Intelligent recommendation method and apparatus, device, and storage medium

Abstract

The present application discloses an intelligent recommendation method and apparatus, device, and a storage medium. The method comprises: acquiring historical user data and product data corresponding to a diagnostic products; constructing a historical user profile on the basis of b preset usage scenarios and the historical user data, wherein b is a positive integer; determining recommendation rules on the basis of c preset user requirements, the product data, and the historical user profile; classifying the a diagnostic products on the basis of the product data, the historical user profile, and the recommendation rules to obtain a classification label sets; acquiring interaction content between a user and a preset large language model; determining user preference information on the basis of the interaction content; determining a target classification label set from among the a classification label sets on the basis of the user preference information; and outputting, to the user by means of the preset large language model, recommendation content related to a target diagnostic product corresponding to the target classification label set. The present application facilitates improvement of the accuracy of product recommendation.

Description

Intelligent recommendation methods, devices, equipment and storage media Technical Field

[0001] This application relates to the field of product promotion technology, and in particular to an intelligent recommendation method, apparatus, device and storage medium. Background Technology

[0002] As companies produce more and more types of products and target more and more customers, users' need to understand both new and old products, as well as their demand for new purchases and replacements, is gradually increasing.

[0003] Currently, users mainly learn about the products they want to buy by browsing official websites, visiting offline stores, and consulting after-sales follow-ups. Alternatively, businesses can also recommend products that meet user needs through the recommendation algorithms of various e-commerce platforms.

[0004] However, current recommendation algorithms only consider users' browsing and purchase history, which cannot accurately meet all user needs, resulting in low accuracy in product recommendations. Summary of the Invention

[0005] To address the aforementioned problems in the prior art, embodiments of this application provide an intelligent recommendation method, apparatus, device, and storage medium. This method constructs historical user profiles based on preset usage scenarios and historical user data. It then determines recommendation rules by combining preset user needs, historical user profiles, and product information. Furthermore, it categorizes diagnostic products based on product information, historical user profiles, and recommendation rules, obtaining a set of category tags corresponding to each diagnostic product. Based on the user's interaction with a preset large language model, it determines user preference information and identifies a target category tag set corresponding to that preference. Finally, it outputs relevant recommendation content for the target diagnostic products corresponding to the target category tag set to the user through the preset large language model. Thus, by comprehensively considering usage scenarios, historical user profiles, product information, and user preference information, the method recommends diagnostic products to users, improving the accuracy of product recommendations.

[0006] In a first aspect, embodiments of this application provide an intelligent recommendation method, including:

[0007] Retrieve historical user data and product information for 'a' diagnostic products; 'a' is a positive integer.

[0008] Based on b preset usage scenarios and the historical user data, a historical user profile is constructed; b is a positive integer.

[0009] Recommendation rules are determined based on c preset user needs, the product information, and the historical user profiles; c is a positive integer.

[0010] Based on the product information, the historical user profile, and the recommendation rules, the a diagnostic products are classified to obtain a set of category tags; each diagnostic product corresponds to one set of category tags; each set of category tags includes at least one category tag.

[0011] Obtain the interaction content between the user and the preset large language model;

[0012] Based on the interaction content, determine user preference information;

[0013] Based on the user preference information, determine the target category tag set from the a category tag sets;

[0014] The preset large language model outputs relevant recommendations for target diagnostic products to the user, which correspond to the target classification tag set.

[0015] Secondly, embodiments of this application provide an intelligent recommendation device, the device comprising:

[0016] The acquisition unit is used to acquire historical user data and product data corresponding to a diagnostic products; a is a positive integer.

[0017] The processing unit is used to construct a historical user profile based on b preset usage scenarios and the historical user data; b is a positive integer.

[0018] Recommendation rules are determined based on c preset user needs, the product information, and the historical user profiles; c is a positive integer.

[0019] Based on the product information, the historical user profile, and the recommendation rules, the a diagnostic products are classified to obtain a set of category tags; each diagnostic product corresponds to one set of category tags; each set of category tags includes at least one category tag.

[0020] The acquisition unit is used to acquire the interaction content between the user and the preset large language model;

[0021] The processing unit is used to determine user preference information based on the interaction content;

[0022] Based on the user preference information, determine the target category tag set from the a category tag sets;

[0023] The preset large language model outputs relevant recommendations for target diagnostic products to the user, which correspond to the target classification tag set.

[0024] Thirdly, embodiments of this application provide an electronic device, including: a processor and a memory, the processor being connected to the memory, the memory being used to store a computer program, and the processor being used to execute the computer program stored in the memory, so that the electronic device performs the method as described in the first aspect.

[0025] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that is executed by a processor to implement the method described in the first aspect.

[0026] Fifthly, embodiments of this application provide a computer program product, the computer program product including a non-transitory computer-readable storage medium storing a computer program, the computer program being executed by a processor to implement the method as described in the first aspect.

[0027] Implementing the embodiments of this application has the following beneficial effects:

[0028] In this embodiment, historical user data and product data corresponding to *a* diagnostic products are first obtained. Then, historical user profiles are constructed based on *b* preset usage scenarios and historical user data. Recommendation rules are determined based on *c* preset user needs, product data, and historical user profiles. Further, the *a* diagnostic products are categorized based on product data, historical user profiles, and recommendation rules, resulting in a set of category tags for each of the *a* diagnostic products. Next, the interaction content between the user and the preset big oracle model is obtained. Based on the interaction content, user preference information is determined, and a target category tag set is determined from the *a* category tag sets based on the user preference information. Finally, the preset big oracle model outputs relevant recommendation content for the target diagnostic products corresponding to the target category tag set to the user. In this way, usage scenarios can be combined with user data to recommend diagnostic products that meet specific usage scenarios. Furthermore, product data, user profiles, and user needs can be combined to determine recommendation rules that meet different user needs. Diagnostic products can also be categorized by combining product data, historical user profiles, and recommendation rules. Based on the user's interaction with the large model, category tags that match the user's preferences are selected from the category tags to output relevant recommended content corresponding to the target diagnostic product to the user. This comprehensively considers all factors that meet user needs, accurately recommending product content to the user and improving the accuracy of product recommendations. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 is a schematic diagram of an application scenario of an intelligent recommendation method provided in an embodiment of this application;

[0031] Figure 2 is a flowchart illustrating an intelligent recommendation method provided in an embodiment of this application;

[0032] Figure 3 is a schematic diagram of the mapping relationship corresponding to a recommendation rule provided in an embodiment of this application;

[0033] Figure 4 is a flowchart illustrating a method for determining a classification label set according to an embodiment of this application;

[0034] Figure 5 is a schematic diagram of a classification label set provided in an embodiment of this application;

[0035] Figure 6 is a flowchart illustrating a method for determining a target classification label set according to an embodiment of this application;

[0036] Figure 7 is a functional unit block diagram of an intelligent recommendation device provided in an embodiment of this application;

[0037] Figure 8 is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0038] 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.

