Product recommendation method and apparatus, and medium and electronic device

Abstract

A product recommendation method and apparatus, and a medium and an electronic device. The method comprises: in response to an intelligent Internet-of-Things device scanning a product label, determining a target product to be purchased and a target user associated with the intelligent Internet-of-Things device (S110); on the basis of a geographic location associated with the intelligent Internet-of-Things device, determining a target store where the target product is located, querying in the target store for an associated product of a target product type, and taking the target product and the associated product as candidate products (S120); acquiring user description information of the target user (S130); querying product description information of the candidate products by means of a product recommendation model, and screening the candidate products and the product description information of the candidate products on the basis of the user description information of the target user, so as to determine a recommended product and product description information of the recommended product (S140); and on the basis of the product description information of the recommended product, generating recommendation description information for the recommended product (S150).

Description

Commodity recommendation method, device, medium and electronic device

[0001] The present application claims priority to the Chinese patent application No. 202410822091.0, filed on June 24, 2024, with the Chinese Patent Office, the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application relates to the technical field of artificial intelligence, deep learning and commodity recommendation, for example, to a commodity recommendation method, device, medium and electronic device. BACKGROUND

[0003] Commodity recommendation is widely used in online sales channels such as e-commerce websites, and is generally based on certain rules to recommend commodities, for example, according to the target product searched by the user, the commodity is recommended according to the sales volume, price and repurchase rate, etc. To a certain extent, it can improve the effective conversion rate of commodity sales of e-commerce websites and increase commodity sales. However, the rules are relatively fixed, and the personalization of commodity recommendation is insufficient.

[0004] For offline sales channels such as physical stores, there is a lack of effective personalized commodity recommendation. The commodity description information in physical stores is mostly static presentation, and the user needs to actively search for information and compare information by himself. It cannot provide personalized commodity recommendation for users, which affects the user shopping experience and limits the marketing potential of commodities. SUMMARY

[0005] The present application provides a commodity recommendation method, device, medium and electronic device, which can improve the user shopping experience and release the marketing potential of commodities.

[0006] The present application provides a commodity recommendation method, which comprises:

[0007] In response to the scanning of the commodity tag by the intelligent Internet of Things device to determine the target commodity to be selected and the target user associated with the intelligent Internet of Things device;

[0008] Determine the target store where the target commodity is located based on the geographic location associated with the intelligent Internet of Things device, and query the associated commodities of the target commodity in the target store, and take the target commodity and the associated commodities as candidate commodities;

[0009] Obtain the user description information of the target user;

[0010] Query the commodity description information of the candidate commodities through a commodity recommendation model, and filter the candidate commodities and the commodity description information of the candidate commodities based on the user description information of the target user, to determine the recommended commodities and the commodity description information of the recommended commodities;

[0011] generate recommendation description information for the recommended commodity according to the commodity description information of the recommended commodity; wherein the commodity recommendation model is a pre-trained generative artificial intelligence model.

[0012] Embodiments of the present application provide a commodity recommendation device, the device comprising:

[0013] a commodity and user determination module configured to determine a target commodity to be selected and a target user associated with the intelligent Internet of Things device in response to the intelligent Internet of Things device scanning a commodity tag;

[0014] a store and commodity determination module configured to determine a target store where the target commodity is located based on a geographic location associated with the intelligent Internet of Things device, and query associated commodities of the target commodity at the target store, and take the target commodity and the associated commodities as candidate commodities;

[0015] a user description information determination module configured to obtain user description information of the target user;

[0016] a commodity and description information screening module configured to query commodity description information of the candidate commodities through a commodity recommendation model, and screen the candidate commodities and the commodity description information of the candidate commodities based on the user description information of the target user, to determine recommended commodities and commodity description information of the recommended commodities;

[0017] a recommendation description information generation module configured to generate recommendation description information for the recommended commodities according to the commodity description information of the recommended commodities; wherein the commodity recommendation model is a pre-trained generative artificial intelligence model.

[0018] Embodiments of the present application provide a computer readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the commodity recommendation method as described in embodiments of the present application.

[0019] Embodiments of the present application provide an electronic device comprising a memory, a processor, and a computer program stored on the memory and executable by the processor, the processor executing the computer program to implement the commodity recommendation method as described in embodiments of the present application.

[0020] Embodiments of the present application provide a computer program product comprising a computer program, the computer program being executed by a processor to implement the commodity recommendation method as described in embodiments of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0021] FIG. 1 is a flowchart of a commodity recommendation method according to Embodiment One;

[0022] FIG. 2A is a flow chart of a method for recommending a commodity according to an embodiment;

[0023] FIG. 2B is a structural diagram of a commodity recommendation system according to an embodiment of the present application;

[0024] FIG. 3 is a structural diagram of a commodity recommendation apparatus according to an embodiment of the present application;

[0025] FIG. 4 is a structural diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work should fall within the protection scope of the present application.

[0027] The terms "first", "second", "target" and "candidate" and the like in the specification and claims of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] Embodiment One

[0029] FIG. 1 is a flow chart of a method for recommending a commodity according to an embodiment, which can be applied to the case of recommending a commodity in an offline sales scenario. The method can be configured to be executed by a commodity recommendation apparatus, which is realized in the form of hardware and / or software and can be integrated into an electronic device running the system.

[0030] As shown in FIG. 1, the method comprises:

[0031] S110, determining a target commodity to be selected and a target user associated with the smart Internet of Things device in response to the smart Internet of Things device scanning a commodity label.

[0032] S120, determine a target store where the target commodity is located based on the geographic location associated with the intelligent Internet of Things device, and query associated commodities of the target commodity in the target store, and take the target commodity and the associated commodities as candidate commodities.

[0033] S130, obtain user description information of the target user.

