Commodity evaluation content generation method and device, storage medium and computer equipment

Abstract

The invention discloses a commodity evaluation content generation method and device, a storage medium and computer equipment. The method comprises the following steps: a client displays a commodity evaluation page in response to a commodity evaluation request, and displays an intelligent evaluation entry in the commodity evaluation page; the client side responds to a triggering operation of the intelligent evaluation entrance and obtains commodity information of the to-be-evaluated commodity and user evaluation information, and the user evaluation information comprises user scoring information and / or user evaluation keywords; the client sends the commodity information and the user evaluation information to the server; the server responds to the received commodity information of the to-be-evaluated commodity and the user evaluation information, generates commodity evaluation content through a large model based on the commodity information and the user evaluation information, and sends the commodity evaluation content to the client; and the client receives and displays the commodity evaluation content. According to the method, the generation efficiency and the generation quality of the commodity evaluation content can be improved, and the workload and the time cost of comment writing are reduced.

Description

Technical Field

[0001] This invention relates to the field of Internet technology, and in particular to a method, apparatus, storage medium, and computer equipment for generating product review content. Background Technology

[0002] With the rapid development of internet technology, users have long been accustomed to purchasing the goods they need through various e-commerce platforms. Currently, after receiving the goods, users can leave reviews, which can cover various aspects such as the user's experience, areas for improvement for the merchant, or purchasing suggestions for other users.

[0003] For users, the writing requirements for product reviews discourage many from writing them. According to incomplete statistics, on a certain instant delivery e-commerce platform, only less than 40% of users have a habit of actively writing reviews, and more than 60% of these users give up because they dislike writing reviews or find it troublesome. However, for e-commerce platforms, product quality largely depends on reviews for confirmation.

[0004] In existing technologies, to encourage users to write product reviews, e-commerce platforms offer rewards or pre-generate keywords for users to select and directly fill in their reviews. However, even with pre-generated keywords, it's difficult to generate high-quality reviews that benefit other users when they are unwilling to write them. In this scenario, the time cost for users to write reviews is high, resulting in low efficiency and satisfaction, and a lack of motivation to continuously generate high-quality reviews for the platform. Summary of the Invention

[0005] In view of this, this application provides a method, apparatus, storage medium and computer equipment for generating product review content, with the main purpose of solving the technical problems of low efficiency in generating product reviews and poor user review experience.

[0006] According to a first aspect of the present invention, a method for generating product review content is provided, the method comprising:

[0007] The client responds to the product review request by displaying the product review page, and the smart review entry is displayed on the product review page;

[0008] The client responds to the trigger operation of the intelligent evaluation entry to obtain product information and user evaluation information of the product to be evaluated, wherein the user evaluation information includes user rating information and / or user evaluation keywords;

[0009] The client sends the product information and the user review information to the server;

[0010] The server responds to receiving product information and user review information of the product to be evaluated, generates product review content based on the product information and user review information through a large model, and sends the product review content to the client;

[0011] The client receives and displays the product review content.

[0012] Optionally, the client displays the smart review entry on the product review page, including: the client displays the smart review entry at at least one preset position on the product review page through at least one preset component, wherein the preset component is displayed in a static and / or dynamic manner, and the preset component is composed of at least one element among text, image, and icon.

[0013] Optionally, the client displays a smart review entry on the product review page, including: the client responds to the input operation of review content, displays a keyboard component, and displays the smart review entry at a preset position within a preset range of the keyboard component, wherein the smart review entry is removed from display when the keyboard component is hidden.

[0014] Optionally, the user rating information includes first rating information and / or second rating information; then, the client obtaining the user rating information includes: the client obtaining the user's first rating information for the order to be rated and / or the user's second rating information for at least one item in the order to be rated, wherein the user rating information includes at least one of rating value information, rating component operation information, like component operation information, and dislike component operation information.

[0015] Optionally, the client obtains the user evaluation keywords by: the client responding to an input operation and / or selection operation of at least one evaluation keyword to determine the user evaluation keywords, wherein the user evaluation keywords include evaluation content arbitrarily input by the user and / or evaluation keywords of at least one dimension arbitrarily selected by the user.

