Label generation method and related device

Abstract

The embodiment of the invention provides a label generation method and a related device, and relates to the technical field of big data, and the method comprises the steps: obtaining behavior data of a user and attribute data of a commodity; the behavior data comprises historical behavior data and real-time behavior data; obtaining a user portrait label based on the historical behavior data and the real-time behavior data; and obtaining a commodity portrait label based on the attribute data and the real-time behavior data. According to the scheme, the offline data and the real-time behavior data are coordinated, the instant interest of the user and the commodity popularity can be captured in time and the label is generated, the response speed of label updating is effectively improved, different algorithms are adopted for different data to generate the label, and the accuracy of label generation can be effectively improved.

Description

Technical Field

[0001] This application relates to the field of big data technology, and in particular to a tag generation method and related apparatus. Background Technology

[0002] In the field of internet services, the construction of tags used to describe user profiles and product profiles is the core foundation for realizing personalized services.

[0003] Currently, user and product profile tags are mostly generated from batches of offline data. However, because user interests and product popularity can change drastically in a short period of time, the tags generated by the above methods suffer from update delays, cannot respond to instantaneous needs, and the generated user and product profile tags lack accuracy. Summary of the Invention

[0004] This application provides a label generation method and related apparatus to improve the speed of label updates and the accuracy of labels.

[0005] In a first aspect, embodiments of this application provide a label generation method, including:

[0006] Acquire user behavior data and product attribute data; the behavior data includes historical behavior data and real-time behavior data;

[0007] Based on the historical behavior data and the real-time behavior data, user profile tags are obtained;

[0008] Based on the attribute data and the real-time behavior data, product profile tags are obtained.

[0009] In some embodiments, obtaining user profile tags based on the historical behavior data and the real-time behavior data includes:

[0010] Based on the historical behavior data, a first user profile tag is obtained, and based on the real-time behavior data, a second user profile tag is obtained;

[0011] Based on the user identifier, the first user profile tag and the second user profile tag are associated to determine the user profile tag.

[0012] In some embodiments, obtaining product profile tags based on the attribute data and the real-time behavior data includes:

[0013] Based on the attribute data, a first product profile tag is obtained, and based on the attribute data and the real-time behavior data, a second product profile tag is obtained;

[0014] Based on the product identifier, the first product profile tag and the second product profile tag are associated to determine the product profile tag.

[0015] In some embodiments, obtaining a first user profile tag based on the historical behavior data and obtaining a second user profile tag based on the real-time behavior data includes:

[0016] The historical behavior data and the real-time behavior data are preprocessed;

[0017] The preprocessed historical behavior data and real-time behavior data are correlated and fused to obtain user data features;

[0018] The user data features are processed using a first model to obtain a first user profile label; the first model includes multiple sub-models, with different sub-models used for different types of data;

[0019] The real-time behavioral features in the user data are processed using a second model to obtain a second user profile label; the complexity of the first model is higher than that of the second model.

[0020] In some embodiments, obtaining a first product profile tag based on the attribute data, and obtaining a second product profile tag based on the attribute data and the real-time behavior data, includes:

[0021] After preprocessing the basic product data in the attribute data, the third model is used to process the preprocessed basic product data to obtain the first product profile tag.

[0022] Within a preset time window, based on the real-time behavior data, real-time popularity data of the product is obtained, and real-time popularity profile tags are determined based on the real-time popularity data.

[0023] Based on the historical popularity data in the attribute data, determine the historical popularity profile tags of the product;

[0024] The second product profile tag is determined based on the real-time popularity profile tag and the historical popularity profile tag.

[0025] In some embodiments, the method further includes:

[0026] Determine the temperature attribute of each sub-tag in the user profile tag and / or the product profile tag; the temperature attribute is used to indicate the popularity of the tag.

[0027] The storage strategy corresponding to the sub-tag is obtained based on the temperature attribute;

[0028] The sub-tags are stored in the corresponding database based on the storage strategy.

[0029] In some embodiments, the method further includes:

[0030] In response to a query request, a list of user identifiers corresponding to the query request is determined based on the user profile tags; the query request includes at least one query tag.

[0031] Based on the product profile tags, a product recommendation list is generated for each user in the user identifier list.

[0032] In some embodiments, the method further includes:

[0033] Based on the user profile tags and the product profile tags, construct the push strategy graph corresponding to the user;

[0034] In response to a target event, products are pushed to the user based on the push strategy graph.