[0034] S140, query commodity description information of the candidate commodities through a commodity recommendation model, and filter the candidate commodities and the commodity description information of the candidate commodities based on the user description information of the target user, to determine a recommended commodity and commodity description information of the recommended commodity.

[0035] S150, generate recommendation description information for the recommended commodity according to the commodity description information of the recommended commodity.

[0036] The commodity recommendation model is a pre-trained generative artificial intelligence (GenAI) model.

[0037] An intelligent Internet of Things (AIoT) device refers to an intelligent device that is interconnected and works cooperatively through a network. Optionally, the intelligent Internet of Things device can be an intelligent shopping cart, an in-store robot, or a mobile terminal serving an entity store, etc. The intelligent Internet of Things device has a commodity label scanning capability.

[0038] Optionally, the commodity label can be an electronic shelf label (ESL) or a liquid crystal electronic shelf label (LCD-ESL), and the LCD-ESL label refers to an electronic shelf label with an LCD screen. The commodity label supports near field communication (NFC) technology or supports a scanning function of a two-dimensional code / barcode. The commodity label is configured to display commodity label information. The commodity label information refers to content required by the state to be marked on the commodity. Exemplarily, the commodity label information at least includes commodity name, qualification certificate, production address, commodity specification, warning instructions, and use period, etc.

[0039] Based on the commodity label information obtained by the intelligent Internet of Things device through scanning of the commodity label, a target commodity to be selected and purchased can be determined. The target commodity to be selected and purchased refers to a commodity that the target user has an intention to select and purchase. The target user refers to a user who operates the intelligent Internet of Things device to scan the commodity label.

[0040] The intelligent Internet of Things device generally has a determined service place, wherein the target store refers to the service place of the intelligent Internet of Things device. The target store has a target commodity for sale.

[0041] The associated commodity refers to a commodity for sale in the target store that is associated with the target commodity. The association between the associated commodity and the target commodity can be determined by the target store according to example business requirements, which is not limited here. For example, the associated commodity can be of the same type as the target commodity or of a different type from the target commodity. For example, the target commodity is coffee, and the associated commodity can be coffee or a coffee cup. If the associated commodity is of the same type as the target commodity, the associated commodity and the target commodity can be different in style, origin, and manufacturer, etc.

[0042] The candidate commodity refers to a commodity that the target user can have a purchase intention. The candidate commodity includes not only the target commodity but also the associated commodity.

[0043] The user description information of the target user is used to describe the consumption behavior of the target user. Optionally, the user description information of the target user includes the browsing habits, purchase history, personal preferences, and identity information of the target user, etc. The user description information of the target user is obtained under the condition that the collection authorization credential of the user description information is obtained. That is, the user description information of the target user is obtained under the explicit authorization of the target user. Optionally, the user description information of the target user is obtained when the target user requests to use the intelligent Internet of Things device.

[0044] Optionally, the commodity description information of the candidate commodity can describe the candidate commodity from at least two description dimensions of a commodity label dimension and a user feedback dimension. The user feedback dimension can cover user browsing habits, user purchase history, and user personal preferences, etc. For example, the user feedback dimension includes a health dimension, a carpet dimension, and commodity sales. This is because the user purchase history in the user feedback dimension can reflect the sales of the commodity.

[0045] Optionally, the commodity parameters and user parameters in the recommendation guide dialogue template of the commodity recommendation model are instantiated using the candidate commodity and the user description information of the target user. The instantiated recommendation guide dialogue is input into the commodity recommendation model to enable the commodity recommendation model to use the recommendation guide dialogue to perform personalized commodity recommendation for the target user based on the user description information of the target user. The commodity recommendation model is a pre-trained generative artificial intelligence model, which has semantic understanding ability and text generation ability.

[0046] The commodity recommendation model queries the commodity description information of the candidate commodities. As described above, the commodity description information includes at least two description dimensions. Optionally, the at least two description dimensions included in the commodity description information include a commodity label dimension and a user feedback dimension.

[0047] Under each description dimension, other sub-classes can also be subdivided. Optionally, the commodity recommendation model determines the key dimension of interest to the target user based on the user description information of the target user. By comparing the candidate commodities under the key dimension, the recommended commodity can be determined from the candidate commodities. Optionally, the commodity description information belonging to the key dimension is determined as the commodity description information of the recommended commodity. The commodity recommendation model generates recommendation description information for the recommended commodity according to the commodity description information of the recommended commodity. Optionally, the commodity recommendation model processes the commodity description information of the recommended commodity to obtain the recommendation description information of the recommended commodity. The commodity recommendation model processing the commodity description information of the recommended commodity can include adjusting the data modality.

[0048] Optionally, the recommendation description information of the recommended commodity is preferentially displayed, and the commodity description information of the recommended commodity is displayed after the recommendation description information.

[0049] The recommended commodity and the recommendation description information of the recommended commodity are strongly related to the user description information of the target user. The user description information of different users is different, and the key dimension of interest can be different. For example, the key dimension of interest of user A is the commodity price, which is a price-sensitive user, and the recommended commodity is selected from the candidate commodities by comparing the price dimension, and the commodity price of the recommended commodity is preferentially displayed. The key dimension of interest of user B is the commodity ingredient, which is a health-sensitive user, and the recommended commodity is selected from the candidate commodities by comparing the ingredient dimension, and the commodity ingredient of the recommended commodity is preferentially displayed. That is, when user A and user B use the intelligent Internet of Things device to scan the commodity label of the same commodity, the commodities recommended to user A and user B can be different, and the commodity description information displayed to user A and user B will also be different.

