CN121329550A - Product recommendation methods and systems - Google Patents

Abstract

This disclosure discloses a product recommendation method and system, wherein the method includes: generating a first product list based on a first navigation label; the first navigation label represents a navigation label determined from multiple navigation labels based on user operations; updating the first product list based on first interaction information and the first navigation label to obtain a second product list; the first interaction information represents information input by the user through an interaction window provided by an intelligent agent.

Description

Commodity recommendation method and system

Technical Field

The present disclosure relates to, but not limited to, the field of computer technology, and in particular, to a method and system for recommending commodities.

Background

In a traditional electronic commerce system, a navigation tag provided by a user electronic commerce platform performs multiple times of screening on commodities or inputs search conditions in a natural language mode to perform natural language search. However, under the condition of searching by using the navigation tag, the user is required to have clear commodity parameter screening conditions, which is difficult for the user insensitive to commodity parameters or with fuzzy demand, and meanwhile, in the natural language searching mode, the positioning accuracy of the natural language to the commodity tag is lower, so that the searching result is greatly different from the actual intention of the user.

With the development of artificial intelligence technology, the understanding accuracy of natural language input by a user is improved by introducing a large language model in the related technology, so that commodity recommendation according to fuzzy requirements of the user is realized.

Disclosure of Invention

Accordingly, the present disclosure provides at least a commodity recommendation method and system.

The technical scheme of the present disclosure is realized as follows:

in one aspect, the present disclosure provides a method of recommending goods, the method comprising:

generating a first commodity list according to a first navigation tag, wherein the first navigation tag represents a navigation tag determined from a plurality of navigation tags based on user operation;

and updating the first commodity list according to the first interaction information and the first navigation tag to obtain a second commodity list, wherein the first interaction information characterizes information input by a user through an interaction window provided by the intelligent agent.

In some embodiments, updating the first merchandise list based on the first interaction information and the first navigation tag includes:

determining at least one first commodity label from the plurality of commodity labels according to the semantic information of the first interaction information;

and updating the first commodity list according to at least one first commodity label and at least one second commodity label corresponding to the first navigation label to obtain a second commodity list.

In some implementations, the first interactive information includes at least one of natural language, speech, pictures;

determining at least one first commodity label from the plurality of commodity labels according to the semantic information of the first interaction information, wherein the at least one first commodity label comprises at least one of the following:

determining at least one first commodity label matched with the first semantic information from a plurality of commodity labels;

determining at least one first commodity label matched with the second semantic information from a plurality of commodity labels;

carrying out attribute identification on commodity objects in the pictures to obtain a plurality of commodity attributes; at least one first merchandise tag is determined from the plurality of merchandise tags that matches the plurality of merchandise attributes.

In some embodiments, updating the first merchandise list according to at least one first merchandise tag and at least one second merchandise tag corresponding to the first navigation tag to obtain a second merchandise list includes one of:

determining a second merchandise list based on an item in the first merchandise list that matches the at least one first merchandise tag in response to the at least one second merchandise tag including the at least one first merchandise tag;

Determining a second commodity list based on the commodities corresponding to the first commodity list and the third commodity label in response to the at least one first commodity label comprising at least one second commodity label and the third commodity label;

In response to the fourth one of the at least one second merchandise tag being opposite in semantic meaning to the at least one first merchandise tag, a second merchandise list is determined from the at least one first merchandise tag and the at least one second merchandise tag other than the fourth merchandise tag.

In some embodiments, before generating the recommended merchandise list according to the first navigation tag, the method further comprises:

Generating a third commodity list according to second interaction information, wherein the second interaction information characterizes information input by a user through an interaction window provided by the intelligent agent;

and generating a plurality of navigation tags according to the commodity tags corresponding to the plurality of commodities in the third commodity list.

In some embodiments, after updating the first merchandise list according to the first interaction information and the first navigation tag to obtain the second merchandise list, the method further comprises:

determining a third commodity corresponding to the user operation based on the user operation aiming at the second commodity list;

Determining commodity performance requirement information of the third commodity, and displaying and outputting commodity performance information of the third commodity based on a target tool corresponding to the commodity performance requirement information;

or generating commodity recommendation suggestions based on commodity labels corresponding to the third commodity and user portrait information of the user.

In some embodiments, after updating the first merchandise list according to the first interaction information and the first navigation tag to obtain the second merchandise list, the method further comprises:

the third information characterizes navigation tags determined by a user from a plurality of navigation tags or interactive information input by the user through an interactive window provided by an agent;

At least one fourth item to be deleted and/or added is determined from the second item list based on the third information.

In some embodiments, after updating the first merchandise list according to the first interaction information and the first navigation tag to obtain the second merchandise list, the method further comprises:

determining an output strategy for the second commodity list according to user portrait information of the user, wherein the output strategy is characterized in that each commodity displays the type of commodity parameters output;

and displaying and outputting the second commodity list according to the output strategy.

In some embodiments, the display screen of the electronic device includes a first area for displaying an interactive window provided by the agent and a second area for displaying a plurality of navigation tags and a second list of items;

The method further comprises the steps of:

and determining the area ratio of the first area and the second area on the display screen according to the user image information of the user.

In another aspect, the present disclosure further provides a commodity recommendation system, including:

The communication unit is used for acquiring user input information, wherein the user input information comprises a first navigation tag determined by a user from a plurality of navigation tags and first interaction information input by the user through an interaction window provided by an agent;

The commodity recommending unit is used for generating a first recommended commodity list according to the first navigation tag, and updating the first commodity list according to the first interaction information and the first navigation tag to obtain a second commodity list.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the aspects of the disclosure.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the disclosure and together with the description, serve to explain the technical aspects of the disclosure.

Fig. 1 is a schematic diagram of an implementation flow of a commodity recommendation method provided in the present disclosure;

FIG. 2 is a schematic diagram of a display interface in one embodiment provided by the present disclosure;

FIG. 3 is a schematic diagram of making merchandise recommendations in one embodiment provided by the present disclosure;

FIG. 4 is a timing diagram of making a recommendation for a good in one embodiment provided by the present disclosure;

FIG. 5 is a schematic diagram of a commodity recommendation system according to the present disclosure;

Fig. 6 is a schematic hardware structure of an electronic device provided in the present disclosure.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present disclosure more apparent, the technical solutions of the present disclosure are further elaborated below in conjunction with the drawings and the embodiments, and the described embodiments should not be construed as limiting the present disclosure, and all other embodiments obtained by those skilled in the art without making inventive efforts are within the scope of protection of the present disclosure.

In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with one another without conflict.

The term "first/second/third" is merely to distinguish similar objects and does not represent a particular ordering of objects, it being understood that the "first/second/third" may be interchanged with a particular order or precedence where allowed, to enable embodiments of the disclosure described herein to be implemented in other than those illustrated or described herein.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. The terminology used herein is for the purpose of describing the present disclosure only and is not intended to be limiting of the present disclosure.

The present disclosure provides a commodity recommendation method. The method may be performed by an electronic device, which may be a notebook computer, a tablet computer, a desktop computer, a set-top box, a mobile device (e.g., a mobile phone, a portable music player, a personal digital assistant, a dedicated messaging device, a portable game device), etc., or may be implemented as a server. The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, a content distribution network (Content Delivery Network, CDN), basic cloud computing services such as big data and an artificial intelligent platform.

In the following, the technical solutions in the embodiments of the present disclosure will be clearly and completely described with reference to the drawings in the embodiments of the present disclosure.

Fig. 1 is a schematic implementation flow chart of a commodity recommendation method provided in the present disclosure, as shown in fig. 1, the method includes steps S11 to S12 as follows:

And S11, generating a first commodity list according to a first navigation tag, wherein the first navigation tag represents a navigation tag determined from a plurality of navigation tags based on user operation.