[0035] Secondly, embodiments of this application provide a label generation apparatus, comprising:

[0036] The acquisition module is used to acquire user behavior data and product attribute data; the behavior data includes historical behavior data and real-time behavior data.

[0037] The first processing module is used to obtain user profile tags based on the historical behavior data and the real-time behavior data;

[0038] The second processing module is used to obtain product profile tags based on the attribute data and the real-time behavior data.

[0039] Thirdly, this application provides an electronic device, including: a memory and a processor;

[0040] The memory is used to store computer instructions; the processor is used to execute the computer instructions stored in the memory to implement the method of any one of the first aspects.

[0041] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the method of any of the first aspects.

[0042] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the method of any one of the first aspects.

[0043] The tag generation method and related apparatus provided in this application acquire user behavior data and product attribute data; the behavior data includes historical behavior data and real-time behavior data; user profile tags are obtained based on the historical behavior data and the real-time behavior data; and product profile tags are obtained based on the attribute data and the real-time behavior data. This solution, by coordinating offline data with real-time behavior data, allows for the instantaneous capture and tag generation of user instantaneous interests and product popularity, effectively improving the response speed of tag updates. Furthermore, employing different algorithms for tag generation based on different data effectively improves the accuracy of the generated tags. This results in generated tags that better match user needs, effectively enhancing the user experience. Attached Figure Description

[0044] Figure 1 A scenario diagram provided for an embodiment of this application;

[0045] Figure 2 A flowchart illustrating a label generation method provided in this application embodiment. Figure 1 ;

[0046] Figure 3 A flowchart illustrating a label generation method provided in this application embodiment. Figure 2 ;

[0047] Figure 4 This is a schematic diagram of a label generation device provided in an embodiment of this application.

[0048] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0050] In the embodiments of this application, the terms "first" and "second" are used to distinguish identical or similar items with essentially the same function and effect, without limiting their order. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that the terms "first" and "second" do not necessarily imply that they are different.

[0051] It should be noted that, in the embodiments of this application, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0052] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with the relevant laws, regulations, and standards of the relevant countries and regions, have taken necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation access points for users to choose to authorize or refuse.

[0053] Furthermore, the technical solution involved in this application, which involves big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.) and the use of artificial intelligence technology for automated decision-making, and makes decisions that have a significant impact on personal rights based on the results of automated decision-making, provides users with corresponding operation entry points for users to choose to agree to or reject the results of automated decision-making; if the user chooses to reject, the process will proceed to the expert decision-making process.

[0054] In the fields of e-commerce, content recommendation, online advertising, and other internet services, the construction of user profiles and product profiles is the core foundation for realizing personalized services.

[0055] Taking e-commerce platforms as an example, user behavior data (such as clicks, adding to cart, placing orders, and reviews) and product attribute data (such as category, price, and description text) generated when browsing, searching, and purchasing products need to be processed in real time or offline to generate dynamic user profile tags and product profile tags. These tags are used for precision marketing, personalized recommendations, product category management, and user lifecycle management.

[0056] In related technologies, user profile tags and product profile tags rely on T+1 offline data for batch calculations, which cannot capture real-time changes in user interests, resulting in delayed tag updates and an inability to respond to immediate needs. Furthermore, tags are typically generated using single algorithms such as collaborative filtering, which are insufficient to handle diverse business scenarios, leading to inadequate tag accuracy.

[0057] To address the aforementioned issues, this application provides a tag generation method and related apparatus. By coordinating offline data with real-time behavioral data and employing different algorithms for different data to generate tags, the accuracy of tag generation and the response speed of tag updates can be effectively improved, thereby matching user needs and enhancing the user experience.

[0058] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0059] Figure 1 This is a schematic diagram of a scenario provided for an embodiment of this application, such as... Figure 1 As shown, it includes: terminal device 1 and server 2. Terminal device 1 can communicate with server 2 via a wired or wireless communication link.

[0060] Terminal device 1 can have multiple different client applications installed, such as financial service applications, shopping applications, web browser applications, instant messaging tools, email clients, social platform software, etc.

[0061] Server 2 can be a server that provides various services. For example, it can provide background management and support services for converged service applications on user terminal devices. It should be understood that when Server 2 provides background management and support services for different client applications on terminal device 1, the data used and generated is stored in the business database corresponding to each client application.