[0050] The technical scheme of the application applies the commodity recommendation model to an offline sales scene, filters based on user description information of a target user and commodity description information of a candidate commodity through the commodity recommendation model to determine a recommended commodity and commodity description information of the recommended commodity, and generates recommendation description information for the recommended commodity according to the commodity description information of the recommended commodity. In the application, the recommended commodity and the commodity description information of the recommended commodity are strongly related to the user description information of the target user, can best meet the preferences and needs of the target user, and truly achieve one-to-one commodity recommendation, realize personalized commodity recommendation in the offline sales scene, enable the target user to scan a commodity tag of a target commodity by using an intelligent Internet of Things device, know associated commodities of the target commodity in a store, and know the recommended commodity and commodity description information of the commodity of interest, improve the interactivity of user shopping, enrich the user shopping experience, improve the degree of personalization of commodity recommendation, release the commodity marketing potential, and the like.

[0051] In an optional embodiment, the user description information of the target user is acquired by: in response to a device use request of the intelligent Internet of Things device, acquiring user login data associated with the intelligent Internet of Things device; if the user login data includes a collection authorization credential of the user description information, acquiring consumption behavior data of the target user as the user description information of the target user; if the user login data does not include the collection authorization credential of the user description information, determining an information collection area with the target store as a regional center based on a geographical position associated with the intelligent Internet of Things device and a preset information collection range; determining consumption behavior data of a consumption group in the information collection area based on public consumption behavior data in the information collection area and private consumption behavior data in the target store; and taking the consumption behavior data of the consumption group in the information collection area as the user description information of the target user.

[0052] The device use request is used to request to use the smart internet-of-things device for commodity purchase. Optionally, a use request control is arranged on the smart internet-of-things device, wherein the use request control can be a physical key or a virtual key. In response to the use request control being clicked or pressed, a user login page is displayed, and a target user can input user login information on the user login page. The target user login information is used to identify the login identity of the target user, and then determine the granularity of the user description information. The login identity of the target user includes an authenticated user and a visitor user. If the user login information of the target user includes a collection authorization credential, the target user is an authenticated user. If the user login information of the target user does not include a collection authorization credential, the target user is a visitor user. The collection authorization credential refers to a credential that the target user allows access to the user description information thereof and allows the user description information thereof to be used for commodity recommendation for the target user. For the authenticated user, the granularity of the user description information thereof is smaller than the granularity of the user description information of the visitor user. The granularity of the user description information of the authenticated user is personal, and the granularity of the user description information of the visitor user is group. That is, the authenticated user can obtain more accurate commodity recommendation than the visitor user.

[0053] The user description information of the authenticated user is consumption behavior data of the authenticated user. Exemplarily, the consumption behavior data includes browsing habits, purchase history, personal preferences, identity information, and the like. The user description information of the visitor user needs to be determined according to the consumption behavior data of a consumption group in an information collection area. The consumption behavior data of the consumption group has regional characteristics and can include user shopping preferences, shopping frequency, and commodity categories in the information collection area. The consumption behavior data of the consumption group includes publicly disclosed consumption behavior data in the information collection area and privately owned consumption behavior data in the target store.

[0054] The publicly disclosed consumption behavior data is consumption behavior data that is publicly disclosed and available to the public. The privately owned consumption behavior data in the target store refers to consumption behavior data that is not publicly disclosed and is owned by the target store.

[0055] The information collection area refers to an area in which consumption behavior data is collected with the target store as the regional center. The information collection area is determined according to a geographic location associated with the smart internet-of-things device and a preset information collection range. The geographic location associated with the smart internet-of-things device is used to determine the regional center of the information collection area. The preset information collection range is used to determine the regional range of the information collection area. Exemplarily, the preset information collection range can be determined by a regional radius, for example, an area within 5 kilometers from the target store as the regional center is determined as the information collection area. The preset information collection range is determined according to actual business needs, which is not limited here.

[0056] The technical solution provides a feasible user description information acquisition scheme, provides data support for using the user description information for commodity recommendation, and can obtain the user description information when the collection authorization credential of the user description information is obtained, and can still determine the user description information of the target user by means of the consumption behavior data of the consumption group when the collection authorization credential of the user description information is not obtained, thereby improving the usability of the commodity recommendation method.

[0057] In an optional embodiment, the method further comprises: acquiring, by the intelligent Internet of Things device, demand supplement information of the target user for the target commodity; querying, by the commodity recommendation model, a matching commodity based on commodity description information of the target commodity and the demand supplement information; and generating, by the commodity recommendation model, a commodity matching reason based on the commodity description information of the target commodity, the user description information of the target user, and commodity description information of the matching commodity.

[0058] The demand supplement information refers to some personalized demands, expectations or preferences additionally provided by the target user.

[0059] The matching commodity refers to a commodity on sale in the target store and capable of meeting the demands of the target user by matching the target commodity. The commodity recommendation model queries the matching commodity based on the commodity description information of the target commodity, the user description information of the target user, and the demand supplement information.

[0060] The commodity matching reason is related to the demand supplement information, and is used to give a reason for selecting the matching commodity from the commodity description information of the target commodity and the demand supplement information. For example, the target commodity is steak, and the demand supplement information of the target user for the target commodity is “today is the wedding anniversary, and I want to have a sumptuous dinner.” The matching commodity queried by the commodity recommendation model can be red wine. The commodity recommendation model can understand the characteristics of different wines, such as wine body, acidity, sweetness, tannin content, etc., thereby providing a scientific and taste-coordinated matching suggestion. The commodity recommendation model also considers the shopping preferences of the target user, and performs personalized recommendation based on the purchase history and commodity evaluation of the target user, to generate the commodity matching reason.