[0039] The terms "first," "second," "third," and "fourth," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0040] In this document, the term "embodiment" means that a particular feature, result, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0041] First, referring to Figure 1, Figure 1 is a schematic diagram of an application scenario of an intelligent recommendation method provided by an embodiment of this application. The scenario shown in Figure 1 includes an intelligent recommendation system, which includes a server, an interactive device, and a database.

[0042] The server primarily handles data requests, data caching, and program execution, including application servers such as Java application servers and .NET application servers. The server may also include a web server for publishing and managing websites, receiving HTTP requests from client browsers, and returning corresponding web page content. The server may also include a cloud server for cloud computing and big data processing. Furthermore, the server may utilize virtual server software to create multiple isolated Virtual Private Servers (VPSs) on a single physical server. This application does not specifically limit the type of server. Optionally, the server may also include a communication gateway for route calculation and routing message forwarding, such as a Vehicle Communication Interface (VCI) in the automotive electronics field.

[0043] Interactive devices primarily include electronic devices that can be used for data communication and are operable to enable user interaction, including touch-screen smart communication devices such as tablets and smartphones. Interactive devices may also include computing devices that interact with users through external devices, such as computers and smart appliances. Interactive devices may also include operable smart wearable devices, such as smartwatches and VR glasses. This application does not specifically limit the type of interactive device. Optionally, interactive devices may include vehicle diagnostic equipment for automobiles. It should be noted that a product intelligent recommendation program is installed on the interactive device. Users can access the product intelligent recommendation program by operating the interactive device and interact with the preset large language model embedded in the product intelligent recommendation program. During the interaction, the interactive device can send an interaction request to the server, and the server returns the interaction content to the interactive device by executing the product intelligent recommendation program and calling the preset large language model.

[0044] The database may include servers for data storage and management, such as file servers, database servers, cloud servers, etc. The database may also include storage for data reading and writing functions, including memory, disks, etc. This application does not specifically limit the type of database. In this embodiment, the database is mainly used to store product information and historical user information corresponding to diagnostic products. Diagnostic products mainly include products used to detect the technical condition of vehicles, find the causes of faults, and evaluate vehicle performance, such as vehicle diagnostic equipment, VCI, test benches, etc. The database can read historical user information and product information from the storage of the server or interactive device, and send the historical user information and product information to the server or interactive device when the server or interactive device sends a request to the database.

[0045] In some traditional product recommendation methods, a user's browsing and purchase history for diagnostic products on their interactive device can be stored in a database. The server can then retrieve this browsing and purchase history and recommend relevant products to the user based on it. However, a user's actual needs for purchasing diagnostic products may vary each time, and relying solely on browsing and purchase history cannot accurately meet all of a user's needs, resulting in low accuracy in product recommendations.

[0046] Therefore, in the intelligent recommendation method provided in this application, the server in the aforementioned intelligent recommendation system obtains historical user data and product data corresponding to a diagnostic products; where a is a positive integer.

[0047] The server constructs historical user profiles based on b preset usage scenarios and historical user data; b is a positive integer.

[0048] The server determines the recommendation rules based on c preset user needs, product information, and historical user profiles; c is a positive integer.

[0049] The server categorizes 'a' diagnostic products based on product information, historical user profiles, and recommendation rules, resulting in 'a' category tag sets. Each diagnostic product corresponds to one category tag set, and each category tag set includes at least one category tag.

[0050] The server obtains the interaction content between the user and the preset large language model;

[0051] The server determines user preference information based on the content of the interaction;

[0052] The server determines the target category tag set from a set of category tags based on user preference information;

[0053] The server uses a pre-defined large language model to output relevant recommendations for target diagnostic products that correspond to the target classification tag set.

[0054] As can be seen, in the above scenario, the server can obtain historical user data and product data corresponding to the diagnostic products from the database. Then, based on the preset usage scenarios and historical user data, a historical user profile is constructed, and recommendation rules are determined based on preset user needs, product data, and historical user profiles. Next, the diagnostic products are classified according to product data, historical user profiles, and recommendation rules, resulting in a set of product category tags. Furthermore, the server can determine user preference information based on the user's interaction with the preset large language model, thereby selecting a target category tag set that matches the user's preference information from the category tag set, and then outputting relevant recommendation content for the target diagnostic products corresponding to the target category tag set to the user through the preset large language model. In this way, by comprehensively considering usage scenarios, historical user profiles, product data, and user preference information, the server recommends diagnostic products to users, taking into account both the user's historical preferences and the user's actual needs, thus improving the accuracy of product recommendations.

[0055] The method provided in the embodiments of this application will be described below. Referring to Figure 2, Figure 2 is a flowchart illustrating an intelligent recommendation method provided in the embodiments of this application. This method is applied to the server in the above scenario, and the method includes, but is not limited to, the following steps:

[0056] 201: Obtain historical user data and product data corresponding to diagnostic products (a).

[0057] In this embodiment, 'a' is a positive integer. Historical user data may include basic information registered by the user when using the product intelligent recommendation program on the interactive device, as well as historical behavioral information and historical interaction content during the use of the product intelligent recommendation program. Product data may include information such as the product functions and features corresponding to 'a' diagnostic products.

[0058] It should be noted that historical user data and product data corresponding to the 'a' diagnostic products can be uploaded in advance by administrators and stored in the database or the memory of the interactive device. Furthermore, when historical user data and product data are updated, the new data uploaded by administrators can be synchronously updated to the database or the memory of the interactive device. The server can directly retrieve historical user data and product data corresponding to the 'a' diagnostic products from the database or the memory of the interactive device.

[0059] 202: Based on b preset usage scenarios and historical user data, construct historical user profiles.

[0060] In this embodiment, b is a positive integer. The preset use scenario can be a specific scenario in which a user uses the diagnostic product, such as an individual checking the operating status of a car, repairing a car brake pad at a small repair shop, repairing an engine at a large repair shop, and so on.

[0061] Understandably, users have different needs for diagnostic products in different use cases. Based on historical user data, we can analyze the diagnostic products that users are interested in. By combining use cases with historical user data, we can analyze the product suitability of the diagnostic products that users are interested in for each use case.