[0062] The terminal devices provided in this application embodiment may include smartphones, tablets, laptops, and desktop computers, etc. The server may be a local server, a cloud server, or a server cluster. This application embodiment does not limit the types of terminal devices and servers.

[0063] Below, in Figure 1 Based on the illustrated embodiment, with Figure 1 The server shown is the execution entity, and the tag generation method provided in the embodiments of this application will be described.

[0064] Figure 2 This is a flowchart illustrating a label generation method provided in an embodiment of this application, as shown below. Figure 2 As shown, it includes:

[0065] S201. Obtain user behavior data and product attribute data.

[0066] In some embodiments, user behavior data refers to user behavior data when using a client application. For example, data such as a user's searches, clicks, adding items to cart, placing orders, and reviews when using a shopping application.

[0067] User behavior data can include historical behavior data and real-time behavior data. Real-time behavior data refers to the user's behavior while currently using the client application, while historical behavior data refers to the user's behavior when using the client application in the past.

[0068] Product attribute data can refer to data used to describe a product, such as the product's price, category, description information, sales-related data, etc.

[0069] In some embodiments, historical user behavior data can be obtained from the business database corresponding to the client application.

[0070] In some embodiments, real-time user behavior data can be obtained from a client application. For example, the client may have data tracking points set up to collect user operation time, which can be used to collect real-time user behavior data.

[0071] It should be understood that user behavior data may include user behavior data corresponding to the use of multiple client applications.

[0072] In some embodiments, product attribute data can be obtained from the business database corresponding to the product.

[0073] S202. Obtain user profile tags based on historical and real-time behavioral data.

[0074] In some embodiments, user profile tags may include long-term tags and real-time tags.

[0075] For example, a first user profile tag can be obtained based on historical behavior data, and a second user profile tag can be obtained based on real-time behavior data; the first user profile tag and the second user profile tag can be associated based on the user identifier to determine the user profile tag.

[0076] The first user profile tag can refer to a user's long-term tags, such as demographic attribute tags, long-term interest preference tags, and purchasing power tags. The second user profile tag can refer to a user's real-time tags, such as instantaneous intent tags.

[0077] In some embodiments, historical behavior data can be processed based on a pre-trained deep learning model or a preset rule set to obtain the user's first profile label.

[0078] Alternatively, data computing engines (such as Spark, Flink, etc.) can be used to process historical data, generate user behavior summaries, construct guiding words, and use a large language model to process the behavior summaries to obtain the first user profile label.

[0079] In some embodiments, the user's real-time behavior data can be processed based on pre-trained lightweight rules or models (such as Transformer) to obtain a second user profile label.

[0080] In some embodiments, after obtaining the first user profile tag and the second user profile tag, the first user profile tag and the second user profile tag can be associated based on the user's identifier to obtain a user profile tag. For example, the first user profile tag and the second user profile tag can be stored as attributes with the user ID as the primary key.

[0081] S203. Obtain product profile tags based on attribute data and real-time behavior data.

[0082] In some embodiments, product profile tags may include basic attribute tags and popularity tags.

[0083] For example, a first product profile tag is obtained based on attribute data, and a second product profile tag is obtained based on attribute data and real-time behavior data; the first product profile tag and the second product profile tag are associated based on the product identifier to determine the product profile tag.

[0084] Among them, the first product profile tag can refer to the basic attribute tag of the product, and the second product profile tag can refer to the popularity tag of the product.

[0085] For example, attribute data can be processed based on a pre-trained deep learning model or a preset rule set to obtain the first product profile label.

[0086] Alternatively, data computing engines (such as Spark, Flink, etc.) can be used to process the attribute data, generate attribute summaries, construct guide words, and use a large language model to process the attribute summaries to obtain the first product profile label.

[0087] For example, sales-related data (such as sales volume and add-to-cart volume) in attribute data can be analyzed to obtain historical popularity data for products. Real-time user behavior data (such as pageviews) can be used to calculate real-time popularity data for products. A pre-trained popularity analysis model based on the dotted lines can then be used to process the historical and real-time popularity data to generate a second product profile tag.

[0088] In some embodiments, after obtaining the first product profile tag and the second product profile tag, the first product profile tag and the second product profile tag can be associated based on the product's identifier to obtain a product profile tag. For example, the first product profile tag and the second product profile tag can be stored as attributes, with the product ID as the primary key.

[0089] It should be understood that when generating user profile tags, user behavioral data can refer to the user's own behavioral data. When generating product image tags, user behavioral data can refer to the individual user's behavioral data related to that product.