Here, the Navigation Tab (Navigation Tab) refers to an important design element provided in an application interface of a shopping application or a web page of a shopping website for classifying and organizing commodity information, guiding a user to screen commodities. In some implementations, the navigation tag may refer to a hardware parameter tag generated based on hardware parameters of the article. For example, in a scenario where the commodity is a computer, the navigation tag may be tag information generated by classifying the computer according to the memory size of the computer, the type of a central processing unit (Central Processing Unit, CPU), the type of an image processor (Graphics Processing Unit, GPU), the type of a graphics card, and the like. In some implementations, the navigation tag can be a scenerized tag generated based on an application scene of the merchandise. For example, in the case where the commodity is a notebook computer, the navigation tag may be a tabby book, a master book, a designer book, a game book, or the like.

Shopping applications or shopping websites typically provide users with a plurality of navigation tags so that the user can screen items from multiple angles according to personal needs.

A first navigation tag characterizing a navigation tag determined from a plurality of navigation tags based on user operation. Here, the user operation may be any type of operation, for example, a click operation for the first navigation tag, a selection operation for the first navigation tag with a preset gesture, a voice command including a semantic meaning to select the first navigation tag, and the like.

Here, after the user determines the first navigation tag, the plurality of articles are screened according to the first navigation tag, thereby generating a first article list.

In some embodiments, in the case that the first navigation tag is a hardware parameter tag, the plurality of commodities may be filtered based on a hardware parameter corresponding to the hardware parameter tag, so as to generate the first commodity list.

In some embodiments, in the case that the first navigation tag is a scenerized tag, the at least one commodity tag corresponding to the first navigation tag may be determined according to a pre-established mapping relationship between the plurality of navigation tags and the plurality of commodity tags.

In some embodiments, a diversified merchandise tag may be generated for each merchandise, such that a first merchandise list corresponding to the first navigation tag is determined based on the diversified merchandise tag. Here, the diversified merchandise tags may be any type of tag, for example, parameter tags (e.g., "i7 processor", "32G memory", "RTX video card", etc.) describing hardware parameters of merchandise, and scenerized merchandise tags (e.g., "student book", "tabby book", "game book", "office book", "designer high-color-gamut screen", etc.) describing scenes of merchandise usage. In this way, by generating diverse merchandise tags from multiple perspectives, the merchandise tags may be enabled to express a broader semantic context associated with the merchandise.

Therefore, when the navigation tag is a scenerified navigation tag, the mapping relation between the navigation tag and the scenerified commodity tag can be established, and the mapping relation between the navigation tag and the hardware parameter tag can also be established. For example, when the navigation tag is "designer book", a mapping relationship between "designer book" and hardware parameter "color gamut not less than 100% Adobe RGB" may be established, and a mapping relationship between "designer book" and scene tag "office" and "light and thin" may be also established.

In some embodiments, in the case that the navigation tag is a hardware parameter tag, a mapping relationship between the navigation tag and the scenerified commodity tag may also be established, for example, a mapping relationship between the hardware parameter tag "RTX 4060" and the scenerified commodity tag "1080P fluency" is established.

In some embodiments, a commodity label library for storing diversified commodity labels may be searched based on the first navigation label, so as to determine a plurality of commodity labels matched with the first navigation label, and further generate a first commodity list based on commodities corresponding to the plurality of commodity labels.

And step S12, updating the first commodity list according to the first interaction information and the first navigation tag to obtain a second commodity list, wherein the first interaction information characterizes information input by a user through an interaction window provided by the agent.

Here, the agent refers to an intelligent entity for assisting in performing a commodity recommendation task. In some embodiments, the agent can perform semantic understanding on the interaction information input by the user by calling the model, so that accuracy of understanding the intention of the user is improved. In some embodiments, the agent may also implement operations on the application interface or web page by calling tools, e.g., the agent may call application programming interfaces (Application Programming Interface, APIs), function libraries, plug-ins, etc.

The interactive window provided by the agent is a window that can be used to receive user input information and display output feedback. In some embodiments, the interactive window provided by the agent may be displayed in parallel with the merchandise display window in the shopping application or shopping webpage, so that the user may browse the interactive information, merchandise screening result information, and a plurality of navigation tags with the agent at the same time. In some embodiments, the interactive window provided by the agent may be displayed in a floating window form on the upper layer of the merchandise display window, so that the user may freely move the interactive window according to the browsing requirement.

The first interactive information refers to interactive information input by a user through an interactive window. In some implementations, the first interaction information may be any type of information, such as text type, picture type, audio type, video type, or multimodal type of information. In this way, the user can input a natural language description of the commodity into a text input box in the interactive window through a text form, or input a voice description of the commodity through a voice input control provided in the interactive window, or upload a picture or video containing commodity information to the interactive window through a picture or video uploading control provided in the interactive window, and the like.

And after the first interaction information is determined, continuing to execute commodity recommendation according to the first interaction information and the first navigation tag, and updating the first commodity list according to the recommendation result to obtain a second commodity list.

In some implementations, the first list of items may be updated in any suitable manner based on the first interaction information and the first navigation tag.

In some embodiments, the plurality of commodities in the first commodity list may be screened according to the semantic information of the first interaction information, so as to obtain the second commodity list.

In some embodiments, the semantic information of the first interaction information and the semantic information of the first navigation tag may be fused, and commodity recommendation is performed according to the fused semantic information, so as to obtain the second commodity list.

In some embodiments, first, a plurality of commodity labels may be screened according to semantic information of first interaction information to obtain a plurality of commodity labels corresponding to the first interaction information, then, the plurality of commodity labels corresponding to the first interaction information and the plurality of commodity labels corresponding to the first navigation label are integrated to obtain integrated commodity labels, and finally, a second commodity list is generated according to the integrated commodity labels.

In some embodiments, a new scenerised merchandise tag may be generated for the merchandise using a tag generation model based on the interaction results of the navigation tag interaction (i.e., the user performs merchandise screening by manipulating the navigation tag) and the agent interaction (i.e., the user performs merchandise screening by entering interaction information in the interaction window). For example, in the commodity recommendation process, when a plurality of users are found to search for a notebook computer which meets the requirements of video clip, light weight, thin weight and portability at the same time, the user requirements can be understood and summarized by using a label generation model, so that a new scene commodity label of video clip portability is generated, and the label is associated with a plurality of commodities. Thus, based on the newly generated scenerised commodity label, the target commodity can be positioned more quickly and accurately.

In the embodiment provided by the disclosure, first, a first commodity list is generated according to a first navigation tag, wherein the first navigation tag represents a navigation tag determined from a plurality of navigation tags based on user operation, and then, the first commodity list is updated according to first interaction information and the first navigation tag to obtain a second commodity list, and the first interaction information represents multi-mode information input by a user through an interaction window provided by an intelligent agent. Therefore, in the scheme provided by the disclosure, when the user sends out the interaction information (namely, the commodity recommendation task) through the intelligent agent, the commodity recommendation can be performed by combining the previous navigation tag operation of the user, so that the previous navigation tag interaction information is integrated into the context information of the intelligent agent interaction, the dual-mode collaborative interaction of the navigation tag interaction and the intelligent agent interaction is realized, the high efficiency of the navigation tag on commodity screening is combined with the flexibility of the intelligent agent interaction and the advantage of supporting fuzzy interaction, the user can rapidly position the commodity meeting expectations by utilizing two interaction modes in a complex commodity screening scene, and the commodity recommendation efficiency and the user experience are improved.

In some embodiments, the updating the first merchandise list according to the first interaction information and the first navigation tag, that is, the step S12 may be implemented as the following steps S121 to S122:

step S121, determining at least one first merchandise tag from the plurality of merchandise tags according to the semantic information of the first interaction information.

The plurality of article tags refers to diversified tags generated for a plurality of articles. As described above, the plurality of merchandise tags may be any type of tag, for example, parameter tags generated based on merchandise hardware parameters, scenic merchandise tags generated based on the application scene of the merchandise, and the like.

The semantic information of the first interaction information may be feature vector information corresponding to the first interaction information. In some implementations, a feature vector of the first interaction information can be generated using the first model. The first Model may be any type of Model, such as a large Language Model (Large Language Model, LLM), a visual-Language Model (VLM), or a voice Model.