[0062] In some feasible embodiments, constructing a historical user profile based on b preset usage scenarios and historical user data may include:

[0063] Based on the basic information, determine the user's basic information tags;

[0064] Based on historical behavior information, determine the user's historical consumption information for diagnostic products a;

[0065] Based on basic information and historical consumption information, determine the user's historical spending power tag;

[0066] Based on historical behavior information, determine the user's historical attention to diagnostic product a;

[0067] Perform semantic recognition on historical interaction content to determine the user's historical needs;

[0068] Based on historical consumption information, historical attention, and historical needs, determine the user's historical interest tags;

[0069] Based on basic information tags, historical spending power tags, and historical interest tags, determine the product adaptation tags for users in b preset usage scenarios;

[0070] Based on basic information tags, historical spending power tags, historical interest tags, and product compatibility tags, historical user profiles are determined.

[0071] Historical user data can include: basic information, historical behavior information, and historical interaction content. Basic information includes personal information registered by the user when using the product intelligent recommendation program, such as age, gender, and occupation. Historical behavior information can include the user's browsing history, consumption history, and historical interaction records with diagnostic products, such as likes and comments. Historical interaction content includes all interactions between the user and the preset large language model. The server can construct a historical user profile based on b preset usage scenarios and historical user data. This historical user profile is used to label the user's basic information, historical spending power, historical interest in diagnostic products, and product suitability.

[0072] Specifically, the server first labels the user with basic information based on their registration details. This label can include information such as the user's age, gender, and occupation. It should be noted that users of different ages, genders, and occupations have different interests in different diagnostic products. Therefore, the user's basic information can serve as one indicator to determine the diagnostic products the user is interested in, allowing for more accurate product recommendations.

[0073] Then, the server can obtain the user's historical consumption information for 'a' diagnostic products when using the intelligent product recommendation program from historical behavior information. This historical consumption information includes the user's purchase records for diagnostic products. It is understandable that the user's historical consumption information can reflect the user's spending power and budget range for purchasing diagnostic products. Based on this, accurate recommendations for diagnostic products can be achieved by recommending diagnostic products that match the user's spending power.

[0074] Next, the server can label the user's historical spending power based on the user's basic information and historical consumption information, obtaining a historical spending power tag. Specifically, the server can determine the spending power score corresponding to the user's basic information. Different age, gender, occupation, and other information correspond to different spending power scores, and the mapping relationship between basic information and spending power scores can be pre-defined. Furthermore, based on the user's historical consumption information for 'a' diagnostic products and the corresponding consumption range, the server can determine the user's historical spending power score, and the mapping relationship between consumption range and spending power score can also be pre-defined. Based on the spending power scores corresponding to basic information and historical consumption information, the user's historical spending power can be labeled, obtaining a historical spending power tag. For example, this historical spending power tag can include the user's historical budget range and spending power score.

[0075] Then, the server can determine the user's historical attention level towards 'a' diagnostic products based on historical behavior information. Specifically, the server can obtain the user's historical interaction records with the diagnostic products, such as likes, comments, and favorites. A corresponding historical attention level is set for each type of historical interaction record, and a corresponding historical attention level is also set for the specific content of the comments. For example, the historical attention level corresponding to positive reviews is different from that corresponding to negative reviews. Based on this, the server can determine the user's historical attention level towards 'a' diagnostic products according to the user's historical interaction records.

[0076] Furthermore, the server can perform semantic recognition on historical interaction content to determine the user's historical needs. Optionally, a pre-set large language model can be used to perform semantic recognition on historical interaction content to determine the semantic information corresponding to the historical interaction content, and based on this semantic information, determine the user's historical needs. For example, the pre-set large language model can be any of the following: Generative Pre-trained Transformer (GPT), Bidirectional Encoder Representation from Transformers (BERT), etc.

[0077] Furthermore, the server can determine the user's historical interest tags for the diagnostic products (a) based on the user's historical consumption information, historical attention to these products, and historical needs. These historical interest tags are used to label the user's interest value for each of the diagnostic products (a). For example, the server can determine the interest value corresponding to historical consumption information based on the number of times the user has purchased each diagnostic product, the interest value corresponding to historical attention based on the user's historical attention to each product, and the interest value corresponding to historical needs based on the user's historical needs for each product. The sum of these interest values—the historical consumption value, the historical attention value, and the historical needs value—is then used as the user's interest value for each diagnostic product, thus labeling the user's historical interest tags.

[0078] Furthermore, the server can determine product compatibility tags for a user in b preset usage scenarios based on basic information tags, historical spending power tags, and historical interest tags. Specifically, based on historical interest tags, diagnostic products that the user is interested in can be identified. Then, based on historical spending power tags, diagnostic products matching the user's historical spending power are selected from the diagnostic products that match the user's historical spending power. Next, based on the user's basic information, diagnostic products matching the user's age, gender, occupation, etc., are selected from the diagnostic products matching the user's historical spending power. Finally, based on the compatibility between the diagnostic products and the b preset usage scenarios, the compatibility between the user and the diagnostic products is labeled, resulting in product compatibility tags for the b preset usage scenarios. The compatibility between the diagnostic products and the b preset usage scenarios can be preset.

[0079] Finally, basic information tags, historical spending power tags, historical interest tags, and product compatibility tags are used as user tags to obtain historical user profiles.

[0080] As can be seen, by analyzing users' historical behavior information, we can obtain their historical consumption information of diagnostic products. Based on the user's basic information and historical consumption information, we can determine the user's historical spending power tags and their historical attention to diagnostic products. Then, based on the user's historical interaction content with the preset large language model, we can determine the user's historical needs. Therefore, based on historical consumption information, historical attention, and historical needs, we can analyze and obtain the user's historical interest tags. Furthermore, based on basic information tags, historical spending power tags, and historical interest tags, we can determine the product compatibility tags for each diagnostic product in b preset usage scenarios, thus obtaining a historical user profile. Based on this, we can determine the diagnostic products that users are interested in and the compatibility of each diagnostic product, thereby recommending diagnostic products to users more accurately and improving the accuracy of product recommendations.

[0081] 203: Determine the recommendation rules based on c preset user needs, product information, and historical user profiles.

[0082] In this embodiment, c is a positive integer. Preset user requirements are pre-defined user requirements corresponding to product information. Recommendation rules are used to represent the mapping relationship between preset user requirements, product information, and historical user profiles.

[0083] For example, product information may include: product functional information and non-product functional information. Each preset user requirement may include product functional requirements and budget requirements. Product functional information is used to indicate the product functions corresponding to the diagnostic product. Non-product functional information may include information such as the product's price and applicable vehicle models.

[0084] In some feasible embodiments, the recommendation rules are determined based on c preset user needs, product information, and historical user profiles, and may include:

[0085] Based on the product function information and the product function requirements corresponding to each of the c preset user requirements, determine the diagnostic product set corresponding to each preset user requirement, and obtain c first diagnostic product sets.