[0090] Optionally, once user profile tags and product profile tags are obtained, personalized product recommendations can be made to users.

[0091] Optionally, upon obtaining product profile tags, the product database can be dynamically categorized and managed in a multi-dimensional and cross-cutting manner based on these tags, going beyond fixed administrative categories and enhancing the flexibility of product display.

[0092] The tag generation method provided in this application obtains user behavior data and product attribute data. The behavior data includes historical behavior data and real-time behavior data. Based on the historical and real-time behavior data, user profile tags are obtained; based on the attribute data and real-time behavior data, product profile tags are obtained. This solution, by coordinating offline data with real-time behavior data, allows for the instantaneous capture of user interests and product popularity to generate tags, effectively improving the response speed of tag updates. Furthermore, by employing different algorithms for tag generation based on different data, the accuracy of the generated tags can be effectively improved. This results in generated tags that better match user needs, effectively enhancing the user experience.

[0093] Below, based on the above embodiments, and in conjunction with Figure 3 The label generation method provided in the embodiments of this application will be further described.

[0094] Figure 3 A flowchart illustrating a label generation method provided in this application embodiment. Figure 2 ,like Figure 3 As shown, it includes:

[0095] S301. Obtain user behavior data and product attribute data.

[0096] The specific implementation method shown in S301 of this application embodiment is the same as Figure 2 The specific implementation of S201 in the illustrated embodiment is similar and will not be repeated here.

[0097] S302. Preprocess historical behavior data and real-time behavior data; and then correlate and fuse the preprocessed historical behavior data and real-time behavior data to obtain user data features.

[0098] In some embodiments, preprocessing includes data cleaning, normalization, and other operations. For example, removing duplicate, erroneous, and invalid data from historical and real-time behavioral data. Standardizing the format of data from different sources, such as unifying timestamp formats and currency units.

[0099] In some embodiments, after data preprocessing, historical behavior data and real-time behavior data can be associated based on the user's identifier (such as user ID) and vectorized to obtain user data features.

[0100] Optionally, after obtaining user data features, redundant features can be removed from the data features to improve the efficiency of subsequent processing. For example, correlation analysis and recursive feature elimination methods can be used to filter features.

[0101] S303. Process the user data features to obtain the first user profile label and the second user profile label, and associate the first user profile label and the second user profile label to obtain the user profile label.

[0102] In some embodiments, multiple different pre-trained models can be used to process user data features to obtain a first user profile label and a second user profile label.

[0103] For example, the first model is used to process user data features to obtain a first user profile label; the first model includes multiple sub-models, and different sub-models are used for different types of data; the second model is used to process real-time behavioral features in user data features to obtain a second user profile label; the complexity of the first model is higher than that of the second model.

[0104] For example, the first model may include collaborative filtering model, matrix factorization model, topic model, clustering model, statistical or machine learning model, etc.

[0105] Collaborative filtering and matrix factorization models can be used to process user data features, analyze long-term user behavior sequences, and generate long-term user interest and preference tags. Topic models (such as LDA models) can be used to process user data features to obtain basic user attribute tags (such as age distribution). Clustering models can be used to process user data features to obtain user grouping tags. Statistical or machine learning models can be used to process user data features to generate user purchasing power tags, etc. It should be understood that a user's first profile tag includes all of the above tags, also known as sub-tags.

[0106] Optionally, to further improve the model's processing efficiency, multiple different feature sets can be extracted and constructed from user data features, with each feature set corresponding to a different model. When generating the first profile label, different feature sets can be input into the corresponding models, thereby reducing the amount of data processed by each model and improving the efficiency of the model in generating labels.

[0107] For example, the second model can be a time-based deep learning model (such as GRU, Transformer, etc.).

[0108] Real-time behavioral features from user data can be input into a time-series-based deep learning model to obtain a second profile label.

[0109] In some embodiments, the generation of the first profile tag and the second profile tag can be automated through scheduled tasks. For example, the generation of the first profile tag can be performed at a fixed period (e.g., every morning), using data up to the previous day. The generation of the second profile tag can be triggered based on real-time user behavior data (e.g., clicks on searches), using real-time streaming data.

[0110] In some embodiments, the process of associating the first user profile tag and the second user profile tag to obtain the user profile tag can refer to the process in the above embodiments, and will not be repeated here.