[0016] Optionally, the method further includes: the client responding to the trigger operation of the intelligent evaluation entry point by receiving at least one dimension of evaluation keywords sent by the server; the client displaying the at least one dimension of evaluation keywords, wherein the dimensions of the evaluation keywords include at least one of the following: product dimension, ingredients dimension, attribute dimension, packaging dimension, delivery dimension, user experience dimension, and language style dimension.

[0017] Optionally, the client displays evaluation keywords of at least one dimension, including: the client performing structured processing on the evaluation keywords according to the dimensions of the evaluation keywords, and displaying the structured evaluation keywords; and / or the client arranging the evaluation keywords of the at least one dimension according to preset arrangement rules and / or random sorting rules, and displaying the arranged evaluation keywords.

[0018] Optionally, the client displays evaluation keywords of at least one dimension, including: the client scrolls to display the evaluation keywords of at least one dimension in response to a sliding operation of the evaluation keywords; and / or, the client scrolls to display the evaluation keywords of at least one dimension at a preset scrolling speed.

[0019] Optionally, the client sends the product information and the user rating information to the server, including: the client responding to the trigger operation of the smart rating entry to send the product information and the user rating information to the server; and / or, the client responding to the trigger operation of the rating content generation control to send the input and / or selected user rating keywords to the server.

[0020] Optionally, the client displays the product review content, including: the client displays first product review content and / or second product review content, wherein the first product review content is generated based on the product information and the user rating information, and the second product review content is generated based on the product information, the user rating information, and the user review keywords.

[0021] Optionally, the client displays the product review content, including: the client displays at least one product review content; the method further includes: in response to a selection operation of any product review content, the client fills the selected product review content into the product review page.

[0022] Optionally, the method further includes: the client responding to the editing operation of the product review content by displaying the edited product review content.

[0023] Optionally, the client displays the product review content, including: the client displays at least one product review in at least one language style, wherein the product review content is allocated according to a preset language style occupancy ratio.

[0024] Optionally, the method further includes: the client displaying at least one language style keyword; and the client, in response to the selection of any one of the language style keywords, displaying at least one product review corresponding to the selected language style keyword.

[0025] Optionally, the method further includes: the client displaying evaluation keywords of at least one dimension; the client, in response to the selection operation of any one of the evaluation keywords, sending the selected target evaluation keyword to the server; the server, in response to receiving the target evaluation keyword, generating product evaluation content through a large model based on the product information, the user evaluation information, and the target evaluation keyword, and sending the updated product evaluation content to the client; and the client receiving and displaying the updated product evaluation content.

[0026] Optionally, the method further includes: the client responding to a request to change the product review content by sending the request to change the product review content to the server; the server responding to receiving the request to change the product review content by regenerating the product review content based on the product information and the user review information using a large model, and sending the updated product review content to the client; the client receiving and displaying the updated product review content.

[0027] Optionally, the server generates product review content based on the product information and the user review information using a large model, including: the server generating at least one dimension of review keywords based on the product information and the user review information using the large model, and generating at least one product review based on the at least one dimension of review keywords; and / or, the server matching at least one product review in the review database based on the product information and the user review information, wherein the review database stores historical product review content generated by the large model based on historical product information and historical user review information.

[0028] Optionally, the server generates at least one product review based on the evaluation keywords of the at least one dimension, including: the server generates at least one product review based on at least one language style keyword and the evaluation keywords of the at least one dimension.

[0029] Optionally, the method further includes: the server matching at least one dimension of evaluation keywords in a keyword database based on the product information, and sending the evaluation keywords to the client; and / or, the server generating at least one dimension of evaluation keywords based on the product information and the user evaluation information using the big model, and sending the evaluation keywords to the client.

[0030] Optionally, the training method of the large model includes: the server acquiring multiple historical user reviews corresponding to stores of multiple product categories; the server extracting review keywords from the historical user reviews and setting at least one dimension of review tags for the review keywords; the server pre-training the original large model based on the review keywords and the review tags corresponding to the review keywords to obtain the trained large model.

[0031] Optionally, the training method for the large model further includes: the server acquiring multiple historical user reviews in various language styles, wherein each historical user review is assigned a corresponding language style tag; the server pre-training the original large model based on the historical review content and the corresponding language style tags to obtain a trained large model.