[0086] Based on non-product function information, determine the price of each diagnostic product in the c first diagnostic product sets;

[0087] Based on the price, historical spending power tag, and budget requirement corresponding to each preset user need of each diagnostic product in the c first diagnostic product sets, the diagnostic products in the c first diagnostic product sets are adjusted to obtain c second diagnostic product sets.

[0088] Based on historical interest tags and product adaptation tags, the diagnostic products in the c second diagnostic product sets are adjusted to obtain c target diagnostic product sets;

[0089] Determine the mapping relationship between c preset user needs, historical spending power tags, historical interest tags, product adaptation tags, and c target diagnostic product sets, and determine the mapping relationship as recommendation rules.

[0090] Each first diagnostic product set includes at least one diagnostic product. Each second diagnostic product set includes at least one diagnostic product. Each preset user requirement corresponds to a target diagnostic product set. Each target diagnostic product set includes at least one diagnostic product. The c target diagnostic product sets include a diagnostic products.

[0091] Specifically, the server first selects a set of diagnostic products from a set of diagnostic products based on product feature information and the product feature requirements corresponding to each of c preset user needs, resulting in c first diagnostic product sets. Diagnostic products in any two first diagnostic product sets can be duplicated. Then, the server obtains the price corresponding to each diagnostic product in the c first diagnostic product sets from non-product feature information. It matches the price of each diagnostic product in the c first diagnostic product sets with historical spending power tags to adjust the diagnostic products in the c first diagnostic product sets, resulting in c adjusted first diagnostic product sets. Finally, it matches the price of each diagnostic product in the adjusted c first diagnostic product sets with the budget requirements corresponding to the preset user needs, further adjusting the diagnostic products in the adjusted c first diagnostic product sets to obtain c second diagnostic product sets.

[0092] Furthermore, the server filters the diagnostic products in the c second diagnostic product sets based on historical demand tags to ensure that the diagnostic products in the second diagnostic product sets match the historical demand tags, resulting in c adjusted second diagnostic product sets. Based on product adaptation tags, the adaptation degree of each diagnostic product in the c adjusted second diagnostic product sets is determined. Then, based on the adaptation degree of each diagnostic product in the c adjusted second diagnostic product sets, the diagnostic products in the c adjusted second diagnostic product sets are filtered to ensure that the adaptation degree of each diagnostic product in the second diagnostic product sets is greater than a preset adaptation degree, thus obtaining c target diagnostic product sets. In this way, c preset user needs, historical spending power tags, historical interest tags, product adaptation tags, and c target diagnostic product sets can be mapped one-to-one, establishing the mapping relationship shown in Figure 3. Here, preset user needs, historical spending power tags, historical interest tags, product adaptation tags, and target diagnostic product sets are mutually mapped. The mapping relationship between the c preset user needs, historical spending power tags, historical interest tags, product adaptation tags, and c target diagnostic product sets is used as a recommendation rule.

[0093] As can be seen, in this embodiment, by matching product function information with product function requirements corresponding to preset user needs, c first diagnostic product sets are determined. The price of each diagnostic product in the c first diagnostic product sets is determined based on non-product function information. Based on the price of each diagnostic product, the user's historical spending power tags are matched with the budget requirements corresponding to each preset user need to determine c second diagnostic product sets. Then, through historical interest tags and product compatibility tags, the diagnostic products in the c second diagnostic product sets are filtered to obtain c target diagnostic product sets. This allows for the mapping relationship between preset user needs, historical spending power tags, historical interest tags, product compatibility tags, and target diagnostic product sets, and this mapping relationship is used as a recommendation rule. Thus, by determining the mapping relationship between user needs and historical user profiles and product data, diagnostic products matching user needs can be recommended more accurately, improving the accuracy of product recommendations.

[0094] 204: Based on product information, historical user profiles, and recommendation rules, classify a diagnostic products to obtain a set of classification tags.

[0095] In this embodiment, each diagnostic product corresponds to a set of category tags. Each set of category tags includes at least one category tag. The server can classify and label 'a' diagnostic products based on the mapping relationship between product information, historical user profiles, and tags corresponding to recommendation rules, thus obtaining 'a' set of category tags.

[0096] For example, as shown in Figure 4, based on product information, historical user profiles, and recommendation rules, a diagnostic products are classified to obtain a set of classification tags, which may include:

[0097] 401: Based on the recommendation rules, determine the preset user demand set corresponding to each of the a diagnostic products, and obtain a preset user demand sets;

[0098] 402: Based on product information and a preset user demand sets, determine the functional category tag set and price category tag set corresponding to each of the a diagnostic products, and obtain a functional category tag set and a price category tag set;

[0099] 403: Based on historical user profiles and a preset user demand sets, determine the user category tag set corresponding to each of the a diagnostic products, and obtain a user category tag sets;

[0100] 404: Based on a set of a preset user needs, a set of a functional category tags, a set of a price category tags, and a set of a user category tags, determine a set of a category tags.

[0101] In this embodiment, each preset user requirement set includes at least one preset user requirement. Each function category tag set includes at least one function category tag. Each user category tag set includes at least one user category tag.

[0102] Specifically, the server first matches the diagnostic products in the target diagnostic product set with the preset user needs according to the mapping relationship between the preset user needs in the recommendation rules and the target diagnostic product set. It can be understood that c preset user needs include a combination of all product function requirements and budget requirements corresponding to a diagnostic products. Therefore, each of the a diagnostic products can be matched with at least one preset user need. Thus, a set of a preset user needs corresponding to a diagnostic products can be obtained.

[0103] Then, the server can, based on the preset user demand set corresponding to each of the *a* diagnostic products, search the product data for the product function corresponding to each preset user demand in the preset user demand set, and determine the function category tag set corresponding to each of the *a* diagnostic products based on the product function, thus obtaining *a* function category tag sets. Based on the preset user demand set corresponding to each of the *a* diagnostic products, the server can search the product data for the product price corresponding to each budget demand in the preset user demand set, and determine the price category tag corresponding to each of the *a* diagnostic products based on the product price, thus obtaining *a* price category tags.

[0104] Furthermore, based on historical user profiles and *a* preset user demand sets, the server determines the user category tag set corresponding to each of the *a* diagnostic products. Specifically, the server matches the preset user demands in the preset user demand set corresponding to each of the *a* diagnostic products with the basic information tags, historical spending power tags, historical interest tags, and product compatibility tags corresponding to the historical user profiles. All the basic information tags, historical spending power tags, historical interest tags, and product compatibility tags corresponding to each diagnostic product are then used as the user category tags for each diagnostic product, resulting in the user category tag set for each diagnostic product, thus obtaining *a* user category tag sets.

