtain the product feature information corresponding to each product identifier from the product feature nodes in the merchant-specific product feature cluster. When the merchant-specific product feature cluster does not exist, the range of product identifiers that each product identifier hits in the product feature node pool is detected, and the target product feature node corresponding to each product identifier range is determined. Detect whether there is product feature information corresponding to the product identifier in each of the target product feature nodes. If there is, obtain the product feature information corresponding to the product identifier from the target product feature node. The product feature node stores the product feature information corresponding to the high-popularity product. If the target product feature node does not contain the product feature information corresponding to the product identifier, then the product feature information corresponding to the product identifier is obtained from the medium-hot, medium-performance library or the low-hot, low-performance library in the medium-low-hot partition storage cluster. The products are sorted according to the product feature information corresponding to each product identifier, a corresponding product sorting list is generated, and the product sorting list is pushed to the corresponding user terminal.

2. The product feature recall method according to claim 1, characterized in that, In response to the coarse product recall information pushed by the product search service, the following steps are taken to obtain the merchant identifier and multiple product identifiers corresponding to the coarse product recall information: Respond to product search requests pushed by the user and obtain the product search terms and merchant identifiers corresponding to the product search requests; Obtain the product database corresponding to the merchant identifier, and query one or more product objects corresponding to the product search term from the product database; Obtain the product identifier corresponding to each of the product objects, and generate coarse search product recall information with each of the product identifiers and each of the merchant identifiers.

3. The product feature recall method according to claim 1, characterized in that, Before detecting whether a merchant-specific product feature cluster corresponding to the merchant identifier exists, and if so, before obtaining the product feature information corresponding to each product identifier from the product feature nodes in the merchant-specific product feature cluster, the following steps are included: Obtain store popularity information from multiple merchant users on the platform. The store popularity information includes store product sales volume, store order volume, store visits, or store exposure. The merchant user popularity analysis algorithm is invoked to determine the store popularity score of each merchant user based on the store popularity information of each merchant user; Identify popular merchant users whose store popularity score exceeds the store popularity score threshold, generate a merchant-specific product feature cluster with multiple product feature nodes for the popular merchant users, and store the product feature information of the popular merchant users in each of the product feature nodes. The merchant-specific product feature cluster is marked with the merchant identifier of the popular merchant user.

4. The product feature recall method according to claim 1, characterized in that, The process involves detecting whether product feature information corresponding to the product identifier exists in each of the target product feature nodes. If it exists, the product feature information corresponding to the product identifier is obtained from the target product feature node. Alternatively, if the product feature information corresponding to the product identifier does not exist in the target product feature node, the process involves retrieving the product feature information from the medium- or low-heat database in the medium- or low-heat partition storage cluster. Before this, the process includes the following steps: Obtain product popularity information for multiple product objects on the platform, including product sales volume, product visits, or product click-through rate; The popular product analysis algorithm is invoked to determine the popularity level of each product object based on the product popularity information of each product object. The popularity level of the product is divided into high popularity products, medium popularity products and low popularity products. Identify high-popularity product objects with high popularity, obtain the product identifier and product feature information of each high-popularity product object, determine the target product feature node within the product identifier range hit in the product feature node pool for each product identifier, and store each product feature information into the target product feature node corresponding to its product identifier; The popularity of the products is determined to be medium-popularity products and low-popularity products, respectively. The product identifiers and product feature information of each medium-popularity product and each low-popularity product are stored in the medium-popularity medium-performance database and the low-popularity low-performance database in the medium-low popularity partitioned storage cluster, respectively.

5. The product feature recall method according to claim 1, characterized in that, Detecting whether the corresponding product feature information of the product identifier exists in each of the target product feature nodes includes the following steps: The preset hash algorithm is invoked to calculate the hash value corresponding to the product identifier; Obtain the product feature storage hash table of the target product feature node corresponding to the product identifier, and query the storage element corresponding to the hash value in the product feature storage hash table; When the storage element is indicated to exist, it indicates that the target product feature node contains product feature information corresponding to the product identifier, and the product feature information corresponding to the storage element is obtained from the target product feature node.

6. The product feature recall method according to claim 1, characterized in that, If the target product feature node does not contain product feature information corresponding to the product identifier, then after obtaining the product feature information corresponding to the product identifier from the medium-popularity medium-performance library or the low-popularity low-performance library in the medium-low popularity partition storage cluster, the following steps are included: When there is no corresponding product feature information in the target product feature node corresponding to the product identifier, check whether the product identifier hits the product feature storage range of the medium-hot, medium-performance library in the medium-low hotness partition storage cluster. When the product identifier matches the product feature storage range, the product feature information corresponding to the product identifier is obtained from the medium-hot, medium-performance library. If it does not match, the product feature information corresponding to the product identifier is obtained from the low-hot, low-performance library of the medium-low-hot partition storage cluster. The acquired product feature information is stored in the target product feature node.

7. The product feature recall method according to claim 1, characterized in that, The products are sorted according to the product feature information corresponding to each product identifier, a corresponding product sorting list is generated, and the product sorting list is pushed to the corresponding user terminal, including the following steps: Obtain product feature information corresponding to multiple product identifiers. The product feature information includes basic product feature information and product business feature information. The basic product feature information includes product color, product size or product brand. The product business feature information includes the target sales audience, product sales volume, product click-through rate and product purchase rate. Obtain the product object to which each product feature information belongs, call the preset product sorting algorithm, sort each product object according to the product feature information of each product object, and generate the corresponding product sorting list; Once the user terminal corresponding to the coarse recall information of the searched products is identified, the product sorting list is pushed to the user terminal.

8. A product feature recall device, characterized in that, include: The product coarse recall module is used to obtain the search product coarse recall information pushed by the product search service, and to obtain the merchant identifier and multiple product identifiers corresponding to the search product coarse recall information. The dedicated cluster detection module is used to detect whether there is a merchant-specific product feature cluster corresponding to the merchant identifier. If it exists, the module obtains the product feature information corresponding to each product identifier from the product feature nodes in the merchant-specific product feature cluster. The target node hitting module is used to detect the range of product identifiers that each product identifier hits in the product feature node pool when the merchant-specific product feature cluster does not exist, and to determine the target product feature node corresponding to each range of product identifiers. The high-temperature feature acquisition module is used to detect whether there is product feature information of the corresponding product identifier in each of the target product feature nodes. If there is, the product feature information corresponding to the product identifier is obtained from the target product feature node. The product feature node stores the product feature information corresponding to the high-temperature product. The cluster feature acquisition module is used to acquire the product feature information corresponding to the product identifier from the medium-hot, medium-performance library or the low-hot, low-performance library in the medium-low-hot partition storage cluster when the product feature information corresponding to the product identifier does not exist in the target product feature node. The product feature sorting module is used to sort products according to the product feature information corresponding to each product identifier, generate a corresponding product sorting list, and push the product sorting list to the corresponding user terminal.

9. A computer device comprising a central processing unit and a memory, characterized in that, The central processing unit is used to invoke and run a computer program stored in the memory to perform the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores, in the form of computer-readable instructions, a computer program implemented according to any one of claims 1 to 7, which, when invoked by a computer, executes the steps included in the corresponding method.

Images