[0061] The technical solution supports the target user to supplement the demands for the target commodity through the intelligent Internet of Things device, fully utilizes the language understanding capability and text generation capability of the commodity recommendation model, generates the commodity matching reason matched with the demand supplement information based on the commodity description information of the target commodity, the user description information of the target user, and the commodity description information of the matching commodity, improves the interactivity of user shopping, enriches the user shopping experience, releases the commodity marketing potential, and stimulates the commodity purchase interest of the user.

[0062] In an optional embodiment, the method further comprises: if the type to which the target commodity belongs is a first type, extracting, by the commodity recommendation model, recommendation reference information related to the first type from the user description information of the target user based on the first type; and generating, by the commodity recommendation model, use prompt information of the target commodity based on the recommendation reference information and commodity description information of the target commodity.

[0063] The first type refers to a commodity type that needs special prompt information. The first type is determined according to actual business needs, which is not limited here. For example, the first type can be medicine, health care products, or food with a high allergy rate.

[0064] The recommendation reference information is used to generate the use prompt information of the target commodity. The recommendation reference information is generated from the user description information, and the recommendation reference information is related to the first type.

[0065] If the type to which the target commodity belongs is the first type, it means that the commodity recommendation model needs to extract recommendation reference information related to the first type from the user description information of the target user based on the first type. For example, if the first type is food with a high allergy rate, the food allergy record of the target user is extracted from the user description information of the target user.

[0066] The commodity recommendation model generates the use prompt information of the target commodity based on the recommendation reference information and the commodity description information of the target commodity. The use prompt information is used to prompt the target user of potential risks in using the target commodity and to show the target user the use precautions of the target commodity.

[0067] The above technical solution sets the first type. In the case where the target commodity is of the first type, the commodity recommendation model generates the use prompt information of the target commodity based on the commodity description information of the target commodity and the recommendation reference information related to the first type in the user description information. This can help the target user make a safer commodity selection and improve the user shopping experience.

[0068] In an optional embodiment, the method further comprises: generating recommendation description information for the recommended commodities according to commodity description information of the recommended commodities, including: performing horizontal comparison on the commodity description information of the recommended commodities according to the key dimensions to obtain dimension comparison descriptions between the recommended commodities; performing vertical comparison on the commodity description information of the recommended commodities according to a time dimension to obtain trend comparison descriptions of the recommended commodities themselves; and generating recommendation description information for the recommended commodities based on the dimension comparison descriptions and the trend comparison descriptions; wherein a data modality corresponding to the recommendation description information includes at least one of a chart modality, a text modality, or a sound modality.

[0069] Optionally, the recommended commodities are at least two, and the target user can make a selection among the at least two recommended commodities. The key dimension refers to a description dimension of interest to the target user. If the key dimension is at least two, the commodity description information of the recommended commodities is compared horizontally under each key dimension. The horizontal comparison is relative to the vertical comparison in the time dimension, and the comparison of the commodity description information of the recommended commodities is made at a certain time point under the key dimension.

[0070] The dimension comparison result refers to the comparison result of the commodity description information between the recommended commodities. For example, the key dimension is the commodity price and the commodity ingredient, and the dimension comparison result is the price comparison result and the ingredient comparison result.

[0071] The commodity description information of each recommended commodity is compared vertically according to the time dimension, the change trend of the commodity description information of each recommended commodity over time is determined, and the trend comparison description of the recommended commodity itself is obtained.

[0072] The dimension comparison result can reflect the individual difference of different recommended commodities in the key dimension, and the trend comparison result can reflect the self-difference of each recommended commodity in the time dimension. Optionally, the commodity recommendation model generates commodity recommendation information for the recommended commodities based on the dimension comparison description and the trend comparison description. The commodity recommendation information is used to provide a reference for the target user to select and purchase commodities.

[0073] Optionally, the commodity recommendation description can be represented in multiple data modalities. The data modality corresponding to the recommendation description information includes at least one of a chart modality, a text modality, or a sound modality. The initial modality of the commodity recommendation description can be a text modality. A text-to-speech conversion technology is used to convert the commodity recommendation information in the text modality into commodity description information in the sound modality. A text-to-chart conversion technology is used to convert the commodity recommendation information in the text modality into commodity description information in the chart modality.

[0074] Optionally, the intelligent Internet of Things device is configured with a display device capable of displaying text charts and a sound producing device capable of playing sounds.

[0075] If the commodity recommendation information is in the chart modality or the text modality, the display device of the intelligent Internet of Things device is controlled to display the commodity recommendation information; if the commodity recommendation information is in the sound modality, the sound producing device of the intelligent Internet of Things device is controlled to broadcast the commodity recommendation information.

[0076] The technical solution above, in the case of recommending commodities and determining commodity description information of the recommended commodities, respectively performs horizontal comparison and vertical comparison on the recommended commodities from a key dimension and a time dimension, obtains individual differences of different recommended commodities in the key dimension and self-differences of each recommended commodity in the time dimension, and displays the differences to a target user, thereby providing a reference for the target user to select and purchase commodities, assisting the target user to make more accurate consumption decisions, and being beneficial to improving user shopping experience.

[0077] Embodiment Two

[0078] FIG. 2A is a flowchart of a commodity recommendation method provided according to Embodiment Two. This embodiment is described based on the above-described embodiments.

[0079] As shown in FIG. 2A, the method comprises:

[0080] S210, determining a target commodity to be selected and purchased and a target user associated with the intelligent Internet-of-Things device in response to the intelligent Internet-of-Things device scanning a commodity tag.

[0081] S220, determining a target store where the target commodity is located based on a geographic location associated with the intelligent Internet-of-Things device, and querying associated commodities of the target commodity in the target store, taking the target commodity and the associated commodities as candidate commodities.

[0082] S230, obtaining user description information of the target user.

[0083] S240, querying commodity description information stored in a commodity description database by using a commodity number of the candidate commodity as a query index through the commodity recommendation model.