[0105] Finally, the server can combine the preset user requirement set, functional category tag set, price category tag set, and user category tag set of each diagnostic product into a category tag set for that diagnostic product, resulting in the category tag set shown in Figure 5. In the diagnostic product's category tag set, there is only one price category tag, and there is at least one preset user requirement tag from the preset user requirement set, one functional category tag from the functional category tag set, and one user category tag from the user category tag set. Optionally, the category tag set may also include only some of the tags mentioned above.

[0106] As can be seen, based on the recommendation rules, each diagnostic product can be matched with preset user needs to obtain a preset user need set corresponding to each diagnostic product. Then, based on product information and each preset user need set, the functional category tag set and price category tag set corresponding to each diagnostic product can be determined. Furthermore, based on historical user profiles and each preset user need set, the user category tag set corresponding to each diagnostic product can be determined, thus obtaining the category tag set corresponding to each diagnostic product. Therefore, diagnostic products can be accurately categorized based on functional needs, price needs, user spending power, user interests, and product suitability to precisely match users' actual needs and recommend diagnostic products accurately, thereby improving the accuracy of product recommendations.

[0107] In some feasible embodiments, classifying *a* diagnostic products based on product information, historical user profiles, and recommendation rules to obtain *a* sets of category labels can be achieved using a pre-set large language model, which can be a GPT model or a BERT model. The pre-set large language model extracts features from product information, historical user profiles, and recommendation rules, breaks down the extracted features into the smallest units, and then classifies these units using a classifier, outputting *a* sets of category labels corresponding to the *a* diagnostic products. Based on this, by pre-training the pre-set large language model, it can accurately classify each diagnostic product, thereby ensuring that the category label sets accurately match the user's actual needs and accurately recommend diagnostic products to the user, improving the accuracy of product recommendations.

[0108] 205: Obtain the interaction content between the user and the preset large language model.

[0109] In this embodiment, a user can interact with a preset large language model on an interactive device. The interaction methods may include text, voice, sign language, etc. The response content output by the preset large language model can guide the user to input their actual needs, previously used products, and products of interest. The server can obtain the aforementioned interaction content.

[0110] 206: Determine user preference information based on the interaction content.

[0111] In this embodiment, user preference information may include: the user's actual needs, the user's historical product usage, and products of interest. The server can perform semantic recognition on the interaction content using a preset large language model to determine the semantic information corresponding to the interaction content, and determine the user preference information based on the semantic information.

[0112] 207: Based on user preference information, determine the target category tag set from a category tag set.

[0113] In this embodiment, the server can match user preference information with the category tags corresponding to each diagnostic product to determine the target category tag set with the highest matching degree with the user preference information.

[0114] For example, user preference information may include: target product feature requirements and target budget requirements, historically used products, and products of interest. Target product feature requirements represent the user's actual needs for product features, and target budget requirements represent the user's actual budget for the product price.

[0115] In some feasible embodiments, as shown in Figure 6, determining the target category tag set from a set of category tags based on user preference information may include:

[0116] 601: Match the target product functional requirements and target budget requirements with a preset user requirement sets to obtain d preset user requirement sets;

[0117] 602: Based on the target budget requirement and the d price category labels corresponding to the d first candidate category label sets, determine e second candidate category label sets;

[0118] 603: Determine the matching degree between the product functional requirements and each functional category label in the e second candidate functional category label sets, and determine the number of labels in each second candidate functional category label set whose matching degree with the product functional requirements is greater than a first threshold, thus obtaining e first quantities;

[0119] 604: Based on e first quantities, determine f third candidate category labels from e sets of second candidate category labels;

[0120] 605: Determine the target category tag set based on historically used products, products of interest, and f user category tag sets corresponding to f third candidate category tag sets.

[0121] Among them, there are d preset user demand sets and d first candidate category label sets, each with a corresponding value. d is a positive integer less than or equal to a. e is a positive integer less than or equal to d. There are e second candidate function category label sets and e second candidate category label sets, each with a corresponding first quantity. f is a positive integer less than or equal to e.

[0122] Specifically, the server first searches for target user needs corresponding to the target product's functional requirements and target budget requirements from the set of a preset user needs corresponding to a diagnostic products, thereby filtering out preset user need sets that include the target user needs, resulting in d preset user need sets. Then, it obtains the category tag set corresponding to each of the d preset user need sets, resulting in d first candidate category tag sets. Next, the server searches for the price category tag corresponding to the target budget requirements from the d price category tags corresponding to the d first candidate category tag sets, resulting in e price category tags, and obtains e second candidate category tag sets corresponding to the e price category tags, with each price category tag corresponding to one second candidate category tag set.

[0123] Furthermore, the server determines the matching degree between the product functional requirements and each functional category tag in the e second candidate functional category tag sets, and identifies the functional category tags in each second candidate functional category tag set whose matching degree with the product functional requirements is greater than a first threshold. The server then counts the number of tags in each second candidate functional category tag set whose matching degree with the product functional requirements is greater than the first threshold, obtaining e first quantities. It can be understood that the larger the first quantity, the higher the matching degree between the user's product functional requirements and the functional category tags. The server then uses the set of category tags in the e second candidate functional category tag sets whose first quantities are greater than a first preset quantity as the third candidate functional category tag set, obtaining f third candidate functional category tag sets.

[0124] Finally, the server obtains the category tag sets corresponding to the user's historically used products and the category tag sets corresponding to the products of interest. It then matches the user's category tags corresponding to these sets with the user category tags in the f third candidate category tag sets. The server determines the number of tags in each third candidate category tag set that match all tags in both the historically used and interested product category tag sets, resulting in f second quantities. The third candidate category tag set with the largest second quantity is selected as the target category tag set.

[0125] Optionally, there can be multiple target category tag sets. The server can use multiple third candidate category tag sets, where the second number is greater than the second preset number, as multiple target category tag sets. Based on the second number, the multiple target category tag sets are sorted, and according to the sorting result, multiple target diagnostic products corresponding to the multiple target category tag sets are recommended to the user in sequence.

[0126] As can be seen, by selecting the user's target product function requirements, target budget requirements, historically used products, and products of interest from a set of a category tags, a target category tag set that meets the user's preferences can be selected sequentially. This allows for accurate matching of the user's actual needs and products of interest, leading to precise recommendations of diagnostic products and improving the accuracy of product recommendations.

[0127] Optionally, each user category tag set may include historical interest tags.

[0128] In some feasible embodiments, determining the target category tag set based on historically used products, products of interest, and f user category tag sets corresponding to f third candidate category tag sets may include:

[0129] By categorizing products based on historical usage and products of interest, interest tags are generated.

[0130] Determine the matching degree between the interest tag and each of the f historical interest tags corresponding to the f user category tag sets, and determine the category tag set with a matching degree greater than the third threshold in the f third candidate category tag sets as the target category tag set.