[0032] According to a second aspect of the present invention, a method for generating product review content is provided, the method comprising:

[0033] In response to a product review request, the product review page is displayed, and the smart review entry is shown on the product review page;

[0034] In response to the triggering operation of the intelligent evaluation entry, the product information and user evaluation information of the product to be evaluated are obtained, and the product information and user evaluation information are sent, wherein the user evaluation information includes user rating information and / or user evaluation keywords;

[0035] Receive and display the product review content, wherein the product review content is generated based on the product information and the user review information through a large model.

[0036] Optionally, displaying the intelligent review entry on the product review page includes: displaying the intelligent review entry at at least one preset position on the product review page through at least one preset component, wherein the preset component is displayed in a static and / or dynamic manner, and the preset component consists of at least one element among text, images, and icons.

[0037] Optionally, displaying the intelligent evaluation entry on the product evaluation page includes: responding to the input operation of evaluation content, displaying a keyboard component, and displaying the intelligent evaluation entry at a preset position within a preset range of the keyboard component, wherein the intelligent evaluation entry is removed from display when the keyboard component is hidden.

[0038] Optionally, the user rating information includes first rating information and / or second rating information; then obtaining the user rating information includes: obtaining the user's first rating information for the order to be rated and / or the user's second rating information for at least one item in the order to be rated, wherein the user rating information includes at least one of rating value information, rating component operation information, like component operation information, and dislike component operation information.

[0039] Optionally, obtaining the user evaluation keywords includes: in response to an input operation and / or selection operation of at least one evaluation keyword, determining the user evaluation keywords, wherein the user evaluation keywords include evaluation content arbitrarily input by the user and / or evaluation keywords of at least one dimension arbitrarily selected by the user.

[0040] Optionally, the method further includes: in response to the triggering operation of the intelligent evaluation entry, receiving at least one dimension of evaluation keywords sent by the server; displaying the at least one dimension of evaluation keywords, wherein the dimensions of the evaluation keywords include at least one of product dimension, ingredient dimension, attribute dimension, packaging dimension, delivery dimension, user experience dimension, and language style dimension.

[0041] Optionally, displaying evaluation keywords of at least one dimension includes: structuring the evaluation keywords according to their dimensions and displaying the structured evaluation keywords; and / or arranging the evaluation keywords of the at least one dimension according to preset arrangement rules and / or random sorting rules and displaying the arranged evaluation keywords.

[0042] Optionally, displaying evaluation keywords of at least one dimension includes: scrolling to display the evaluation keywords of at least one dimension in response to a sliding operation of the evaluation keywords; and / or scrolling to display the evaluation keywords of at least one dimension at a preset scrolling speed.

[0043] Optionally, sending the product information and the user rating information includes: in response to a trigger operation of the intelligent rating entry, sending the product information and the user rating information; and / or, in response to a trigger operation of the rating content generation control, sending the input and / or selected user rating keywords.

[0044] Optionally, displaying the product review content includes: displaying first product review content and / or second product review content, wherein the first product review content is generated based on the product information and the user rating information, and the second product review content is generated based on the product information, the user rating information, and the user review keywords.

[0045] Optionally, displaying the product review content includes: displaying at least one product review content; the method further includes: in response to a selection operation of any product review content, filling the selected product review content into the product review page.

[0046] Optionally, the method further includes: displaying the edited product review content in response to the editing operation of the product review content.

[0047] Optionally, displaying the product review content includes: displaying at least one product review content in at least one language style, wherein the product review content is allocated according to a preset language style occupancy ratio.

[0048] Optionally, the method further includes: displaying at least one language style keyword; and in response to the selection of any one of the language style keywords, displaying at least one product review corresponding to the selected language style keyword.

[0049] Optionally, the method further includes: displaying evaluation keywords for at least one dimension; in response to the selection of any of the evaluation keywords, sending the selected target evaluation keyword; receiving and displaying updated product evaluation content, wherein the updated product evaluation content is generated based on the product information, the user evaluation information, and the target evaluation keyword through a large model.