Thus, after determining the semantic information of the first interaction information, a plurality of first merchandise tags matching the semantic information may be determined from the plurality of merchandise tags. In some embodiments, the plurality of first merchandise tags that match the semantic information of the first interaction information may be determined by calculating semantic similarity of the semantic information of the first interaction information to the semantic information of the plurality of merchandise tags. For example, in the case that the first interaction information input by the user is "i want a notebook capable of cutting video", the semantic analysis of the information can be used to extract the user intention behind the first interaction information, for example, "cutting video" corresponds to the requirement of a high-performance CPU, a large memory, a professional display card, so that the first interaction information can be mapped to a commodity label "multi-core CPU", "large memory", "professional display card" and the like.

Step S122, updating the first commodity list according to the at least one first commodity label and the at least one second commodity label corresponding to the first navigation label, so as to obtain the second commodity list.

As described above, when the first navigation tag is a parameter tag generated based on a hardware parameter of an article, the corresponding at least one second article tag may be directly determined based on the hardware parameter corresponding to the first navigation tag, or the hardware parameter may be mapped to a scenerised article related thereto. For example, when the first navigation tag is the hardware parameter "RTX a5500", the product tag corresponding to the hardware parameter may be used as the second product tag. For another example, in the case where the hardware parameter is "RTX 4060", the hardware parameter may be mapped to the scenerised merchandise tag "1080P fluency". When the first navigation tag is a scenerised tag, the first navigation tag may be mapped to at least one hardware parameter type merchandise tag and/or at least one scenerised merchandise tag.

Therefore, the real-time bidirectional mapping of the agent interaction and the navigation tag interaction is realized by mapping the first interaction information and the first navigation tag of the user to the commodity tag.

In this way, under the condition that the bidirectional mapping can be realized, after at least one first commodity label corresponding to the first interaction information is determined, the at least one first commodity label and at least one second commodity label corresponding to the first navigation label can be fused to obtain a fused commodity label, so that the first commodity list is updated according to the fused commodity label to obtain a second commodity list.

In the embodiment provided by the application, on one hand, compared with the navigation tag, the first interactive information has richer information expression capability, and the diversified commodity tags not only have hardware parameter information and contain more scene description information, so that the plurality of first commodity tags determined based on the first interactive information can describe the real requirements of the user more personally and accurately, and on the other hand, based on the bidirectional mapping of navigation interaction and agent interaction, the information fusion of two interaction modes can be realized. Therefore, the accurate and efficient screening capability of the navigation tag on the commodity is combined with the diversified and personalized expression capability of the first interactive information, so that the screening efficiency of the commodity and the personalized demand on commodity recommendation can be considered, and the commodity recommendation efficiency and the user satisfaction are improved.

In some embodiments, the first interactive information comprises at least one of natural language, voice, and pictures;

in this way, the determining at least one first merchandise tag from the plurality of merchandise tags according to the semantic information of the first interaction information, that is, the above step S121, may be implemented as at least one of the following steps S1211 to S1213:

And step S1211, determining first semantic information corresponding to the natural language, and determining at least one first commodity label matched with the first semantic information from the plurality of commodity labels.

Here, in the case where the first interaction information is a natural language, the first semantic information corresponding to the natural language may be determined using a natural language processing method, for example, a large language model may be used to generate a feature vector corresponding to the natural language, and the feature vector may be used as the first semantic information.

After determining the first semantic information, at least one first merchandise tag having a similarity to the first semantic information greater than a specified threshold may be determined from the plurality of merchandise tags.

And S1212, determining second semantic information corresponding to the voice, and determining at least one first commodity label matched with the second semantic information from the plurality of commodity labels.

Here, in the case where the first interaction information is voice, the second semantic information corresponding to the first interaction information may be determined using a voice recognition technique. For example, the multi-modal model can be utilized to carry out semantic recognition on the voice to obtain second semantic information, the voice recognition model can be utilized to carry out voice recognition on the voice to obtain corresponding text information, and then semantic analysis is carried out on the text information by the semantic analysis model to obtain second semantic information.

After determining the second semantic information corresponding to the voice, at least one first merchandise tag having a similarity with the second semantic information greater than a specified threshold may be determined from the plurality of merchandise tags.

And S1213, carrying out attribute identification on the commodity object in the picture to obtain a plurality of commodity attributes, and determining at least one first commodity label matched with the commodity attributes from the commodity labels.

Here, in the case where the first interaction information is a picture, attribute recognition may be performed on the picture using an image processing technique to obtain a plurality of commodity attribute information. The commodity attribute information may include commodity type, commodity name, brand, model, size, color, material, etc. For example, in the case where a user uploads a photograph of a notebook computer, commodity attributes of the notebook computer including "black", "associated brand", "associated small new series", "14 inches" and the like can be obtained through attribute recognition.

In this way, after determining the plurality of commodity attributes included in the picture, at least one first commodity label matching the plurality of commodity attributes, for example, a commodity type label corresponding to a commodity type, a commodity color label corresponding to a commodity color, or the like, may be determined from among the plurality of commodity labels.

In the embodiment provided by the application, through supporting the user to input the first interactive information of multiple types, the cross-mode semantic understanding can be realized, and the flexibility and the user experience of the commodity recommendation method are further improved.

In some embodiments, the updating the first merchandise list according to the at least one first merchandise tag and the at least one second merchandise tag corresponding to the first navigation tag to obtain the second merchandise list, that is, the step S122 may be implemented as at least one of the following steps S1221 to S1223:

Step S1221, in response to the at least one second merchandise tag including the at least one first merchandise tag, determining the second merchandise list based on an merchandise in the first merchandise list that matches the at least one first merchandise tag.

Here, the at least one second article tag includes at least one first article tag, which may mean that the semantic information of the at least one second article tag may cover the semantic information of the at least one first article tag, or that the semantic information of the at least one second article tag is upper information of the semantic information of the at least one first article tag. In this case, the range of the commodity corresponding to the at least one second commodity label is larger than the range of the commodity corresponding to the at least one first commodity label.

For example, when the second commodity label corresponding to the first navigation label is "weight is less than or equal to 1.3kg", the semantic information of the second commodity label may cover the semantic information of the plurality of first commodity labels under the conditions that the plurality of first commodity labels corresponding to the first interaction information are "multi-core CPU", "large memory", "weight is less than or equal to 1.3 kg". For example, when the second commodity label corresponding to the first navigation label is a "notebook computer", and the plurality of first commodity labels corresponding to the first interaction information are "large memory" and "independent graphics card", the semantic information of the second commodity label is the upper information of the semantic information of the plurality of first commodity labels.

Thus, in the case where the at least one second merchandise tag includes at least one first merchandise tag, the first interaction information characterizing the user input is a further screening of the screening results of the first navigation tag. Therefore, based on the at least one first commodity label, further screening the commodities in the first commodity list to obtain commodities matched with the first commodity label, and accordingly determining a second commodity list. For example, in the above example, based on the plurality of first commodity labels "multi-core CPU", "large memory", "weight is 1.3kg or less", commodities including only the second commodity label "weight is 1.3kg or less" are further screened to obtain the second commodity list.

Step S1222, in response to the at least one first article tag including the at least one second article tag and the third article tag, determining the second article list based on articles corresponding to the first article list and the third article tag.

Here, the at least one first article tag includes at least one second article tag and a third article tag, which may mean that the semantic information of the at least one first article tag may cover the semantic information of the at least one second article tag and the semantic information of the third article tag, where the third article tag is a different article tag from the second article tag. In this case, the range of the article corresponding to the at least one first article tag is larger than the range of the article corresponding to the at least one second article tag.

For example, when the plurality of second commodity labels are "weight 1.3kg or less", "RTX graphic card", and the plurality of first commodity labels are "weight 1.3kg or less", "RTX graphic card", "AMD graphic card", the semantics of the plurality of first commodity labels include the semantics information of the plurality of second commodity labels and the semantics information of the third commodity label.

Because the first interaction information corresponding to at least one first commodity label is information input after the first navigation label, the latest commodity purchasing requirement of a user can be expressed. Therefore, in the case where the at least one first article tag includes at least one second article tag and a third article tag, both the article corresponding to the at least one second article tag (i.e., the article in the first article list) and the article corresponding to the third article tag are recommended articles, i.e., the article corresponding to the third article tag is updated into the first article list, thereby obtaining the second article list.