[0131] Interest tags include all tags corresponding to products used in the past and products of interest.

[0132] Specifically, the server can categorize historically used products and products of interest based on all tags corresponding to those products, thus obtaining interest tags. Then, the server determines the matching degree between each historical interest tag and the interest tag from the f historical interest tags corresponding to the f user category tag sets, where the matching between any two interest tags can be pre-defined. Thus, based on the matching degree between each historical interest tag and the interest tag from the f historical interest tags corresponding to the f user category tag sets, a third candidate category tag set with a matching degree greater than a third threshold can be determined as the target category tag set. Optionally, the server can also use the third candidate category tag set with the highest matching degree as the target category tag set.

[0133] As can be seen, interest tags can be obtained by classifying historically used products and products of interest. By determining the matching degree between the interest tags and each of the f historical interest tags corresponding to the f user category tag sets, the category tag sets with a matching degree greater than a third threshold in the f third candidate category tag sets can be identified as the target category tag set. In this way, the user's historically used products and products of interest can be associated with the target category tag set, thereby recommending diagnostic products of interest to the user and improving the accuracy of product recommendations.

[0134] 208: By using a pre-set large language model, the system outputs relevant recommendations for target diagnostic products to users that correspond to the target classification tag set.

[0135] In this embodiment of the application, the relevant recommended content for the target diagnostic product is the relevant recommended content output by a preset large language model based on product information and user interests. This relevant recommended content can be displayed on interactive devices in the form of text, voice, images, videos, etc.

[0136] For example, by using a pre-defined large language model, relevant recommendations for target diagnostic products corresponding to the target classification tag set are output to the user, including:

[0137] Identify the target product functional information and target non-product functional information in the product documentation that correspond to the target diagnostic product;

[0138] Determine the user's content interests based on the target category tag set and historical interest tags;

[0139] By using a pre-set large language model, relevant recommended content is output to users that corresponds to the target category tag set, target product function information, target non-product function information, and interest content.

[0140] In practice, the server first retrieves the target product function information and target non-product function information corresponding to the target diagnostic product from the product documentation. The target product function information represents the product functions corresponding to the target diagnostic product, such as reading fault codes and reading data streams. The target non-product function information may include information such as the price, performance, and applicable vehicle models of the target diagnostic product.

[0141] Then, the server determines the user's interest tags based on the target category tag set. Based on these interest tags and historical interest tags, it determines the user's content interests. These content interests may include diagnostic products the user is interested in, product features the user is interested in, product prices the user is interested in, and product recommendation formats the user is interested in. Product recommendation formats may include text, images, videos, and audio.

[0142] Finally, the target category tag set, target product function information, target non-product function information, and interest content are input into the preset large language model, and the preset large language model outputs relevant recommended content corresponding to the target diagnostic product.

[0143] Therefore, by comprehensively considering product information and user interests, relevant content corresponding to the target diagnostic product can be recommended to the user. This improves the accuracy of product recommendations by recommending diagnostic products that users are interested in while satisfying their interests.

[0144] In summary, in this embodiment, historical user data and product data corresponding to *a* diagnostic products are first obtained. Then, historical user profiles are constructed based on *b* preset usage scenarios and historical user data. Recommendation rules are determined based on *c* preset user needs, product data, and historical user profiles. Further, the *a* diagnostic products are categorized based on product data, historical user profiles, and recommendation rules, resulting in a set of category tags for each of the *a* diagnostic products. Next, the interaction content between the user and the preset big oracle model is obtained. Based on the interaction content, user preference information is determined, and a target category tag set is determined from the *a* category tag sets based on the user preference information. Finally, the preset big oracle model outputs relevant recommendation content for the target diagnostic products corresponding to the target category tag set to the user. In this way, usage scenarios can be combined with user data to recommend diagnostic products that meet specific usage scenarios. Furthermore, product data, user profiles, and user needs can be combined to determine recommendation rules that meet different user needs. Diagnostic products can also be categorized by combining product data, historical user profiles, and recommendation rules. Based on the user's interaction with the large model, category tags that match the user's preferences are selected from the category tags to output relevant recommended content corresponding to the target diagnostic product to the user. This comprehensively considers all factors that meet user needs, accurately recommending product content to the user and improving the accuracy of product recommendations.

[0145] Referring to Figure 7, Figure 7 is a functional unit block diagram of an intelligent recommendation device provided in an embodiment of this application. The intelligent recommendation device 700 may include a server of any of the above embodiments. Optionally, the intelligent recommendation device 700 may include a VCI of a vehicle diagnostic device. The intelligent recommendation device 700 includes an acquisition unit 701 and a processing unit 702.

[0146] Among them, the acquisition unit 701 is used to acquire historical user data and product data corresponding to a diagnostic products; where a is a positive integer.

[0147] Processing unit 702 is used to construct historical user profiles based on b preset usage scenarios and historical user data; b is a positive integer;

[0148] The recommendation rules are determined based on c preset user needs, product information, and historical user profiles; c is a positive integer.

[0149] Based on product information, historical user profiles, and recommendation rules, classify a diagnostic products to obtain a set of category tags; each diagnostic product corresponds to one set of category tags; each set of category tags includes at least one category tag.

[0150] Acquisition unit 701 is used to acquire the interaction content between the user and the preset large language model;

[0151] The processing unit 702 is used to determine user preference information based on the interaction content;

[0152] Based on user preference information, determine the target category tag set from a set of category tags;

[0153] By using a pre-set large language model, relevant recommendations for target diagnostic products are output to users, corresponding to the target classification tag set.

[0154] In some feasible implementations, historical user data includes: basic information, historical behavior information, and historical interaction content.

[0155] In constructing historical user profiles based on b preset usage scenarios and historical user data, processing unit 702 is specifically used for:

[0156] Based on the basic information, determine the user's basic information tags;

[0157] Based on historical behavior information, determine the user's historical consumption information for diagnostic products a;

[0158] Based on basic information and historical consumption information, determine the user's historical spending power tag;

[0159] Based on historical behavior information, determine the user's historical attention to diagnostic product a;

[0160] Perform semantic recognition on historical interaction content to determine the user's historical needs;

[0161] Based on historical consumption information, historical attention, and historical needs, determine the user's historical interest tags;

[0162] Based on basic information tags, historical spending power tags, and historical interest tags, determine the product adaptation tags for users in b preset usage scenarios;

[0163] Based on basic information tags, historical spending power tags, historical interest tags, and product compatibility tags, historical user profiles are determined.

[0164] In some feasible embodiments, product information includes: product function information and non-product function information; each preset user requirement includes product function requirements and budget requirements;

[0165] In determining recommendation rules based on c preset user needs, product information, and historical user profiles, processing unit 702 is specifically used for:

[0166] Based on the product function information and the product function requirements corresponding to each of the c preset user requirements, determine the diagnostic product set corresponding to each preset user requirement to obtain c first diagnostic product sets; each first diagnostic product set includes at least one diagnostic product.