[0050] Optionally, the method further includes: in response to a request to change product review content, sending a request to change the product review content; receiving and displaying the updated product review content, wherein the updated product review content is regenerated based on the product information and the user review information through a large model.

[0051] According to a third aspect of the present invention, a method for generating product review content is provided, the method comprising:

[0052] In response to receiving product information and user review information of the product to be reviewed, product review content is generated through a large model based on the product information and the user review information. The product information and the user review information are received after being triggered by the smart review entry on the product review page of the client. The user review information includes user rating information and / or user review keywords.

[0053] The product review content is sent to the client so that the client can receive and display the product review content.

[0054] Optionally, in response to receiving the target evaluation keywords sent by the client, product evaluation content is generated through a large model based on the product information, the user evaluation information, and the target evaluation keywords, and the updated product evaluation content is sent to the client; and / or, in response to receiving a request to change the product evaluation content, product evaluation content is generated again through a large model based on the product information and the user evaluation information, and the updated product evaluation content is sent to the client.

[0055] Optionally, generating product review content using a large model based on the product information and user review information includes: generating at least one dimension of review keywords using the large model based on the product information and user review information, and generating at least one product review based on the at least one dimension of review keywords; and / or matching at least one product review in the review database based on the product information and user review information, wherein the review database stores historical product review content generated by the large model based on historical product information and historical user review information.

[0056] Optionally, generating at least one product review based on the evaluation keywords of the at least one dimension includes: generating at least one product review based on at least one language style keyword and the evaluation keywords of the at least one dimension.

[0057] Optionally, the method further includes: matching at least one dimension of evaluation keywords in a keyword database based on the product information, and sending the evaluation keywords to the client; and / or generating at least one dimension of evaluation keywords based on the product information and the user evaluation information using the big model, and sending the evaluation keywords to the client.

[0058] Optionally, the training method of the large model includes: obtaining multiple historical user reviews corresponding to stores of multiple product categories; extracting review keywords from the historical user reviews and setting at least one dimension of review tags for the review keywords; and pre-training the original large model based on the review keywords and the review tags corresponding to the review keywords to obtain the trained large model.

[0059] Optionally, the training method for the large model further includes: acquiring multiple historical user reviews in various language styles, wherein each historical user review is assigned a corresponding language style label; and pre-training the original large model based on the historical review content and the corresponding language style labels to obtain the trained large model.

[0060] According to a fourth aspect of the present invention, a product review content generation apparatus is provided, the apparatus comprising:

[0061] The evaluation entry display module is used to respond to product evaluation requests, display the product evaluation page, and display the intelligent evaluation entry on the product evaluation page;

[0062] The evaluation information acquisition module is used to respond to the trigger operation of the intelligent evaluation entry, acquire the product information and user evaluation information of the product to be evaluated, and send the product information and the user evaluation information, wherein the user evaluation information includes user rating information and / or user evaluation keywords;

[0063] The evaluation content display module is used to receive and display the product evaluation content, wherein the product evaluation content is generated based on the product information and the user evaluation information through a large model.

[0064] According to a fifth aspect of the present invention, a product review content generation apparatus is provided, the apparatus comprising:

[0065] The evaluation content generation module is used to respond to receiving product information and user evaluation information of the product to be evaluated, and to generate product evaluation content based on the product information and the user evaluation information through a large model. The product information and the user evaluation information are received after being triggered by the smart evaluation entry on the product evaluation page of the client. The user evaluation information includes user rating information and / or user evaluation keywords.

[0066] The evaluation content sending module is used to send the product evaluation content to the client so that the client can receive and display the product evaluation content.

[0067] According to a sixth aspect of the present invention, a storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the above-described method for generating product evaluation content.

[0068] According to a seventh aspect of the present invention, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described method for generating product evaluation content.

[0069] This invention provides a method, apparatus, storage medium, and computer device for generating product review content. When a user initiates a product review request, a product review page can be displayed on a client, including a smart review entry point. When the user triggers the smart review entry point, the client responds by obtaining product information and user review information, and sends these to a server. The server then uses this information to generate product review content based on a large-scale model, and sends the generated content back to the client for display. This method automatically initiates product review writing by triggering the smart review entry point during the user's review process, effectively improving the efficiency of product review content generation, reducing the workload and time cost for users, and contributing to a higher proportion of product reviews. Furthermore, by leveraging the text understanding and generation capabilities of a large-scale model, this method effectively improves the quality of generated product review content and enhances the user experience, thus motivating users to continuously produce high-quality reviews.