[0084] The commodity description information of the candidate commodities is pre-stored in the commodity description database. Optionally, the commodity description database is a cloud database. Optionally, a cache technology such as Redis is used to store frequently accessed commodity description data to improve query efficiency. Optionally, the user description information and the commodity description information are both encrypted in the transmission and storage process, so as to fully protect user privacy.

[0085] The commodity description information and the user description information can simultaneously include structured data and unstructured data. Optionally, a dual-database storage strategy is adopted, structured data such as identity information or purchase history in the user description information or commodity price in the commodity description information is stored in a relational database, and unstructured data such as browsing habits and personal preferences in the user description information or user evaluation in the commodity description information is stored in an unstructured or semi-structured database. The dual-database storage strategy can optimize the flexibility of data storage, improve query efficiency, and ensure that various types of data can be processed.

[0086] The query index is used to identify different candidate commodities in the commodity description database. Optionally, the commodities stored in the commodity database are uniformly coded, and the commodity code is taken as the query index of the candidate commodity. The commodity description information is the data content corresponding to the query index.

[0087] S250, determining, by the commodity recommendation model, a key dimension of interest of the target user based on the user description information of the target user and the commodity description information corresponding to the at least two description dimensions.

[0088] The commodity recommendation model has semantic understanding and reasoning ability. The commodity recommendation model performs semantic understanding on the user description information of the target user and the commodity description information corresponding to the at least two description dimensions respectively, and determines the key dimension of interest of the target user based on the semantic understanding result. For example, the user preference included in the user description information is that the user likes litchi. The commodity description information corresponding to the commodity indication dimension includes commodity ingredients, and the content of the commodity ingredients includes litchi juice and juice content. Then, it is determined that the key dimension of interest of the target user is litchi content.

[0089] S260, performing semantic matching between the user description information of the target user and the commodity description information of the candidate commodity under the key dimension, and determining the recommended commodity from the candidate commodity according to the obtained semantic matching result.

[0090] In the case of determining the key dimension, the user description information of the target user is semantically matched with the commodity description information under the key dimension, to obtain the semantic similarity between the user description information of the target user and the commodity description information of the candidate commodity.

[0091] Optionally, a second number of description dimensions are selected as the key dimension based on the order from large to small according to the semantic similarity. The second number is used to determine the number of recommended commodities, and the second number is determined according to actual business requirements, which is not limited here.

[0092] S270, determining, as the commodity description information of the recommended commodity, the commodity description information belonging to the key dimension in the commodity description information of the recommended commodity.

[0093] The commodity description information includes at least two description dimensions of commodity indication dimension and user feedback dimension.

[0094] The commodity identification dimension refers to the content required by the state to be marked on the commodity. For example, the commodity description information under the commodity identification dimension can be the commodity name, the qualification certificate, the production address, the commodity specification, the warning instruction, and the service life, etc. The commodity sales dimension refers to the content related to the commodity sales activities. The user feedback dimension refers to the content related to the user experience. For example, the user feedback dimension includes the user evaluation, the user classification, and the feedback channel, etc. The user feedback dimension can cover the user browsing habit, the user purchase history, and the user personal preference, etc. For example, the user feedback dimension includes the health dimension, the carpet dimension, and the commodity sales volume. This is because the user purchase history in the user feedback dimension can reflect the commodity sales situation.

[0095] S280, generating recommendation description information for the recommended commodity according to the commodity description information of the recommended commodity.

[0096] The technical solution of the present application uses the commodity recommendation model in the offline sales scene, and uses the commodity number of the candidate commodity as a query index to query the commodity description information stored in the commodity description database through the commodity recommendation model. The commodity recommendation model determines the key dimension of interest of the target user based on the user description information of the target user and the commodity description information corresponding to at least two description dimensions; the user description information of the target user and the commodity description information of the candidate commodity are semantically matched under the key dimension, and the recommended commodity is determined from the candidate commodity according to the obtained semantic matching result. The semantic understanding and reasoning ability of the commodity recommendation model are fully utilized, which provides technical support for realizing personalized commodity recommendation in the offline sales scene, is conducive to enriching the user shopping experience, is conducive to improving the personalization degree of commodity recommendation, and is conducive to releasing the commodity marketing potential.

[0097] In an optional embodiment, the commodity description information belonging to the user feedback dimension is obtained by: collecting user evaluation data about the target commodity from at least two data sources, and determining the feedback user corresponding to the user evaluation data; performing semantic analysis on the user evaluation data through the commodity recommendation model to determine the emotional tendency of the feedback user to the target commodity; and taking the emotional tendency of the feedback user to the target commodity, the user description information of the feedback user, and the user evaluation data corresponding to the feedback user as the commodity description information of the user feedback dimension.

[0098] Optionally, the commodity description information and the user description information are collected from multiple data sources. For example, the data sources include but are not limited to social media, online comments, retail platforms, and other non-traditional data sources.

[0099] The user evaluation data is used to reflect the subjective experience of the user using the target commodity. Optionally, the user evaluation data about the target commodity is searched from the user evaluation data included in at least two data sources. The feedback user is a user who publishes product use experience. The feedback user corresponds to the user evaluation data.

[0100] The user evaluation data is analyzed by the commodity recommendation model to determine the emotional tendency of the feedback user to the target commodity. The emotional tendency includes negative and positive. Optionally, the emotional tendency is negative, indicating that the feedback user has a poor experience of using the target commodity, and the emotional tendency is positive, indicating that the feedback user has a good experience of using the target commodity.

[0101] The emotional tendency of the feedback user to the target commodity, the user description information of the feedback user, and the user evaluation data corresponding to the feedback user are used as the commodity description information of the user feedback dimension. In this way, in the case that the user feedback dimension is the key dimension, the target user's user description information can be matched with the commodity description information of the candidate commodity to determine the recommended commodity from the candidate commodity.