Step S1223, in response to the fourth merchandise tag of the at least one second merchandise tag being opposite in semantic meaning to the at least one first merchandise tag, determining the second merchandise list based on the at least one first merchandise tag and at least one second merchandise tag other than the fourth merchandise tag.

Here, the fourth merchandise tag is opposite to the at least one first merchandise tag in terms of its semantic meaning, and the user is represented by the first interaction information by negating a certain merchandise attribute or parameter, resulting in the fourth merchandise tag being determined based on the first navigation tag being opposite to the one first merchandise tag being determined based on the first interaction information.

For example, when the plurality of second merchandise tags are "weight less than or equal to 1.3kg", "RTX graphic card", "AMD graphic card", the first interactive information includes the first merchandise tag "non-AMD graphic card", the "AMD graphic card" in the second merchandise tag is determined as the fourth merchandise tag.

Similarly, since the at least one first merchandise tag may express the latest merchandise purchase demand of the user, in the case that the fourth merchandise tag is included in the at least one second merchandise tag, the second merchandise list is generated according to the at least one first merchandise tag and the at least one second merchandise tag other than the fourth merchandise tag.

In the embodiment provided by the disclosure, the first interaction information input by the user and the first navigation tag selected by the user are mapped to the commodity tag, so that the bidirectional mapping of the intelligent agent interaction mode and the navigation tag interaction mode in the multi-round commodity recommendation process is realized, the commodity recommendation list is dynamically adjusted according to the continuous modification or adjustment of the shopping demand of the user and the combination of the two interaction modes, and the accuracy of commodity recommendation is improved.

In some embodiments, before the generating the recommended merchandise list according to the first navigation tag, that is, before the step S11, the method further includes the following steps S13 to S14:

And step S13, generating a third commodity list according to second interaction information, wherein the second interaction information characterizes information input by a user through an interaction window provided by the intelligent agent.

Here, the second interactive information refers to text, voice, picture, video, and/or multimodal information, etc., which are input by the user through the interactive window provided by the agent.

In this way, after the second interaction information is obtained, semantic analysis is performed on the second interaction information so as to map the second semantic information to a plurality of commodity labels, and then a third commodity list is generated based on the plurality of commodity labels corresponding to the second interaction information.

For example, in the case where the plurality of commodity labels corresponding to the second interaction information are the "multi-core CPU" and the "RTX graphics card", each commodity in the third commodity list has the two commodity labels at the same time.

Step S14, generating a plurality of navigation tags according to the product tags corresponding to the plurality of products in the third product list.

Here, the article tags corresponding to the plurality of articles in the third article list may include all article tags corresponding to each of the plurality of articles.

For example, in the case where the third commodity list is a plurality of commodities having both the "multicore CPU" and the "RTX graphic card" label, the commodity labels corresponding to the plurality of commodities include the "multicore CPU" and the "RTX graphic card", and other all commodity labels, for example, "weight is less than or equal to 1.3kg", "high color gamut", and the like.

In this way, the semantic analysis can be performed on the commodity labels corresponding to the commodities, and the commodities are divided into a plurality of categories according to the semantic analysis result, so that the navigation labels corresponding to each category are determined.

In some embodiments, the plurality of items may be categorized according to any suitable criteria and corresponding navigation tags generated.

In some embodiments, the plurality of merchandise is classified according to attributes, uses, functions, or the like of interest of the user according to merchandise tags and user portraits corresponding to the plurality of merchandise, and corresponding navigation tags are dynamically generated for each class. For example, in a case where the user focuses more on the commodity color, the plurality of commodities may be divided into a plurality of categories according to the commodity color, and corresponding navigation tags may be generated. For another example, when the user is more focused on a customer group of the commodity, the plurality of commodities may be classified into a plurality of categories according to commodity labels related to the commodity user, and corresponding navigation labels may be generated, for example, "student book", "work book", "game book", and the like.

In some embodiments, the plurality of commodities are classified according to the commodity labels corresponding to the plurality of commodities and according to the scores of the plurality of users on a certain attribute of the commodities, and corresponding navigation labels are dynamically generated for each class. For example, the navigation tags "low cost performance", "common cost performance", and "high cost performance" are generated, and the like, by classifying according to the commodity cost performance score.

In the above embodiment provided by the present disclosure, after a round of recommendation is performed on a commodity based on the second interaction information of the user and the agent, the recommended commodity is dynamically classified and navigation tags of each classification are dynamically generated, so that flexibility and practicability of the navigation tags can be improved, and further screening efficiency and shopping experience of the user are improved.

In the above embodiment provided by the present disclosure, after a round of recommendation is performed on the commodity based on the second interaction information of the user and the agent, the recommended commodity is dynamically classified and navigation tags of the respective classifications are dynamically generated. And updating the first commodity list according to the first interaction information of the user and the intelligent agent and the first navigation tag to obtain a second commodity list. Therefore, the user can more accurately find the commodity needed by the user through interaction with the intelligent agent and the navigation tag selection and the mode of repeated cyclic execution, and the user experience is improved.

In some embodiments, after updating the first merchandise list according to the first interaction information and the first navigation tag to obtain the second merchandise list, that is, after the step S12, the method further includes the following steps S15 and S16, or the steps S15 and S17:

Step S15, based on the user operation for the second commodity list, determining a third commodity corresponding to the user operation.

In some embodiments, the user operation for the second product list may refer to a user selection operation of at least one product in the second product list. For example, the user selects at least one commodity in the second commodity list by clicking, sliding, gesture, voice, or the like. In this way, the product selected by the user is taken as the third product.

In some embodiments, the user operation for the second product list may refer to a selection operation of the navigation tag corresponding to the second product list by the user. For example, the user selects the navigation tab "RTX a5500" or the like by clicking, sliding, gesture, voice, or the like. In this way, the product corresponding to the navigation tag selected by the user is used as the third product.

And S16, determining commodity performance requirement information of the third commodity, and displaying and outputting commodity performance information of the third commodity based on a target tool corresponding to the commodity performance requirement information.

Here, after determining the corresponding third commodity by the user operation, commodity performance information of the third commodity is displayed and outputted by the target tool.

And the commodity performance requirement information of the third commodity characterizes the requirement information of the user on the commodity performance such as functions, parameters, usage scenes and the like of the third commodity. In some embodiments, the commodity performance requirement information may be requirement information for any performance, for example, requirement information for hardware parameters, requirement information for rendering effects, requirement information for game frame rates, and the like. In some embodiments, the commodity performance requirement information may be expert evaluation information for commodity performance.

In this way, from the commodity performance requirement information, a target tool for exhibiting the performance may be determined. For example, in the case where commodity performance requirement information is requirement information for hardware parameters, a parameter contrast card may be used as a target tool. For another example, in the case where the commodity performance requirement information is requirement information for a rendering effect or a game frame rate, the scene simulator may be set as a target tool. For another example, in the case where the commodity performance requirement information is expert evaluation information for commodity performance, the expert evaluation system may be regarded as a target tool.

In the embodiment provided by the disclosure, the commodity performance information of the third commodity is displayed and output by using the target tool, so that the user can intuitively check the interested commodity performance, more decision references are provided for the user, and the shopping experience of the user is improved.

And S17, generating commodity recommendation suggestions based on the commodity labels corresponding to the third commodities and the user portrait information of the user.

Here, after determining the third commodity to which the user operates, commodity recommendation suggestions relating to the third commodity are further generated.

Each commodity has diversified commodity labels, and the commodity labels have rich semantic information, and can describe parameter information, attribute information, application scene information and the like of the commodity from multiple angles. Therefore, after the second interaction information input by the user is mapped to the plurality of commodity tags and the third commodity is determined based on the plurality of commodity tags, the commodity recommendation suggestion can be continuously generated based on all commodity tags of the third commodity.