[0070] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0071] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0072] Figure 1 A flowchart illustrating a method for generating product review content according to an embodiment of the present invention is shown;

[0073] Figure 2 A flowchart illustrating another method for generating product review content provided by an embodiment of the present invention is shown;

[0074] Figure 3 The illustration shows a scenario diagram of a product review content generation method provided by an embodiment of the present invention;

[0075] Figure 4 The illustration shows a scenario diagram of a product review content generation method provided by an embodiment of the present invention;

[0076] Figure 5 The illustration shows a scenario diagram of a product review content generation method provided by an embodiment of the present invention;

[0077] Figure 6 The illustration shows a scenario diagram of a product review content generation method provided by an embodiment of the present invention;

[0078] Figure 7 The illustration shows a scenario diagram of a product review content generation method provided by an embodiment of the present invention;

[0079] Figure 8 The illustration shows a scenario diagram of a product review content generation method provided by an embodiment of the present invention;

[0080] Figure 9 The illustration shows a scenario diagram of a product review content generation method provided by an embodiment of the present invention;

[0081] Figure 10 The illustration shows a scenario diagram of a product review content generation method provided by an embodiment of the present invention;

[0082] Figure 11 This diagram illustrates the structure of a product review content generation device provided in an embodiment of the present invention.

[0083] Figure 12 This invention provides a schematic diagram of another product review content generation device according to an embodiment of the invention.

[0084] Figure 13 The diagram shows a schematic representation of a computer device for implementing a method for generating product review content, according to an embodiment of the present invention. Detailed Implementation

[0085] The present invention will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the present application can be combined with each other.

[0086] In one embodiment, such as Figure 1 As shown, a method for generating product review content is provided. Taking the application of this method to computer devices such as client and server as an example, the method includes the following steps:

[0087] 101. The client responds to the product review request, displays the product review page, and shows the smart review entry on the product review page.

[0088] Specifically, after a user purchases a product on an e-commerce platform, they can initiate a product review request via the client after receiving the product. In this scenario, the client responds to the user's review request, displaying a product review page. This page includes input boxes for the user to enter their review content, a rating component for the user to rate the product, and other content, allowing the user to rate and comment on the product. Furthermore, the product review page also displays a smart review entry point, which, when triggered, enters smart review mode.

[0089] In this embodiment, a user can purchase one or more items in the same order and rate at least one item simultaneously, or they can rate each item individually. After the client responds to the user's product rating request, it can either receive the user's rating on a front-end page before displaying the product rating page, or it can directly display the product rating page and receive the user's rating there. Furthermore, the smart rating entry can be displayed in one or more preset locations on the product rating page, and its display method can be designed according to actual conditions. It is understood that this embodiment does not specifically limit the product rating method, the display method of the product rating page, or the display method of the smart rating entry.

[0090] 102. The client responds to the trigger operation of the smart evaluation entry to obtain the product information and user evaluation information of the product to be evaluated. The user evaluation information includes user rating information and / or user evaluation keywords.

[0091] Specifically, in response to the trigger operation of the intelligent evaluation entry, the client can obtain product information and user evaluation information of the product to be evaluated. The user evaluation information may include user rating information and / or user evaluation keywords, while the product information may include product name information, product attribute information, merchant information, and other information related to the purchased product. In this embodiment, product information can be obtained through the product identifier of the product to be evaluated, and user evaluation information can be obtained through the operation information of the rating component, keywords entered in the input box, and keywords selected by the user.

[0092] In this embodiment, user review information can be generated by the user after initiating a product review request, or it can be generated by the user after triggering the smart review entry point, and then retrieved by the client. Specifically, the client can retrieve the product information and user review information immediately after the user triggers the smart review entry point; and / or, it can retrieve them after the user triggers the smart review entry point and then waits for the user to initiate another review content generation request. It is understood that this embodiment does not specifically limit the method of obtaining product information and user review information.

[0093] 103. The client sends product information and user review information to the server.