[0102] The above technical solution provides a practical commodity description information determination scheme for determining commodity description information belonging to the user feedback dimension. Data support is provided for using commodity description information belonging to the user feedback dimension for personalized commodity recommendation for a target user.

[0103] In one embodiment, a commodity recommendation system is provided. The structural diagram of the commodity recommendation system is shown in FIG. 2B.

[0104] Referring to FIG. 2B, the commodity recommendation system includes a commodity label, an intelligent Internet of Things (AloT) device, an electronic price tag controller, an intelligent interactive cloud service system, and a GenAI (commodity recommendation model).

[0105] The intelligent interactive cloud service system is responsible for commodity description information collection, user description information collection, data modal conversion of recommendation description data, and data communication between the GenAI and the electronic price tag controller and the intelligent Internet of Things device.

[0106] The GenAI has data analysis, regional behavior analysis, and user analysis capabilities. Data analysis refers to analyzing commodity description data, regional behavior analysis refers to analyzing consumer behavior data of a consumer group in an information collection area in the case that the login role of a target user is a tourist user, and user analysis refers to analyzing user description data.

[0107] The commodity label can be an ESL or an LCD ESL. The smart Internet of Things device can be a smart shopping cart, a mobile terminal, or a store robot. The electronic price tag controller serves as a signal transmission core and is responsible for managing data communication between the commodity label and the GenAI.

[0108] The smart Internet of Things device supports displaying interactive content generated by the GenAI (commodity recommendation model) through an interface, such as text and charts, and also supports playing voice content generated by the GenAI (commodity recommendation model) through voice.

[0109] The target user touches or scans a two-dimensional code / bar code on the ESL label through the smart Internet of Things device, triggering the acquisition of the following information:

[0110] The target commodity to be selected and the target user associated with the smart Internet of Things device are determined; the target store where the target commodity is located is determined based on the geographic location associated with the smart Internet of Things device, and associated commodities of the target commodity type are queried in the target store, and the target commodity and the associated commodities are taken as candidate commodities; and the user description information of the target user is acquired.

[0111] The electronic price tag controller sends the candidate commodities and the user description of the target user to the GenAI through the smart interaction cloud service system. The GenAI queries the commodity description information of the candidate commodities, and calls the data analysis capability, regional behavior analysis capability, and user analysis capability possessed by the GenAI, filters the candidate commodities and the commodity description information of the candidate commodities based on the user description information of the target user, determines the recommended commodity and the commodity description information of the recommended commodity from the candidate commodities, and then generates the recommendation description information of the recommended commodity according to the commodity description information of the recommended commodity.

[0112] Then, the recommended commodity and the recommendation description information of the recommended commodity are sent to the smart Internet of Things device through the smart interaction cloud service system, so that the smart Internet of Things device displays the interactive content generated by the GenAI (commodity recommendation model) through an interface, such as text and charts, and plays the voice content generated by the GenAI (commodity recommendation model) through voice.

[0113] Embodiment Three

[0114] FIG. 3 is a structural schematic diagram of a commodity recommendation device provided by Embodiment Three of the present application. The present embodiment can be applicable to the case of commodity recommendation in an offline sales scene. The device can be realized by software and / or hardware, and can be integrated into an electronic device such as a smart terminal.

[0115] As shown in FIG. 3, the device can include:

[0116] The commodity and user determination module 310 is configured to determine target commodities to be selected and a target user associated with the smart Internet of Things device in response to the smart Internet of Things device scanning a commodity label.

[0117] The store and commodity determination module 320 is configured to determine a target store where the target commodity is located based on a geographic location associated with the smart Internet of Things device, and query associated commodities of the target commodity in the target store, and take the target commodity and the associated commodities as candidate commodities.

[0118] The user description information determination module 330 is configured to obtain user description information of the target user.

[0119] The commodity and description information screening module 340 is configured to query commodity description information of the candidate commodities through a commodity recommendation model, and screen the candidate commodities and the commodity description information of the candidate commodities based on the user description information of the target user, to determine recommended commodities and commodity description information of the recommended commodities.

[0120] The recommended description information generation module 350 is configured to generate recommended description information for the recommended commodities according to the commodity description information of the recommended commodities; and the commodity recommendation model is a pre-trained generative artificial intelligence model.

[0121] The technical scheme of the present application applies the commodity recommendation model to an offline sales scenario, screens the candidate commodities and the commodity description information of the candidate commodities based on user description information of a target user through the commodity recommendation model, to determine recommended commodities and commodity description information of the recommended commodities, and generates recommended description information for the recommended commodities according to the commodity description information of the recommended commodities. In the present application, the recommended commodities and the commodity description information of the recommended commodities are strongly related to the user description information of the target user, and can best meet the preferences and needs of the target user, truly realizing that thousands of people have thousands of faces in commodity recommendation, and realizing personalized commodity recommendation in an offline sales scenario. The target user can know associated commodities of the target commodity in a store and can know recommended commodities and commodity description information of interest to the target user by scanning a commodity label of the target commodity with a smart Internet of Things device, improving the interactivity of user shopping, enriching the user shopping experience, improving the degree of personalization of commodity recommendation, releasing the potential of commodity marketing, and the like.

[0122] Optionally, the commodity and description information screening module 340 comprises: a description information query submodule configured to query commodity description information stored in a commodity description database by using a commodity number of the candidate commodity as a query index through the commodity recommendation model; a key dimension determination submodule configured to determine a key dimension of interest of the target user based on user description information of the target user and commodity description information corresponding to at least two description dimensions through the commodity recommendation model; a recommended commodity determination submodule configured to perform semantic matching between the user description information of the target user and commodity description information of the candidate commodity under the key dimension, and determine the recommended commodity from the candidate commodity according to a semantic matching result obtained; and a description information determination submodule configured to determine, as commodity description information of the recommended commodity, commodity description information belonging to the key dimension in the commodity description information of the recommended commodity; wherein the commodity description information comprises at least two description dimensions of a commodity identification dimension and a user feedback dimension.