[0111] S304. Process the attribute data and the user's real-time behavior data to obtain the first product profile tag and the second product profile tag, and associate the first product profile tag and the second product profile tag to obtain the product profile tag.

[0112] In some embodiments, different models or algorithms can be used to process attribute data and real-time user behavior data to obtain a first product profile tag and a second product profile tag.

[0113] For example, after preprocessing the basic product data in the attribute data, the third model is used to process the preprocessed basic product data to obtain the first product profile label.

[0114] For example, attribute data can be processed using topic models (such as LDA models) to obtain basic attribute tags for products; clustering models can be used to process attribute data to obtain cluster tags for products; and statistical or machine learning models can be used to process attribute data to generate characteristic tags for products (such as affordability). It should be understood that the first profile tag for a product includes all of the above tags, also known as sub-tags.

[0115] Optionally, to further improve the processing efficiency of the model, multiple different feature sets can be extracted and constructed from the product attribute data, with each feature set corresponding to a different model.

[0116] Alternatively, the models mentioned above can be updated and trained based on newly accumulated data to improve their accuracy.

[0117] For example, within a preset time window, real-time popularity data of a product is obtained based on real-time behavioral data, and real-time popularity profile tags are determined based on the real-time popularity data; historical popularity profile tags of a product are determined based on historical popularity data in attribute data; and a second product profile tag is determined based on the real-time popularity profile tag and the historical popularity profile tag.

[0118] For example, define a 5-minute scrolling window, aggregate by product ID based on real-time behavioral data, and calculate the total number of views and add-to-cart entries for that product within that window. Define a 1-hour sliding window (sliding step of 5 minutes), aggregate by product ID based on real-time behavioral data, and calculate its sales volume. This yields the product's real-time popularity data.

[0119] The popularity data is converted into a corresponding popularity score based on a preset mapping relationship. Then, based on the correspondence between the popularity score and the preset popularity profile tags, the real-time popularity tag for the product is obtained. For example, the top 1% of products in terms of sales in the past hour are tagged with "Real-time Hot Item - Extremely High Popularity".

[0120] For example, historical popularity data in attribute data can refer to data such as the cumulative sales volume of a product over the past 30 days and the average daily sales revenue over the past 7 days. The historical popularity data of any product can be ranked within the popularity data of products in the same dimension to obtain the corresponding historical popularity tag, and / or, the month-on-month comparison data of the product's popularity data can be calculated to obtain the corresponding historical popularity tag for that product.

[0121] Real-time popularity data of a product can be linked and merged with historical popularity data to obtain a second product profile tag.

[0122] In some embodiments, the process of associating the first product image tag and the second product image tag to obtain the product image tag can refer to the process in the above embodiments, and will not be repeated here.

[0123] S305. Determine the temperature attribute of each sub-tag in the user profile tag and / or product profile tag, and store the user profile tag and / or product profile tag based on the temperature attribute.

[0124] In some embodiments, the temperature attribute of each sub-label is used to indicate the heat level at which the label is accessed or the frequency of its use.

[0125] For example, the temperature attribute of each sub-tag in the user profile tag and / or product profile tag is determined; the storage strategy corresponding to the sub-tag is obtained based on the temperature attribute; and the sub-tag is stored in the corresponding database based on the storage strategy.

[0126] For example, the temperature attribute of each sub-tag can be divided into hot tag, warm tag, and cold tag.

[0127] In some embodiments, temperature matching can be performed on each sub-label based on a predefined set of rules (based on prior knowledge) to obtain the temperature attribute of each sub-label.

[0128] In some embodiments, a preset label temperature prediction model can be trained based on big data, and the temperature of each sub-label can be predicted based on the trained model to obtain the temperature attribute of each sub-label.

[0129] In some embodiments, tags with different temperature attributes have different storage strategies. For example, hot tags are stored in databases such as Redis to achieve millisecond-level response; warm tags are stored in databases such as HBase or ClickHouse, which can be quickly queried by primary key; and cold tags are stored in databases such as HDFS or S3, which can be used for batch analysis via Spark or Hive.

[0130] Alternatively, all tags, such as user profile tags and product profile tags, can be stored in data such as Hive or ClickHouse, while real-time behavioral data tags, such as second user profile tags and second product profile tags, can be stored in key-value databases such as Redis or HBase.

[0131] By storing tags in different databases, a good balance can be achieved between cost, performance, and functionality.

[0132] S306. Make personalized recommendations to users based on user profile tags and product profile tags.