[0094] Specifically, after the client obtains the product information and user review information of the product to be reviewed, it can send the obtained product information and user review information to the server. In this embodiment, the client can obtain the product information and user review information and send them to the server immediately after the user triggers the smart review entry; and / or, it can wait for the user to initiate another review content generation request after the user triggers the smart review entry before obtaining the product information and user review information and sending them to the server. It is understood that this embodiment does not specifically limit the method of sending product information and user review information.

[0095] 104. Upon receiving the product information and user review information of the product to be evaluated, the server generates product review content based on the product information and user review information through a large model, and sends the product review content to the client.

[0096] Among them, large models refer to advanced artificial intelligence models that can handle complex tasks and large amounts of data, such as large language models (LLM), computer vision models (CV), and large multimodal models (LMMs).

[0097] Specifically, after receiving product information and user reviews of the product to be evaluated, the server can generate product review content based on the product information and user reviews using a pre-trained large model, and then send the generated product review content to the client. In this embodiment, the large model can generate review content corresponding to the product information and user reviews using pre-designed prompts. For example, the large model can first generate multiple review keywords for multiple dimensions such as product packaging, taste, and ingredients using the product information and user reviews, and then generate product review content corresponding to the product and user reviews using the review keywords for each dimension. It is understood that this embodiment does not specifically limit the specific method by which the large model generates product review content.

[0098] In this embodiment, a large model generates product review content based on product information and user reviews. On the one hand, the large model's text understanding and text generation capabilities can be used to generate high-quality product review content, thus improving the quality of the product review content. On the other hand, the generated product review content can always revolve around the product being reviewed, as well as the user's rating of the product and / or the user's review keywords, so that the product review content can meet the user's expectations, thereby improving the user's satisfaction with the product review content.

[0099] 105. The client receives and displays product review content.

[0100] Specifically, after receiving product review content from the server, the client can display the product review content. In this embodiment, the client can display one or more product reviews sent by the server. If the client receives multiple product reviews, it can display some of them or all of them. The product review content can be displayed in a newly opened smart review page or directly on the product review page. Furthermore, the component for displaying product review content can be designed according to actual needs. It is understood that this embodiment does not specifically limit the display method of product review content.

[0101] The product review content generation method provided in this embodiment can automatically initiate the writing of product review content by triggering an intelligent review entry during the user's review writing process. This effectively improves the efficiency of product review content generation, reduces the workload and time cost for users writing reviews, and helps increase the proportion of product reviews. Furthermore, by leveraging the text understanding and text generation capabilities of a large model to generate product review content, the method can effectively improve the quality of generated product review content and enhance the user's review writing experience, thereby motivating users to continuously output high-quality review content.

[0102] In one embodiment, the method of displaying the smart review entry on the product review page in step 101 can be implemented in the following way: the client displays the smart review entry in at least one preset position on the product review page through at least one preset component, wherein the preset component can be displayed in a static and / or dynamic manner, and the preset component can be composed of at least one element among text, image, and icon.

[0103] In the above embodiments, reference is made to Figure 3 , Figure 5 and Figure 8 As shown by the dashed box in the scene diagram on the left, the client can display the smart review entry point in at least one preset location on the product review page. The component used to display the smart review entry point can consist of at least one element selected from text, images, and icons. Furthermore, the smart review entry point can be displayed statically in a fixed preset location on the product review page, or it can be dynamically displayed in response to changes in other components on the page, or its display can change according to the component's own dynamic display rules. These methods effectively improve the recognizability of the smart review entry point, thereby increasing user engagement with the smart review function.

[0104] In one embodiment, the method of displaying the smart evaluation entry on the product evaluation page in step 101 can also be implemented in the following way: the client responds to the input operation of the evaluation content, displays the keyboard component, and displays the smart evaluation entry at a preset position within a preset range of the keyboard component, wherein the smart evaluation entry can be eliminated when the keyboard component is hidden.