[0167] Based on non-product function information, determine the price of each diagnostic product in the c first diagnostic product sets;

[0168] Based on the price, historical spending power tag, and budget requirement corresponding to each preset user need of each diagnostic product in the c first diagnostic product sets, the diagnostic products in the c first diagnostic product sets are adjusted to obtain c second diagnostic product sets; each second diagnostic product set includes at least one diagnostic product.

[0169] Based on historical interest tags and product adaptation tags, the diagnostic products in the c second diagnostic product sets are adjusted to obtain c target diagnostic product sets; each preset user requirement corresponds to one target diagnostic product set; each target diagnostic product set includes at least one diagnostic product; the diagnostic products in the c target diagnostic product sets include a diagnostic products;

[0170] Determine the mapping relationship between c preset user needs, historical spending power tags, historical interest tags, product adaptation tags, and c target diagnostic product sets, and determine the mapping relationship as recommendation rules.

[0171] In some feasible embodiments, in classifying a diagnostic products based on product information, historical user profiles, and recommendation rules to obtain a set of classification tags, the processing unit 702 is specifically used for:

[0172] Based on the recommendation rules, determine the preset user demand set corresponding to each of the a diagnostic products, and obtain a preset user demand set; each preset user demand set includes at least one preset user demand.

[0173] Based on product information and a preset user demand sets, determine the functional category tag set and price category tag set corresponding to each of the a diagnostic products, resulting in a functional category tag set and a price category tag set; each functional category tag set includes at least one functional category tag.

[0174] Based on historical user profiles and a preset user demand sets, determine the user category tag set corresponding to each of the a diagnostic products, resulting in a user category tag set; each user category tag set includes at least one user category tag.

[0175] Based on a set of pre-defined user needs, a set of functional category tags, a set of price category tags, and a set of user category tags, determine a set of category tags.

[0176] In some feasible implementations, user preference information includes: target product feature requirements and target budget requirements, historically used products, and products of interest;

[0177] In determining the target category tag set from a set of a category tags based on user preference information, processing unit 702 is specifically used for:

[0178] The target product functional requirements and target budget requirements are matched with a preset user requirement sets to obtain d preset user requirement sets; each of the d preset user requirement sets corresponds one-to-one with a first candidate category tag set; d is a positive integer less than or equal to a.

[0179] Based on the target budget requirement and the d price category labels corresponding to the d first candidate category label sets, determine e second candidate category label sets; e is a positive integer less than or equal to d;

[0180] Determine the matching degree between the product functional requirements and each functional category label in the e second candidate functional category label sets, and determine the number of labels in each second candidate functional category label set whose matching degree with the product functional requirements is greater than a first threshold, thus obtaining e first quantities; the e second candidate functional category label sets correspond one-to-one with the e second candidate category label sets; each second candidate functional category label set corresponds to one first quantity;

[0181] Based on e first quantities, determine f third candidate category labels from e second candidate category label sets; f is a positive integer less than or equal to e.

[0182] The target category tag set is determined based on historical product usage, products of interest, and f user category tag sets corresponding to f third candidate category tag sets.

[0183] In some feasible implementations, each user category tag set includes historical interest tags;

[0184] In determining the target category tag set based on historically used products, products of interest, and f user category tag sets corresponding to f third candidate category tag sets, processing unit 702 is specifically used for:

[0185] Products used in the past and products of interest are categorized to obtain interest tags; interest tags include all tags corresponding to products used in the past and products of interest.

[0186] Determine the matching degree between the interest tag and each of the f historical interest tags corresponding to the f user category tag sets, and determine the category tag set with a matching degree greater than the third threshold in the f third candidate category tag sets as the target category tag set.

[0187] In some feasible embodiments, in outputting relevant recommended content of the target diagnostic product corresponding to the target classification tag set to the user through a preset large language model, the processing unit 702 is specifically used for:

[0188] Identify the target product functional information and target non-product functional information in the product documentation that correspond to the target diagnostic product;

[0189] Determine the user's content interests based on the target category tag set and historical interest tags;

[0190] By using a pre-set large language model, relevant recommended content is output to users that corresponds to the target category tag set, target product function information, target non-product function information, and interest content.

[0191] Referring to Figure 8, which is a schematic diagram of the structure of an electronic device provided in an embodiment of this application, the electronic device 800 includes a transceiver 801, a processor 802, and a memory 803. These are connected via a bus 804. The memory 803 stores computer programs and data, and can transmit data stored in the memory 803 to the processor 802. The electronic device 800 may include an intelligent recommendation device 700. The electronic device 800 may also be a server in any of the above embodiments. The electronic device 800 may also include a vehicle diagnostic interface (VCI).

[0192] Processor 802 is used to read the computer program in memory 803 and perform the following operations:

[0193] Retrieve historical user data and product information for 'a' diagnostic products; 'a' is a positive integer.

[0194] Based on b preset usage scenarios and historical user data, construct historical user profiles; b is a positive integer;

[0195] The recommendation rules are determined based on c preset user needs, product information, and historical user profiles; c is a positive integer.

[0196] Based on product information, historical user profiles, and recommendation rules, classify a diagnostic products to obtain a set of category tags; each diagnostic product corresponds to one set of category tags; each set of category tags includes at least one category tag.

[0197] Obtain the interaction content between the user and the preset large language model;

[0198] Determine user preference information based on the interaction content;

[0199] Based on user preference information, determine the target category tag set from a set of category tags;

[0200] By using a pre-set large language model, relevant recommendations for target diagnostic products are output to users, corresponding to the target classification tag set.

[0201] The above primarily describes the solutions of the embodiments of this application from the perspective of the method execution process. It is understood that, in order to achieve the above functions, the electronic device includes corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments provided herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0202] This application also provides a computer-readable storage medium storing a computer program that is executed by a processor to implement some or all of the steps of any of the methods described in the above method embodiments.

[0203] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments.

[0204] 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.

[0205] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0206] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0207] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0208] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software program module.

[0209] If the integrated unit 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 described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0210] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0211] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. An intelligent recommendation method, characterized in that, include: Retrieve historical user data and product information for 'a' diagnostic products; 'a' is a positive integer. Based on b preset usage scenarios and the historical user data, a historical user profile is constructed; b is a positive integer. Recommendation rules are determined based on c preset user needs, the product information, and the historical user profiles; c is a positive integer. Based on the product information, the historical user profile, and the recommendation rules, the a diagnostic products are classified to obtain a set of category tags; each diagnostic product corresponds to one set of category tags; each set of category tags includes at least one category tag. Obtain the interaction content between the user and the preset large language model; Based on the interaction content, determine user preference information; Based on the user preference information, determine the target category tag set from the a category tag sets; The preset large language model outputs relevant recommendations for target diagnostic products to the user, which correspond to the target classification tag set.