[0133] In some embodiments, personalized recommendations can be made based on requests from the front end.

[0134] For example, in response to a query request, a list of user identifiers corresponding to the query request is determined based on user profile tags; the query request includes at least one query tag; and a list of recommended products corresponding to each user in the user identifier list is generated based on product profile tags.

[0135] For example, business users can use AND, OR, and NOT logic on the front end to combine query tags such as "female," "browsed luxury goods in the last 7 days," and "high spending power." The system backend can then quickly retrieve user profile tags based on these tags and return a list of users matching the query tags. Once the user list is obtained, a product recommendation list that the user might like can be generated based on the product profile tags.

[0136] In some embodiments, personalized recommendations can be made automatically.

[0137] For example, a push strategy graph corresponding to a user is constructed based on user profile tags and product profile tags; in response to a target event, products are pushed to the user based on the push strategy graph.

[0138] The target event could be when a user opens an application or performs a product search.

[0139] For example, user tags and product tags are mapped to a unified semantic space and converted into numerical vectors; using users, products, and tags as nodes, three types of key edges are established: user-tag edges (representing user attribute preferences with weights), product-tag edges (representing product feature attribution), and tag-tag edges (representing semantic associations), generating a push strategy graph. The graph is stored in a graph database (such as Neo4j).

[0140] When the target event occurs, the system extracts relevant subgraphs from the graph: locates user nodes and associated tags; finds matching product nodes along the tag edges; calculates the matching score (based on path length, edge weight, and tag overlap); generates a push list by sorting the scores, and pushes products accordingly.

[0141] In summary, the tag generation method provided in this application, through multi-source data fusion, dynamic algorithm adaptation, and automated task scheduling, solves the problems of tag lag, algorithm rigidity, inefficient processes, and rigid applications in the prior art, providing an efficient and accurate profile management solution for the Internet service field. Specifically, through offline and real-time task collaboration, users' instantaneous interests and product popularity can be captured instantly and tags generated, effectively improving the real-time performance of tag updates; dynamically selecting algorithms based on data types avoids the performance bottleneck of a single algorithm, improving tag accuracy; from data access, feature extraction, model training to tag deployment, everything is completed automatically by the system, reducing manual intervention and improving operational efficiency; through hierarchical storage and logical combination interfaces, business users can quickly achieve complex user group selection and dynamic product classification, directly driving precision marketing and personalized recommendations.

[0142] Based on the above embodiments, this application also provides a label generation device.

[0143] Figure 4 This is a schematic diagram of the structure of the label generation apparatus 40 provided in the embodiments of this application, as shown below. Figure 4 As shown, it includes:

[0144] The acquisition module 401 is used to acquire user behavior data and product attribute data; the behavior data includes historical behavior data and real-time behavior data.

[0145] The first processing module 402 is used to obtain user profile tags based on historical behavior data and real-time behavior data.

[0146] The second processing module 403 is used to obtain product profile tags based on attribute data and real-time behavior data.

[0147] In some embodiments, the first processing module 402 is configured to obtain a first user profile tag based on historical behavior data, and obtain a second user profile tag based on real-time behavior data; and associate the first user profile tag and the second user profile tag based on the user identifier to determine the user profile tag.

[0148] In some embodiments, the second processing module 403 is used to obtain a first product profile tag based on attribute data, and to obtain a second product profile tag based on attribute data and real-time behavior data; and to associate the first product profile tag and the second product profile tag based on the product identifier to determine the product profile tag.

[0149] In some embodiments, the first processing module 402 is used to preprocess historical behavior data and real-time behavior data; to correlate and fuse the preprocessed historical behavior data and real-time behavior data to obtain user data features; to process the user data features using a first model to obtain a first user profile label; the first model includes multiple sub-models, with different sub-models used for different types of data; to process the real-time behavior features in the user data features using a second model to obtain a second user profile label; the complexity of the first model is higher than that of the second model.

[0150] In some embodiments, the second processing module 403 is used to preprocess the basic product data in the attribute data, and then use a third model to process the preprocessed basic product data to obtain a first product profile label; within a preset time window, based on real-time behavior data, obtain real-time popularity data of the product, and determine a real-time popularity profile label based on the real-time popularity data; based on historical popularity data in the attribute data, determine a historical popularity profile label of the product; and based on the real-time popularity profile label and the historical popularity profile label, determine a second product profile label.