[0123] Optionally, the commodity description information belonging to the user feedback dimension is obtained by: collecting user evaluation data about the target commodity from at least two data sources, and determining a feedback user corresponding to the user evaluation data; performing semantic analysis on the user evaluation data through the commodity recommendation model, and determining an emotional tendency of the feedback user to the target commodity; and taking the emotional tendency of the feedback user to the target commodity, user description information of the feedback user, and user evaluation data corresponding to the feedback user as commodity description information of the user feedback dimension.

[0124] Optionally, the recommended description information generation module 350 comprises: a dimension comparison submodule configured to perform horizontal comparison of commodity description information of the recommended commodity according to a key dimension, to obtain dimension comparison descriptions between the recommended commodities; a trend comparison submodule configured to perform vertical comparison of commodity description information of the recommended commodity according to a time dimension, to obtain trend comparison descriptions of the recommended commodities themselves; and a recommended information generation module configured to generate recommended description information for the recommended commodities based on the dimension comparison descriptions and the trend comparison descriptions; wherein a data modality corresponding to the recommended description information comprises at least one of a chart modality, a text modality, or a sound modality.

[0125] Optionally, the user description information determining module 330 comprises: a login data obtaining sub-module, configured to obtain user login data associated with the smart internet-of-things device in response to a device use request of the smart internet-of-things device; a first description information determining sub-module, configured to obtain consumption behavior data of a target user as user description information of the target user if the user login data comprises a collection authorization credential of user description information; an information collection area determining sub-module, configured to determine an information collection area with the target store as a regional center based on a geographical location associated with the smart internet-of-things device and a preset information collection range if the user login data does not comprise a collection authorization credential of user description information; a consumption behavior data determining sub-module, configured to determine consumption behavior data of a consumer group in the information collection area based on public consumption behavior data in the information collection area and private consumption behavior data in the target store; and a second description information determining sub-module, configured to take the consumption behavior data of the consumer group in the information collection area as the user description information of the target user.

[0126] Optionally, the apparatus further comprises: a demand supplement information obtaining module, configured to obtain demand supplement information of the target user for the target product through the smart internet-of-things device; a matching product querying module, configured to query a matching product matching the demand of the target user through the product recommendation model based on product description information of the target product and the demand supplement information; and a matching reason generating module, configured to generate a product matching reason matching the demand supplement information through the product recommendation model based on product description information of the target product, user description information of the target user, and product description information of the matching product.

[0127] Optionally, the apparatus further comprises: a recommendation reference information determining module, configured to extract recommendation reference information related to a first type from the user description information of the target user through the product recommendation model based on the first type if the target product belongs to the first type; and a use prompt information generating module, configured to generate use prompt information of the target product through the product recommendation model based on the recommendation reference information and the product description information of the target product.

[0128] The product recommendation apparatus provided by the embodiments of the present application can execute the product recommendation method provided by any of the embodiments of the present application, and has a performance module corresponding to the product recommendation method.

[0129] In the technical solution of the present application, the collection, storage, use, processing, transmission, provision and disclosure of user data involved in the technical solution comply with relevant laws and regulations and do not violate public order and good customs.

[0130] Example Four

[0131] According to the embodiments of the present application, the present application further provides an electronic device, a readable storage medium and a computer program product.

[0132] FIG. 4 shows a structural schematic diagram of an electronic device 410 that can be used to implement the embodiments. The electronic device 410 includes at least one processor 411, and a memory, such as a Read-Only Memory (ROM) 412, a Random Access Memory (RAM) 413, etc., which is communicatively connected to the at least one processor 411, wherein the memory stores a computer program that can be executed by the at least one processor. The processor 411 can perform various appropriate actions and processes according to the computer program stored in the Read-Only Memory (ROM) 412 or the computer program loaded from the storage unit 418 to the Random Access Memory (RAM) 413. In the RAM 413, various programs and data required for the operation of the electronic device 410 can also be stored. The processor 411, the ROM 412, and the RAM 413 are connected to each other through a bus 414. An Input / Output (I / O) interface 415 is also connected to the bus 414.

[0133] Various components in the electronic device 410 are connected to the I / O interface 415, including: an input unit 416, such as a keyboard, a mouse, etc.; an output unit 417, such as various types of displays, a loudspeaker, etc.; a storage unit 418, such as a magnetic disk, an optical disk, etc.; and a communication unit 419, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 419 allows the electronic device 410 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0134] The processor 411 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 411 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various special-purpose Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, a Digital Signal Processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 411 performs various methods and processes described above, such as the commodity recommendation method.

[0135] In some embodiments, the item recommendation method can be implemented as a computer program tangibly embodied in a computer readable storage medium, e.g., storage unit 418. In some embodiments, parts or all of the computer program can be loaded and / or installed onto electronic device 410 via, e.g., ROM 412 and / or communication unit 419. When the computer program is loaded onto RAM 413 and executed by processor 411, one or more steps of the item recommendation method described above can be performed. Alternatively, in other embodiments, processor 411 can be configured to perform the item recommendation method by way of other means (e.g., via firmware).

[0136] Various implementations of the systems and techniques described above can be realized in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on chip (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0137] Computer programs used to implement the methods of the present application can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the computer program, when executed, can implement the functions / acts specified in the flowcharts and / or block diagrams. The computer program can be executed entirely on a machine, partially on a machine, partially on a machine and partially on a remote machine or entirely on a remote machine or server.