2. The method according to claim 1, characterized in that, The historical user data includes: basic information, historical behavior information, and historical interaction content; The step of constructing historical user profiles based on b preset usage scenarios and the historical user data includes: Based on the aforementioned basic information, determine the user's basic information tags; Based on the historical behavior information, determine the user's historical consumption information for the a diagnostic products; Based on the basic information and the historical consumption information, the user's historical spending power tag is determined; Based on the historical behavior information, determine the user's historical attention to the a diagnostic products; Perform semantic recognition on the historical interaction content to determine the user's historical needs; Based on the historical consumption information, historical attention, and historical needs, the user's historical interest tags are determined; Based on the basic information tags, the historical spending power tags, and the historical interest tags, determine the product compatibility tags for the user in the b preset usage scenarios; The historical user profile is determined based on the basic information tags, the historical spending power tags, the historical interest tags, and the product compatibility tags.

3. The method according to claim 2, characterized in that, The product information includes: product function information and non-product function information; each preset user requirement includes product function requirements and budget requirements. The step of determining recommendation rules based on c preset user needs, the product information, and the historical user profiles includes: Based on the product function information and the product function requirements corresponding to each of the c preset user requirements, a diagnostic product set corresponding to each preset user requirement is determined to obtain c first diagnostic product sets; each first diagnostic product set includes at least one diagnostic product. Based on the non-product function information, determine the price of each diagnostic product in the c first diagnostic product sets; Based on the price of each diagnostic product in the c first diagnostic product sets, the historical spending power tag, and the budget requirement corresponding to each preset user need, the diagnostic products in the c first diagnostic product sets are adjusted to obtain c second diagnostic product sets; each second diagnostic product set includes at least one diagnostic product. Based on the historical interest tags and the product adaptation tags, the diagnostic products in the c second diagnostic product sets are adjusted to obtain c target diagnostic product sets; each preset user requirement corresponds to one target diagnostic product set; each target diagnostic product set includes at least one diagnostic product; the diagnostic products in the c target diagnostic product sets include the a diagnostic products; Determine the mapping relationship between the c preset user needs, the historical spending power tags, the historical interest tags, the product adaptation tags, and the c target diagnostic product sets, and determine the mapping relationship as the recommendation rule.

4. The method according to claim 3, characterized in that, The step involves classifying the *a* diagnostic products based on the product information, the historical user profiles, and the recommendation rules to obtain *a* classification tag sets, including: Based on the recommendation rules, a preset user demand set corresponding to each of the a diagnostic products is determined, resulting in a preset user demand set; each preset user demand set includes at least one preset user demand. Based on the product information and the a preset user demand sets, determine the functional category tag set and price category tag set corresponding to each of the a diagnostic products, thus obtaining a functional category tag set and a price category tag set; each functional category tag set includes at least one functional category tag. Based on the historical user profiles and the a preset user demand sets, determine the user category tag set corresponding to each of the a diagnostic products, and obtain a user category tag sets; each user category tag set includes at least one user category tag. Based on the a preset user demand sets, the a functional category tag sets, the a price category tags, and the a user category tag sets, determine the a category tag sets.

5. The method according to claim 4, characterized in that, The user preference information includes: target product feature requirements and target budget requirements, historically used products, and products of interest; The step of determining the target category tag set from the a category tag sets based on the user preference information includes: The target product functional requirements and the target budget requirements are matched with the a preset user requirement sets to obtain d preset user requirement sets; the d preset user requirement sets correspond one-to-one with the d first candidate category tag sets; d is a positive integer less than or equal to a; Based on the target budget requirement and the d price category labels corresponding to the d first candidate category label sets, determine e second candidate category label sets; e is a positive integer less than or equal to d; Determine the matching degree between the product functional requirements and each functional category label in the e second candidate functional category label sets, and determine the number of labels in each second candidate functional category label set whose matching degree with the product functional requirements is greater than a first threshold, thus obtaining e first quantities; the e second candidate functional category label sets correspond one-to-one with the e second candidate functional category label sets; each second candidate functional category label set corresponds to one first quantity; Based on the e first quantities, f third candidate classification label sets are determined from the e second candidate classification label sets; f is a positive integer less than or equal to e; The target category tag set is determined based on the historically used products, the products of interest, and the f user category tag sets corresponding to the f third candidate category tag sets.

6. The method according to claim 5, characterized in that, Each user category tag set includes the historical interest tags; The step of determining the target category tag set based on the historically used products, the products of interest, and the f user category tag sets corresponding to the f third candidate category tag sets includes: The historically used products and the products of interest are categorized to obtain interest tags; the interest tags include all tags corresponding to the historically used products and the products of interest. Determine the matching degree between the interest tag and each of the f historical interest tags corresponding to the f user category tag sets, and determine the category tag set in the f third candidate category tag sets whose matching degree with the interest tag is greater than a third threshold as the target category tag set.

7. The method according to claim 6, characterized in that, The step of outputting relevant recommendations for the target diagnostic product corresponding to the target classification tag set to the user through the preset large language model includes: Determine the target product functional information and target non-product functional information in the product data that correspond to the target diagnostic product; Based on the target category tag set and the historical interest tags, determine the user's interest content; The preset large language model outputs relevant recommended content to the user, corresponding to the target category tag set, the target product function information, the target non-product function information, and the content of interest.

8. An intelligent recommendation device, characterized in that, The device includes: The acquisition unit is used to acquire historical user data and product data corresponding to a diagnostic products; a is a positive integer. The processing unit is used to construct a historical user profile based on b preset usage scenarios and the historical user data; b is a positive integer. Recommendation rules are determined based on c preset user needs, the product information, and the historical user profiles; c is a positive integer. Based on the product information, the historical user profile, and the recommendation rules, the a diagnostic products are classified to obtain a set of category tags; each diagnostic product corresponds to one set of category tags; each set of category tags includes at least one category tag. The acquisition unit is used to acquire the interaction content between the user and the preset large language model; The processing unit is used to determine user preference information based on the interaction content; Based on the user preference information, determine the target category tag set from the a category tag sets; The preset large language model outputs relevant recommendations for target diagnostic products to the user, which correspond to the target classification tag set.

9. An electronic device, characterized in that, include: A processor and a memory, the processor being connected to the memory, the memory being used to store a computer program, and the processor being used to execute the computer program stored in the memory to cause the electronic device to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is executed by a processor to implement the method as described in any one of claims 1-7.

Images