[0151] In some embodiments, the second processing module 403 is configured to determine the temperature attribute of each sub-tag in the user profile tag and / or product profile tag; the temperature attribute is used to indicate the popularity of the tag; obtain the storage strategy corresponding to the sub-tag based on the temperature attribute; and store the sub-tag in the corresponding database based on the storage strategy.

[0152] In some embodiments, the label generation device 40 further includes a recommendation module 404.

[0153] The recommendation module 404 is used to respond to a query request by determining a list of user identifiers that meet the query request based on user profile tags; the query request includes at least one query tag; and generates a list of recommended products for each user in the user identifier list based on product profile tags.

[0154] In some embodiments, the recommendation module 404 is used to construct a push strategy graph corresponding to the user based on user profile tags and product profile tags; and to push products to the user based on the push strategy graph in response to a target event.

[0155] The label generation apparatus provided in this application embodiment can execute the label generation method shown in any of the above embodiments, and its principle and technical effect are similar, so it will not be described again here.

[0156] This application also provides an electronic device.

[0157] Figure 5 This is a schematic diagram of the structure of the electronic device 50 provided in the embodiments of this application, such as... Figure 5 As shown, the electronic device may include: transceiver 501, processor 502, and memory 503.

[0158] Processor 502 executes computer execution instructions stored in memory, causing processor 502 to perform the scheme in the above embodiments. Processor 502 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0159] The memory 503 is connected to the processor 502 via the system bus and completes communication between them. The memory 503 is used to store computer program instructions.

[0160] Transceiver 501 can perform the functions of receiving and sending data and instructions.

[0161] Optionally, the electronic device 50 may also include a communication interface 504, which allows communication and interaction with external or internal devices via the communication interface 503. External devices may be, for example, client devices (e.g., mobile phones, tablets). In specific implementations, if the communication interface 504, memory 503, and processor 502 are implemented independently, they can be interconnected via a bus to complete communication with each other.

[0162] The system bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The system bus can be divided into address bus, data bus, control bus, etc. For ease of representation, only one thick line is used in the diagram, but this does not indicate that there is only one bus or one type of bus. Transceivers are used to enable communication between database access devices and other computers (e.g., clients, read-write libraries, and read-only libraries). Memory may include random access memory (RAM) and may also include non-volatile memory.

[0163] Optionally, in a specific implementation, if the communication interface 504, memory 503, and processor 502 are integrated on a single chip, then the communication interface 504, memory 503, and processor 502 can communicate through an internal interface.

[0164] This application also provides a chip for executing instructions, which is used to execute the technical solutions of the methods described in the above embodiments.

[0165] This application also provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the technical solution of the above method embodiment. Its implementation principle and technical effect are similar, and will not be repeated here.

[0166] In one possible implementation, a computer-readable medium may include random access memory (RAM), read-only memory (ROM), compact discread-only memory (CD-ROM) or other optical disc storage, disk storage or other magnetic storage devices, or any other medium targeted to carry or to store the required program code in the form of instructions or data structures, and accessible by a computer. Furthermore, any connection is appropriately referred to as a computer-readable medium. For example, if software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL), or wireless technologies such as infrared, radio, and microwave, then coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of medium. As used herein, disks and optical discs include optical discs, laser discs, optical discs, Digital Versatile Discs (DVDs), floppy disks, and Blu-ray discs, where disks typically reproduce data magnetically, while optical discs optically reproduce data using lasers. Combinations of the above should also be included within the scope of computer-readable media.

[0167] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the technical solution of the above method embodiments. Its implementation principle and technical effects are similar, and will not be repeated here.

[0168] In the specific implementation of the aforementioned terminal device or server, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this application can be directly manifested as execution by a hardware processor, or execution by a combination of hardware and software modules within the processor.

[0169] Those skilled in the art will understand that all or part of the steps in any of the above method embodiments can be implemented by hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium, and when the program is executed, all or part of the steps in the above method embodiments are performed.

[0170] If the technical solution of this application is implemented in software form and sold or used as a product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the technical solution of this application can be embodied in the form of a software product, which is stored in a storage medium and includes a computer program or several instructions. This computer software product enables a computer device (which may be a personal computer, server, network device, or similar electronic device) to execute all or part of the steps of the methods in the embodiments of this application.

[0171] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0172] It should be further noted that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0173] It should be understood that the above-described device embodiments are merely illustrative, and the device of this application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units, modules, or components may be combined, or integrated into another system, or some features may be ignored or not executed.