User portrayal information refers to a virtual user model constructed by collecting, integrating and analyzing multidimensional data of a user, such as a personalized description model constructed based on the data of historical behavior, preference characteristics, equipment usage habits, purchase records and the like of the user. In a shopping scene, the user portrait information can be used for describing points of interest, demand trends and interaction modes of the user, and is a basis for providing personalized recommendation for the user. In some embodiments, the user profile information may include a variety of label combinations of price sensitive, performance priority, brand loyalty, student groups, office usage, gaming usage, and the like.

And when the commodity recommendation suggestion is generated based on the commodity label corresponding to the third commodity, referring to the user portrait information of the user at the same time. For example, in the case where the user portrayal information characterizes that the user is a price-sensitive user, a commodity recommendation suggestion may be generated based on the price-related commodity label of the third commodity, for example, a suggestion "if a product is purchased, the cost performance is higher" may be generated, a price graph of a plurality of third commodities may also be generated, or the like.

In one embodiment, where the user operation for the second merchandise list is a navigation tab operation, and the navigation tab operation characterizes the user dragging the price filter bar from 8000 to 6000, first, the merchandise budget is determined to drop 25%, then the hardware performance profile associated to the current price range (i.e., 6,000-8,000) and the user portrayal information of the user (e.g., once focused on RTX 40 series graphics cards), then a hierarchical early warning prompt is automatically generated for the user, such as:

"Red alert zone ('6,000-6,500') with 40% reduced performance;

yellow advice area (, 6,500-7,000) recommends sacrificing storage capacity retention core performance;

Green security zone (> 7,000), maintains full game performance. ",

Meanwhile, in the case of keeping the original commodity recommendation list, a performance loss prediction map is displayed and output alternatives are displayed for the user, for example, an "installment" and "old swap" entry are popped up.

Also for example, in the above example, if the user increases the budget, e.g., drags the price filter bar from 8000 to 10000, a merchandise recommendation suggestion may be generated suggesting which merchandise parameters the user references with emphasis.

In some embodiments, user portrait information may be updated based on various types of user behavior collected during interactions.

In some embodiments, user portrayal information may be modified by collecting explicit feedback behavior of the user. Here, explicit feedback behavior may include user scoring, commentary, modification of screening conditions, etc. of the merchandise. For example, the user adjusts the filtering condition about the storage capacity of the hard disk through natural language or navigation tags, and the user portrait information related to "storage capacity" can be modified accordingly, for example, the user portrait is updated to a "storage sensitive" user.

In some embodiments, implicit feedback behavior of the user may be collected to modify the user profile information. Here, implicit feedback actions may include user interaction bias (e.g., more toward navigation tab interactions), dwell time on different hardware parameter pages (e.g., dwell time on "heat sink" pages), and so forth.

In the above embodiment provided by the present disclosure, since the solution provided by the present disclosure establishes diversified merchandise tags for each merchandise in advance, merchandise recommendation suggestions can be generated according to the screening conditions of the user and by combining a plurality of merchandise tags of the third merchandise with the portrait information of the user, thereby implementing deep analysis and personalized recommendation of the third merchandise, providing richer and personalized decision reference information for the user, and improving shopping experience of the user.

In some embodiments, after updating the first merchandise list according to the first interaction information and the first navigation tag to obtain the second merchandise list, i.e. after the step S12, the method further includes the following steps S18 to S19:

And S18, acquiring third information, wherein the third information represents navigation tags determined by a user from a plurality of navigation tags or interaction information input by the user through an interaction window provided by the agent.

After generating the second list of items, the user may continue to refine or modify the screening criteria for the items through navigation tag interactions or agent interactions.

Here, in response to an operation of the plurality of navigation tags by the user, the navigation tag corresponding to the operation is taken as third information, or in response to an input action for use in an interactive window provided by the agent, information received through the interactive window is taken as third information.

Step S19, determining at least one fourth commodity to be deleted and/or added from the second commodity list based on the third information.

And under the condition that the third information is obtained, recalculating the matching degree between the plurality of commodities and the third information, and deleting and/or adding partial commodities in the second commodity list.

The method comprises the steps of firstly, carrying out semantic analysis on third information to determine semantic information of the third information, then determining a plurality of commodity labels corresponding to the third information according to the semantic information of the third information, and finally, determining at least one fourth commodity to be deleted and/or added from a second commodity list according to the commodity labels corresponding to the third information.

In the above embodiment provided by the present disclosure, by continuously acquiring the third information, the user may continuously perform the merchandise screening in a manner of navigation tag interaction or agent interaction, thereby implementing a continuous interaction process of dual-mode collaboration, and referring to the previous interaction information in each interaction mode, so as to improve the merchandise recommendation efficiency, accuracy and continuity of multiple interactions.

In some embodiments, after updating the first product list according to the first interaction information and the first navigation tag to obtain the second product list, that is, after the step S12, the method further includes the following steps S110 to S111:

and step S110, determining an output strategy for the second commodity list according to the user portrait information of the user, wherein the output strategy is characterized in that each commodity displays the commodity parameter type output, and different user portrait information corresponds to different output strategies.

After the second commodity list is generated, an output strategy for displaying and outputting the second commodity list is further determined according to the user portrait so that the display and output mode of the second commodity list is matched with the user portrait information.

The output strategy refers to the commodity parameter type displayed and output for each commodity. Because different users have different personalized requirements, different output strategies are determined for users of different user portrait information. For example, for price sensitive users, the output policy may emphasize price related parameters such as price, promotional activity, and historical price trend, while for parameter sensitive users, the output policy may emphasize hardware parameters such as CPU model, graphics card specification, memory size, etc. For example, for novice users or small white users, scene case videos can be prominently displayed and more specialized technical parameters can be folded, and for specialized users, detailed parameter tables can be unfolded to provide custom screening.

Meanwhile, each commodity parameter type may be used as part of an output policy that determines whether the commodity parameter type is displayed on the current commodity card. For example, in the student book classification, the output policy may be set to display commodity parameters of types such as a screen color gamut, weight, duration, etc., while hiding commodity parameters of types such as the number of interfaces or a heat dissipation scheme, etc., where the user's attention is low.

And step S111, displaying and outputting the second commodity list according to the output strategy.

Here, displaying the output refers to a process of rendering and presenting the second item list on the user interface according to the determined output policy.

After the output strategy is determined, corresponding commodity parameters are obtained from a back-end database according to the commodity parameter types in the output strategy, and rendering is carried out according to the field sequence and format configured by the strategy. For example, for a game book preference user, the system may preferentially display the display card model, frame rate prediction, heat dissipation design, etc. commodity parameters on the interface, and highlight these parameters with a striking color so that the user can quickly identify the core selling point of the commodity.

In the embodiment provided by the disclosure, the output strategy is dynamically generated based on the user portrait information of the user, so that the user can intuitively and conveniently acquire the interested commodity information, thereby realizing personalized commodity recommendation effect and improving shopping experience of the user.

In some embodiments, the display screen of the electronic device includes a first area for displaying an interactive window provided by the agent and a second area for displaying the plurality of navigation tags and the second list of items.

Here, the display screen of the electronic device is divided into two areas to respectively display the interactive window, the navigation tag and the commodity list provided by the intelligent agent, so that the user can perform commodity screening through two interactive modes.

The first area is an area for displaying an agent interaction window, presents an agent-guided dialog interface, for example, presents contents such as a recommendation statement, a user demand analysis result, a product recommendation suggestion, and the like, and a user information input window. In some embodiments, the first region may be located at any position of the display screen, for example, a right side position, a left side position, an upper side position, a lower side position, or a center position of the screen, etc. In some embodiments, the presentation position of the first region on the screen may be autonomously determined by the user according to the interaction habit.

And the second area is used for displaying the traditional navigation tag and the commodity list. In some embodiments, the second region may be located at any position of the display screen, for example, a right side position, a left side position, an upper side position, a lower side position, or a center position of the screen, etc. In some embodiments, the position of the second region in the display screen may be dynamically adjusted according to the position of the first region selected by the user, or the presentation position of the second region on the screen may be autonomously determined by the user according to the interaction habit.