[0138] In the context of this application, a computer readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer readable storage medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, a computer readable storage medium can be a machine readable signal medium. Examples of a machine readable signal medium will include one or more of: a carrier wave, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM) or flash memory, an optical fiber, a compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0139] To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having a display device (e.g., a Cathode Ray Tube (CRT) or a Liquid Crystal Display (LCD) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0140] The systems and techniques described herein can be implemented in a computing system that includes a back end component (e.g., as a commodity recommendation server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), blockchain network, and the Internet.

[0141] The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. A server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system, and solves the defects of large management difficulty and weak business scalability in traditional physical host and virtual private server (VPS) services.

[0142] The embodiments of the present application further disclose a computer program product, which comprises a computer program, and the computer program, when executed by a processor, implements the commodity recommendation method provided by any of the embodiments of the present application. The program product and the commodity recommendation method disclosed in the embodiments of the present application belong to the same inventive concept, and thus will not be repeated here.

[0143] It should be understood that the steps shown in the above can be reordered, added, or deleted. For example, the steps described in the present application can be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions of the present application can be achieved, and the present application is not limited herein.

Claims

1. A product recommendation method, comprising: In response to the smart IoT device scanning the product tag to determine the target product to be purchased and the target user associated with the smart IoT device; Based on the geographic location associated with the smart IoT device, the target store where the target product is located is determined, and the related products of the target product are queried in the target store, and the target product and the related products are selected as candidate products; Obtain the user description information of the target user; The product recommendation model queries the product description information of the candidate products, and filters the candidate products and their product description information based on the user description information of the target user to determine the recommended products and their product description information. Based on the product description information of the recommended products, recommendation description information is generated for the recommended products; wherein, the product recommendation model is a pre-trained generative artificial intelligence model.

2. The method according to claim 1, wherein, The step of querying the product description information of the candidate products through a product recommendation model, and filtering the candidate products and their product description information based on the user description information of the target user to determine the recommended products and their product description information, includes: The product recommendation model uses the product ID of the candidate product as a query index to query the product description information stored in the product description database. The product recommendation model determines the key dimensions that the target user is interested in based on the user description information of the target user and the product description information corresponding to at least two description dimensions. Under the key dimensions, the user description information of the target user is semantically matched with the product description information of the candidate products, and the recommended product is determined from the candidate products based on the obtained semantic matching results; The product description information belonging to the key dimension in the product description information of the recommended product is determined as the product description information of the recommended product; The product description information includes at least two descriptive dimensions: product labeling and user feedback.

3. The method according to claim 2, wherein, Product description information belonging to the aforementioned user feedback dimension is obtained through the following methods: Collect user review data about the target product from at least two data sources, and identify the feedback users corresponding to the user review data; The product recommendation model is used to perform semantic analysis on the user review data to determine the user's sentiment towards the target product. The emotional inclination of the feedback user towards the target product, the user description information of the feedback user, and the user evaluation data corresponding to the feedback user are used as the product description information of the user feedback dimension.

4. The method according to claim 2, wherein, Based on the product description information of the recommended products, generate recommendation description information for the recommended products, including: The product description information of the recommended products is compared horizontally according to the key dimensions to obtain the dimensional comparison description between the recommended products; The product description information of the recommended products is compared longitudinally according to the time dimension to obtain the trend comparison description of the recommended products themselves; Based on the dimensional comparison description and the trend comparison description, recommendation description information is generated for the recommended products; The data modality corresponding to the recommended description information includes at least one of chart modality, text modality, or audio modality.

5. The method according to claim 1, wherein, The step of obtaining the user description information of the target user includes: In response to the device usage request of the smart IoT device, obtain the user login data associated with the smart IoT device; In response to the collection authorization credential that includes user description information in the user login data, the consumption behavior data of the target user is obtained as the user description information of the target user; In response to the fact that the user login data does not include the authorization credentials for collecting user description information, an information collection area centered on the target store is determined based on the geographical location associated with the smart IoT device and the preset information collection range. Based on publicly available consumer behavior data in the information collection area and private consumer behavior data in the target store, the consumer behavior data of the consumer group in the information collection area is determined. The consumption behavior data of the consumer groups in the information collection area are used as the user description information of the target users.

6. The method according to claim 1, further comprising: The smart IoT device is used to obtain supplementary information about the target user's needs for the target product. Based on the product description information of the target product and the supplementary demand information, the product recommendation model queries the target store for matching products that match the needs of the target user. The product recommendation model generates product pairing reasons that match the supplementary demand information based on the product description information of the target product, the user description information of the target user, and the product description information of the complementary products.

7. The method according to claim 1, further comprising: In response to the target product belonging to the first type, the product recommendation model extracts recommendation reference information related to the first type from the user description information of the target user based on the first type; The product recommendation model generates usage prompts for the target product based on the recommendation reference information and the product description information of the target product.

8. A product recommendation device, comprising: The product and user identification module is configured to respond to the smart IoT device scanning the product tag to identify the target product to be purchased and the target user associated with the smart IoT device; The store and product determination module is configured to determine the target store where the target product is located based on the geographical location associated with the smart IoT device, and query the related products of the target product in the target store, and select the target product and the related products as candidate products; The user description information determination module is configured to obtain the user description information of the target user; The product and description information filtering module is configured to query the product description information of the candidate products through the product recommendation model, and filter the candidate products and their product description information based on the user description information of the target user, so as to determine the recommended products and their product description information. The recommendation description information generation module is configured to generate recommendation description information for the recommended products based on the product description information of the recommended products; wherein, the product recommendation model is a pre-trained generative artificial intelligence model.

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

10. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein, When the processor executes the computer program, it implements the product recommendation method as described in any one of claims 1-7.

Images