[0174] Furthermore, unless otherwise specified, the functional units / modules in the various embodiments of this application can be integrated into one unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated together. The integrated units / modules described above can be implemented in hardware or as software program modules.

[0175] When integrated units / modules are implemented in hardware, the hardware can be digital circuits, analog circuits, etc. The physical implementation of the hardware structure includes, but is not limited to, transistors, memristors, etc. Unless otherwise specified, the processor can be any suitable hardware processor, such as a CPU, GPU, FPGA, DSP, and ASIC, etc. Unless otherwise specified, the storage unit can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc.

[0176] If the integrated unit / module is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0177] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0178] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A label generation method characterized by, The method comprises: acquiring behavior data of a user and attribute data of a commodity; the behavior data comprises historical behavior data and real-time behavior data; acquiring a user portrait label based on the historical behavior data and the real-time behavior data; acquiring a commodity portrait label based on the attribute data and the real-time behavior data.

2. The method of claim 1, wherein, The acquiring of the user portrait label based on the historical behavior data and the real-time behavior data comprises: acquiring a first user portrait label based on the historical behavior data, and acquiring a second user portrait label based on the real-time behavior data; associating the first user portrait label and the second user portrait label based on a user identifier to determine the user portrait label.

3. The method of claim 1, wherein, The acquiring of the commodity portrait label based on the attribute data and the real-time behavior data comprises: acquiring a first commodity portrait label based on the attribute data, and acquiring a second commodity portrait label based on the attribute data and the real-time behavior data; associating the first commodity portrait label and the second commodity portrait label based on a commodity identifier to determine the commodity portrait label.

4. The method of claim 2, wherein, The acquiring of the first user portrait label based on the historical behavior data and the acquiring of the second user portrait label based on the real-time behavior data comprise: preprocessing the historical behavior data and the real-time behavior data; associating and fusing the preprocessed historical behavior data and real-time behavior data to obtain user data features; processing the user data features by using a first model to obtain a first user portrait label; the first model comprises a plurality of sub-models, and different sub-models are used for data of different types; processing real-time behavior features in the user data features by using a second model to obtain a second user portrait label; the complexity of the first model is higher than that of the second model.

5. The method of claim 3, wherein, The acquiring of the first commodity portrait label based on the attribute data and the acquiring of the second commodity portrait label based on the attribute data and the real-time behavior data comprise: preprocessing commodity basic data in the attribute data, and processing the preprocessed commodity basic data by using a third model to obtain the first commodity portrait label; acquiring real-time heat data of a commodity based on the real-time behavior data within a preset time window, and determining a real-time heat portrait label based on the real-time heat data; determining a historical heat portrait label of the commodity based on historical heat data in the attribute data; determining the second commodity portrait label based on the real-time heat portrait label and the historical heat portrait label.

6. The method according to any one of claims 1 to 5, characterized in that, The method further comprises: determining temperature attributes of each sub-label in the user portrait label and / or the commodity portrait label; the temperature attribute is used to indicate the heat of a label being accessed; acquiring a storage strategy corresponding to the sub-label based on the temperature attribute; storing the sub-label in a corresponding database based on the storage strategy.

7. The method according to any one of claims 1 to 5, characterized in that, The method further comprises: in response to a query request, determining a user identifier list corresponding to the query request based on the user portrait label; the query request comprises at least one query label. Based on the commodity image label, a commodity recommendation list corresponding to each user in the user identification list is generated.

8. The method of claim 7, wherein, The method further comprises: Based on the user image label and the commodity image label, a push strategy graph corresponding to the user is constructed; In response to a target event, the user is pushed with commodities based on the push strategy graph.

9. A label generating apparatus characterized by comprising: It comprises: An acquisition module is configured to acquire behavior data of a user and attribute data of commodities; the behavior data comprises historical behavior data and real-time behavior data; A first processing module is configured to acquire a user image label based on the historical behavior data and the real-time behavior data; A second processing module is configured to acquire a commodity image label based on the attribute data and the real-time behavior data.

10. An electronic device, comprising: It comprises: A memory is configured to store a computer program; A processor is configured to execute the computer program to implement the method in any one of claims 1-8.

11. A computer readable storage medium, characterized in that, A computer program is stored thereon, and the computer program is executed by a processor to implement the method in any one of claims 1-8.

12. A computer program product, characterised in that, A computer program is stored thereon, and the computer program is executed by a processor to implement the method in any one of claims 1-8.

Images