By performing agent interactions (e.g., natural language interactions) in the first region, the user's personalized shopping experience may be enhanced, meeting the needs of fuzzy demand type users. Through executing the navigation tag interaction in the second area, the operation habit of the user with clear requirements can be met, the user can conveniently and quickly locate the target commodity, and the browsing efficiency is improved. The advantages of the two interaction modes, namely the high efficiency of the interaction of the navigation tag and the individuation of the interaction of the natural language, can be considered through simultaneously displaying and outputting the agent interaction window and the navigation tag interaction window, so that the effects of the interaction synergy and enhancement of the two modes are realized.

As shown in fig. 2, in one embodiment of the present disclosure, a first region 210 and a second region 220 are included in a display screen 200 of an electronic device.

The first area 210 provides an interactive window for the agent. In the first area 210, a process of interaction between the user and the agent using natural language is displayed, wherein the information 211 is natural language interaction information input by the user, and the information 212 is recommendation suggestion generated by the agent for the user according to the information 211. Meanwhile, the first area 210 further includes an information input box 213 so that a user can input interactive information through the information input box 213.

The second area 220 is a navigation tab interaction window. The second area 220 includes a navigation tag area 221 and a commodity list area 222, a plurality of navigation tags, that is, "brand", "model", "weight", "graphic card", "CPU", are displayed in the navigation tag area 221, and commodities 2221 and 2222 currently recommended for the user are displayed in the commodity list area 222.

In actual use, the information in the first area 210 and the second area 220 are in a linkage state, that is, the interactive information input by the user in the first area 210 can update the commodity list displayed in the second area 220, and meanwhile, the navigation tag operation and commodity selection of the user in the second area 220 can also be used as context information of the interaction of the agent, for example, as a basis for generating commodity recommendation suggestions by the agent, as a basis for executing commodity recommendation by the agent according to the interactive information input by the user, and the like.

In fig. 2, the relative positions of the first region 210 and the second region 220 are arranged side by side, but in practical applications, the relative positions of the first region 210 and the second region 220 may be any other type, for example, the second region 220 is displayed above or below the first region 210.

In some embodiments, the method further comprises the step of S112 of:

And step S112, determining the area occupation ratio of the first area and the second area on the display screen according to the user portrait information of the user.

The area ratio refers to the area ratio of the first area and the second area in the display screen, for example, 1:1, 2:1, 1:2, and the like.

Here, the area ratio of the first area and the second area on the display screen is determined according to the user image information of the user. For example, for users who prefer agent interactions (e.g., natural language interactions), the area of the first region may be enlarged appropriately to enhance the visibility of the AI-directed, while for users who prefer traditional navigation tag interactions, the area of the first region may be reduced, and the area of the second region increased, to facilitate the user's quick positioning of the desired merchandise using the navigation tag.

In some embodiments, the area occupation of the first area and the second area in the display screen can be dynamically adjusted according to the change of the user portrait information. For example, as a small white user, a larger first area and a smaller second area can be set, and as the user related knowledge increases, the area of the first area can be reduced and the area of the second area can be enlarged as the small white user is changed into a professional user.

In the above embodiment provided by the present disclosure, the area occupation ratio of the first area and the second area is dynamically adjusted according to the user image information, so that personalized interface adaptation can be implemented for the user, and the interaction process is more fit with the user characteristics, thereby improving the shopping experience of the user.

Next, a flow of commodity recommendation by using the commodity recommendation system in one embodiment provided by the present disclosure will be described with reference to fig. 3. The commodity recommendation system comprises a communication unit 310, a routing engine 320, a first parsing unit 330, a second parsing unit 340, a semantic mapping engine 350, a knowledge-graph database 370, a decision unit 360 and a user center 380. Next, this embodiment will be described in conjunction with the following steps S301 to S307:

Step S301, the communication unit 310 receives the first user information sent by the terminal 390 and sends the first user information to the routing engine 320, and then step S302 follows;

step S302, the routing engine 320 sends the first user information to the first parsing unit 330 or the second parsing unit 340 according to the type of the first user operation information, and then, step S303 is executed;

Here, the first parsing unit 330 is used for parsing interactive information input by a user through an interactive window provided by the agent, and the second parsing unit 340 is used for parsing operation information of the user on a plurality of navigation tags.

In some implementations, the first parsing unit 330 may be implemented as a hybrid model of a bi-directional transformer coded representation model (Bidirectional Encoder Representations from Transformers, BERT) and a graph-sampling aggregation model (GRAPH SAMPLE AND AGGREGATE).

Step S303, when the first user information is the third interaction information input by the user, the first parsing unit 330 generates the third semantic information corresponding to the third interaction information and sends the third semantic information to the semantic mapping engine 350, or when the first user information is the operation information of the user on a plurality of navigation tags, the second parsing unit 340 determines the second navigation tag corresponding to the operation information and sends the second navigation tag information to the semantic mapping engine 350;

Step S304, the semantic mapping engine 350 maps the received third semantic information or the second navigation tag to a plurality of commodity tags and sends the commodity tags to the decision unit 360, and then, the step S305 is executed;

Step S305, the decision unit 360 queries the knowledge graph database 370 to determine a fourth commodity list, and at the same time, the decision unit 360 obtains the user portraits information of the user from the user center 380 to determine a commodity output strategy according to the user portraits information;

in some implementations, user center 380 may be implemented as a hybrid database scheme of a key value store database (Redis) and a Neo4j image database.

In some implementations, the decision unit 360 employs a furin (APACHE FLINK) stream processing model as the real-time decision engine.

Step S306, the decision unit 360 generates a commodity output list based on the commodity output strategy and the fourth commodity list, and sends the commodity output list to the user terminal 390, and then, step S307 is executed;

in some implementations, visual presentation of 3D graphics and data is implemented in a web browser using a simplified API of three.js and hardware accelerated rendering capabilities of a web graphic library (WebGL (Web Graphics Library)).

Step S307, in the case that the user transmits new user information through the terminal 390, the above steps S301 to S307 are repeatedly performed, and the user portrait information in the user center 380 is updated according to the new operation information.

Next, a timing chart corresponding to the above embodiment will be described with reference to fig. 4. As shown in fig. 4, this embodiment includes the following steps S401 to S421:

step S401, the terminal 390 sends the first user information to the routing engine 320 through the communication unit 310, after which step S402 is performed;

Here, the first text information input by the user will be described by taking the first user information as an example, for example, "help me find a notebook computer that can clip a 4K video".

Note that, the communication process between the terminal 390 and each unit or engine in the commodity recommendation system is implemented by the communication unit 310, and for brevity, the description about the communication unit 310 is omitted here.

Step S402, the routing engine 320 sends the first text information input by the user to the first parsing unit 330, and then, step S403 is executed;

Step S403, the first parsing unit 330 sends the semantic information corresponding to the first text information to the semantic mapping engine 350, and then, step S404 is executed;

the first parsing unit 330 parses the first text information of the user to obtain corresponding semantic information.

Step S404, the semantic mapping engine 350 sends a plurality of commodity labels to the decision unit 360, and then, step S405 and step S407 are executed;

here, after the semantic mapping engine 350 maps the received semantic information to a plurality of commodity tags, the plurality of commodity tags are transmitted to the decision unit 360.

Step S405, the decision unit 360 sends a plurality of commodity labels to the knowledge-graph database 370, and then step S406 is executed;

step S406, the knowledge-graph database 370 sends a list of recommended products to be outputted to the decision unit 360, and then step S409 is executed;

Here, the knowledge-graph database 370 determines a fourth item list to be output according to the plurality of item tags, and transmits the fourth item list to the decision unit 360.

Step S407, the decision unit 360 sends the user portrait inquiry information to the user center 380, and then, step S408 is executed;

Step S408, the user center 380 sends user portrait information to the decision unit 360, and then step S409 is executed;

Here, the user center 380 inquires user portrait information of the user, and transmits the user portrait information to the decision unit 360.

Step S409, the decision unit 360 determines a commodity output strategy according to the user portrait information, generates a commodity output list according to the commodity output strategy, and sends the commodity output list to the terminal 390;

Step S410, in response to the user' S operation on the navigation tag, the terminal 390 sends second user information to the routing engine 320, after which step S411 is performed;

here, the second user letter characterizes the operation information of the user clicking on the navigation tab "RTX a 5500".

Step S411, the routing engine 320 sends the second user information to the second parsing unit 340, and then, step S412 is performed;

Step S412, the second parsing unit 340 sends the target navigation tag of the user operation, i.e., "RTX A5500", to the semantic mapping engine 350, after which step S413 is performed;

Step S413, the semantic mapping engine 350 sends the plurality of commodity labels to the decision unit 360, and then step S414 is performed;

here, after the semantic mapping engine 350 maps the received target navigation tag to a plurality of commodity tags, the plurality of commodity tags are transmitted to the decision unit 360.

Step S414, the decision unit 360 updates the context information of the user interaction according to the received plurality of commodity labels and generates a comparison report;

here, based on the "graphic card RTX a5500" clicked by the user, the decision unit 360 generates a performance comparison report of a plurality of commodities including the graphic card RTX a 5500.

Step S415, decision unit 360 sends a comparison report to terminal 390, after which step S416 is performed;

Step S416, in response to the user entering third user information via the interactive window provided by the agent, terminal 390 sends the third user information to routing engine 320;

step S417, the routing engine 320 sends the third user information to the first parsing unit 330, and then step S418 is performed;

step S418, the first parsing unit 330 sends semantic information corresponding to the user' S natural language information to the decision unit 360, and then, step S419 is performed;

the first parsing unit 330 obtains corresponding fourth semantic information, for example, "greater storage is required" for the third user information of the user.

Step S419, the commodity recommendation system sends fourth semantic information to the user center 380 and updates the commodity list to be recommended, and then step S420 and step S421 are executed;

step S420, the user center 380 updates the user portrait information according to the fourth semantic information;

here, the user center 380 increases the weight of the "storage sensitivity" feature in the user portrait information by, for example, 20%.

In step S421, the commodity recommendation center generates a commodity output list according to the output policy corresponding to the user portrait information from the updated commodity list to be recommended, and sends the commodity output list to the terminal 390.

Here, according to the latest input information of the user, the 2TB SSD configuration model having a larger storage capacity may be displayed in a limited manner in the commodity output list.

Based on the foregoing embodiments, the present disclosure provides a commodity recommendation system. As shown in fig. 5, the commodity recommendation system 500 includes:

The communication unit 510 is configured to obtain user input information, where the user input information includes a first navigation tag determined by a user from a plurality of navigation tags and first interaction information input by the user through an interaction window provided by the agent;

The commodity recommending unit 520 is configured to generate a first recommended commodity list according to the first navigation tag, and update the first commodity list according to the first interaction information and the first navigation tag to obtain a second commodity list.

In some embodiments, the commodity recommendation unit 520 is configured to:

determining at least one first commodity label from a plurality of commodity labels according to the semantic information of the first interaction information;

And updating the first commodity list according to the at least one first commodity label and at least one second commodity label corresponding to the first navigation label to obtain the second commodity list.

In some embodiments, the first interactive information comprises at least one of natural language, voice, and pictures;

the commodity recommendation unit 520 is configured to perform at least one of:

Determining at least one first commodity label matched with the first semantic information from the plurality of commodity labels;

determining at least one first commodity label matched with the second semantic information from the plurality of commodity labels;

carrying out attribute identification on commodity objects in the pictures to obtain a plurality of commodity attributes; at least one first merchandise tag that matches the plurality of merchandise attributes is determined from the plurality of merchandise tags.

In some embodiments, the commodity recommendation unit 520 is configured to perform one of the following:

Determining the second commodity list based on the commodities in the first commodity list matched with the at least one first commodity label in response to the at least one second commodity label containing the at least one first commodity label;

Determining, in response to the at least one first merchandise tag comprising the at least one second merchandise tag and a third merchandise tag, the second merchandise list based on merchandise corresponding to the first merchandise list and the third merchandise tag;

In response to a fourth merchandise tag of the at least one second merchandise tag being opposite in semantic meaning to the at least one first merchandise tag, determining the second merchandise list from the at least one first merchandise tag and at least one second merchandise tag other than the fourth merchandise tag.

In some embodiments, the commodity recommendation unit 520 is further configured to:

Generating a third commodity list according to second interaction information, wherein the second interaction information characterizes information input by a user through an interaction window provided by the intelligent body;

And generating a plurality of navigation tags according to the commodity tags corresponding to the commodities in the third commodity list.

In some embodiments, the commodity recommendation unit 520 is further configured to:

determining a third commodity corresponding to the user operation based on the user operation for the second commodity list;

Determining commodity performance requirement information of the third commodity, and displaying and outputting commodity performance information of the third commodity based on a target tool corresponding to the commodity performance requirement information;

Or alternatively

And generating commodity recommendation suggestions based on the commodity labels corresponding to the third commodities and the user portrait information of the user.

In some embodiments, the commodity recommendation unit 520 is further configured to:

The method comprises the steps of obtaining third information, wherein the third information represents navigation tags determined by a user from a plurality of navigation tags or interactive information input by the user through an interactive window provided by an agent;

And determining at least one fourth commodity to be deleted and/or added from the second commodity list based on the third information.

In some embodiments, the commodity recommendation unit 520 is further configured to:

Determining an output strategy for the second commodity list according to the user portrayal information of the user, wherein the output strategy is characterized in that each commodity displays the type of the commodity parameter output;

And displaying and outputting the second commodity list according to the output strategy.

In some embodiments, the display screen of the electronic device comprises a first area and a second area, wherein the first area is used for displaying an interactive window provided by the intelligent agent;

The commodity recommending unit 520 is further configured to determine an area ratio of the first area and the second area on the display screen according to the user portrait information of the user.

The description of the system embodiments above is similar to that of the method embodiments above, with similar benefits as the method embodiments. In some embodiments, the functions or units included in the system provided by the embodiments of the present disclosure may be used to perform the methods described in the method embodiments, and for technical details not disclosed in the embodiments of the system of the present disclosure, please understand with reference to the description of the embodiments of the method of the present disclosure.

If the technical scheme of the disclosure relates to personal information, the product applying the technical scheme of the disclosure clearly informs the personal information processing rule before processing the personal information, and obtains personal autonomous consent. If the technical scheme of the disclosure relates to sensitive personal information, the product applying the technical scheme of the disclosure obtains individual consent before processing the sensitive personal information, and simultaneously meets the requirement of 'explicit consent'. For example, a clear and obvious mark is set at a personal information acquisition device such as a camera to inform that the personal information acquisition range is entered, personal information is acquired, if the personal voluntarily enters the acquisition range, the personal information is considered as consent to acquire the personal information, or if a clear mark/information is used on a personal information processing device to inform that the personal information processing rule is used, personal authorization is obtained through popup information or a mode of requesting the personal information to upload the personal information by the personal, wherein the personal information processing rule can comprise information such as a personal information processor, a personal information processing purpose, a processing mode, a processed personal information type and the like.

It should be noted that, in the embodiment of the present disclosure, if the above-mentioned commodity recommendation method is implemented in the form of a software function module, and is sold or used as a separate product, the commodity recommendation method may also be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the embodiments of the present disclosure may be essentially or portions contributing to the related art, and the software product may be stored in a storage medium, including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the methods described in the embodiments of the present disclosure. The storage medium includes various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read Only Memory (ROM), a magnetic disk, or an optical disk. Thus, embodiments of the present disclosure are not limited to any specific hardware, software, or firmware, or any combination of the three.

The disclosed embodiments provide a computer device comprising a memory storing a computer program executable on the processor and a processor implementing some or all of the steps of the above method when the processor executes the program.

The disclosed embodiments provide a computer readable storage medium having stored thereon a computer program which when executed by a processor performs some or all of the steps of the above method. The computer readable storage medium may be transitory or non-transitory.

The disclosed embodiments provide a computer program comprising computer readable code which, when run in a computer device, performs some or all of the steps for implementing the methods described above.

Embodiments of the present disclosure provide a computer program product comprising a non-transitory computer-readable storage medium storing a computer program which, when read and executed by a computer, performs some or all of the steps of the above-described method. The computer program product may be realized in particular by means of hardware, software or a combination thereof. In some embodiments, the computer program product is embodied as a computer storage medium, and in other embodiments, the computer program product is embodied as a software product, such as a software development kit (Software Development Kit, SDK), or the like.

It should be noted herein that the above description of various embodiments is intended to emphasize the differences between the various embodiments, and that the same or similar features may be referred to each other. The above description of apparatus, storage medium, computer program and computer program product embodiments is similar to that of method embodiments described above, with similar advantageous effects as the method embodiments. For technical details not disclosed in the embodiments of the disclosed apparatus, storage medium, computer program and computer program product, please refer to the description of the embodiments of the disclosed method.

It should be noted that fig. 6 is a schematic diagram of a hardware entity of an electronic device in the present disclosure, as shown in fig. 6, the hardware entity of the electronic device 600 includes a processor 601, a communication interface 602, and a memory 603, where:

the processor 601 generally controls the overall operation of the electronic device 600.

The communication interface 602 may enable the electronic device to communicate with other terminals or servers over a network.

The memory 603 is configured to store instructions and applications executable by the processor 601, and may also cache data (e.g., image data, audio data, voice communication data, and video communication data) to be processed or processed by various modules in the processor 601 and the electronic device 600, which may be implemented by a FLASH memory (FLASH) or a random access memory (Random Access Memory, RAM). Data transfer may be performed between the processor 601, the communication interface 602, and the memory 603 via the bus 604.

It should be appreciated that reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present disclosure. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. It should be understood that, in various embodiments of the present disclosure, the size of the sequence numbers of the steps/processes described above does not mean the order of execution, and the order of execution of the steps/processes should be determined by their functions and inherent logic, and should not constitute any limitation on the implementation of the embodiments of the present disclosure. The foregoing embodiment numbers of the present disclosure are merely for description and do not represent advantages or disadvantages of the embodiments.

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

In the several embodiments provided in the present disclosure, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is merely a logical function division, and there may be additional divisions of actual implementation, such as multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. In addition, the various components shown or discussed may be coupled or directly coupled or communicatively coupled to each other via some interface, whether indirectly coupled or communicatively coupled to devices or units, whether electrically, mechanically, or otherwise.

The units described as separate components may or may not be physically separate, and components displayed as units may or may not be physical units, may be located in one place or distributed on a plurality of network units, and may select some or all of the units according to actual needs to achieve the purpose of the embodiment.

In addition, each functional unit in each embodiment of the disclosure may be integrated in one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated in one unit, where the integrated units may be implemented in a form of hardware or a form of hardware plus a form of software functional unit.

It will be appreciated by those of ordinary skill in the art that implementing all or part of the steps of the above method embodiments may be implemented by hardware associated with program instructions, where the above program may be stored in a computer readable storage medium, where the program when executed performs the steps comprising the above method embodiments, where the above storage medium includes various media that may store program code, such as a removable storage device, a Read Only Memory (ROM), a magnetic disk, or an optical disk.

Or the integrated units of the present disclosure may be stored in a computer readable storage medium if implemented in the form of software functional modules and sold or used as a stand-alone product. Based on such understanding, the technical solutions of the present disclosure may be embodied essentially or in part in a form of a software product stored in a storage medium, including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the methods described in the embodiments of the present disclosure. The storage medium includes various media capable of storing program codes such as a removable storage device, a ROM, a magnetic disk, or an optical disk.

The foregoing is merely an embodiment of the present disclosure, but the protection scope of the present disclosure is not limited thereto, and any person skilled in the art can easily think about the changes or substitutions within the technical scope of the present disclosure, and should be covered by the protection scope of the present disclosure.

Claims (10)

1. A merchandise recommendation method comprising:

Generating a first commodity list according to a first navigation tag, wherein the first navigation tag represents a navigation tag determined from a plurality of navigation tags based on user operation;

And updating the first commodity list according to the first interaction information and the first navigation tag to obtain a second commodity list, wherein the first interaction information characterizes information input by a user through an interaction window provided by the intelligent agent.

2. The method of claim 1, the updating the first merchandise list based on the first interaction information and the first navigation tag, comprising:

determining at least one first commodity label from a plurality of commodity labels according to the semantic information of the first interaction information;

And updating the first commodity list according to the at least one first commodity label and at least one second commodity label corresponding to the first navigation label to obtain the second commodity list.

3. The method of claim 2, wherein the first interactive information comprises at least one of natural language, voice, and pictures;

the determining at least one first commodity label from the plurality of commodity labels according to the semantic information of the first interaction information comprises at least one of the following steps:

Determining at least one first commodity label matched with the first semantic information from the plurality of commodity labels;

determining at least one first commodity label matched with the second semantic information from the plurality of commodity labels;

carrying out attribute identification on commodity objects in the pictures to obtain a plurality of commodity attributes; at least one first merchandise tag that matches the plurality of merchandise attributes is determined from the plurality of merchandise tags.

4. The method of claim 2, wherein updating the first merchandise list according to the at least one first merchandise tag and the at least one second merchandise tag corresponding to the first navigation tag to obtain the second merchandise list comprises one of:

Determining the second commodity list based on the commodities in the first commodity list matched with the at least one first commodity label in response to the at least one second commodity label containing the at least one first commodity label;

Determining, in response to the at least one first merchandise tag comprising the at least one second merchandise tag and a third merchandise tag, the second merchandise list based on merchandise corresponding to the first merchandise list and the third merchandise tag;

In response to a fourth merchandise tag of the at least one second merchandise tag being opposite in semantic meaning to the at least one first merchandise tag, determining the second merchandise list from the at least one first merchandise tag and at least one second merchandise tag other than the fourth merchandise tag.

5. The method of claim 1, further comprising, prior to generating the list of recommended items according to the first navigation tag:

Generating a third commodity list according to second interaction information, wherein the second interaction information characterizes information input by a user through an interaction window provided by the intelligent body;

And generating a plurality of navigation tags according to the commodity tags corresponding to the commodities in the third commodity list.

6. The method according to any one of claims 1 to 5, further comprising, after updating the first commodity list according to the first interaction information and the first navigation tag to obtain the second commodity list:

determining a third commodity corresponding to the user operation based on the user operation for the second commodity list;

Determining commodity performance requirement information of the third commodity, and displaying and outputting commodity performance information of the third commodity based on a target tool corresponding to the commodity performance requirement information;

Or alternatively

And generating commodity recommendation suggestions based on the commodity labels corresponding to the third commodities and the user portrait information of the user.

7. The method according to any one of claims 1 to 5, further comprising, after updating the first commodity list according to the first interaction information and the first navigation tag to obtain the second commodity list:

The method comprises the steps of obtaining third information, wherein the third information represents navigation tags determined by a user from a plurality of navigation tags or interactive information input by the user through an interactive window provided by an agent;

And determining at least one fourth commodity to be deleted and/or added from the second commodity list based on the third information.

8. The method according to any one of claims 1 to 5, further comprising, after updating the first commodity list according to the first interaction information and the first navigation tag to obtain a second commodity list:

Determining an output strategy for the second commodity list according to the user portrayal information of the user, wherein the output strategy is characterized in that each commodity displays the type of the commodity parameter output;

And displaying and outputting the second commodity list according to the output strategy.

9. The method of any one of claims 1 to 5, wherein a display screen of an electronic device comprises a first area for displaying an interactive window provided by the agent and a second area for displaying the plurality of navigation tags and the second list of items;

The method further comprises the steps of:

and determining the area occupation ratio of the first area and the second area on the display screen according to the user portrait information of the user.

10. A merchandise recommendation system comprising:

the communication unit is used for acquiring user input information, wherein the user input information comprises a first navigation tag determined by a user from a plurality of navigation tags and first interaction information input by the user through an interaction window provided by an agent;

the commodity recommending unit is used for generating a first recommended commodity list according to the first navigation tag, and updating the first commodity list according to the first interaction information and the first navigation tag to obtain a second commodity list.