[0105] In the above embodiments, reference is made to Figure 5 As shown by the dotted box in the scenario diagram on the left, when a user enters a review on the product review page, the input box is triggered, displaying a cursor. Simultaneously, the keyboard component also appears on the product review page. To improve the recognizability of the smart review entry point in this scenario, it can be displayed at a preset location within a pre-defined range of the keyboard component. For example, it could be displayed above or inside the keyboard component, or at least in one of these locations. Users can then access the smart review mode by triggering this entry point. Furthermore, the smart review entry point can be hidden when the user hides the keyboard component, thus improving the page's simplicity. These methods effectively enhance the recognizability of the smart review entry point, thereby increasing user engagement with the smart review function.

[0106] In one embodiment, the method for obtaining user rating information in step 102 can be implemented as follows: the client obtains the user's first rating information for the order to be rated and / or the user's second rating information for at least one item in the order to be rated, wherein the user rating information may include the first rating information and / or the second rating information, and the user rating information may include at least one of the following: rating value information, rating component operation information, like component operation information, and dislike component operation information.

[0107] Specifically, users can rate orders and / or products after initiating a product review request, or after triggering the smart review entry. Correspondingly, the client can obtain the user rating information after the user rates the order and / or product. This user rating information can include ratings in at least one dimension; for example, user rating information could include the following: "Overall order 5 stars, packaging 5 stars, quality 5 stars, dish XXX is excellent."

[0108] In the above embodiments, reference is made to Figure 3 As shown in the scenario diagram on the left, users can rate the overall order, packaging, quality, and other dimensions using the rating component (i.e., the first rating information). This rating can be obtained through the operation information of the rating component. Additionally, users can also rate the quality of individual items in the order using the "like" and "dislike" components (i.e., the second rating information). This rating can be obtained through the operation information of the "like" and / or "dislike" components. In this embodiment, users can rate the order and the actual condition of the items, and this rating will serve as an important basis for the large model to generate product reviews. For example, when a user gives the overall order a score of 1, the large model can generate reviews such as "not fresh" or "not tasty"; when a user gives the overall order a score of 5, the large model can generate reviews such as "tasty" or "superb". By obtaining user reviews of the order and the items, the large model can generate product reviews that better match the user's intentions, thereby improving the accuracy of the product reviews and user satisfaction.

[0109] In one embodiment, the method for obtaining user evaluation keywords in step 102 can be implemented as follows: the client responds to an input operation and / or selection operation of at least one evaluation keyword to determine the user evaluation keywords, wherein the user evaluation keywords may include any evaluation content input by the user and / or at least one dimension of evaluation keywords arbitrarily selected by the user.

[0110] In the above embodiments, reference is made to Figure 6 and Figure 9As shown by the dashed box in the scene diagram on the left, the client can determine user review keywords based on user input and / or selected review keywords. In this embodiment, user review keywords can serve as the foundational information for the large model to generate product review content; that is, the large model can expand and enrich based on user review keywords. This approach preserves the original meaning of the user review content while improving its richness and quality. Simultaneously, it reduces the time and tediousness of users manually writing reviews, increasing the enjoyment and satisfaction of user product reviews.

[0111] In one embodiment, the above-mentioned method for generating product review content may further include the following steps:

[0112] 201. The client responds to the trigger operation of the smart evaluation entry and receives evaluation keywords for at least one dimension sent by the server.

[0113] 202. The client displays evaluation keywords for at least one dimension, wherein the dimensions of the evaluation keywords include at least one of the following dimensions: product dimension, ingredients dimension, attribute dimension, packaging dimension, delivery dimension, user experience dimension, and language style dimension.

[0114] In the above embodiments, reference is made to Figure 5 and Figure 8 As shown in the scene diagram on the right, after the user triggers the intelligent evaluation entry, the client can receive evaluation keywords of at least one dimension sent by the server and display the evaluation keywords of each dimension. In this embodiment, the client can display evaluation keywords of multiple dimensions related to product information, such as product name, product ingredients, product temperature, product taste, product quantity, product freshness, etc., and can also display evaluation keywords related to the language style of the product evaluation content, such as more interesting, more poetic, more concise, more professional, etc. Users can select evaluation keywords themselves as user evaluation keywords based on the product itself and their experience using the product, so that the large model can generate product evaluation content that better meets the user's expectations based on the evaluation keywords selected by the user. The above method, by displaying evaluation keywords of at least one dimension for the user to select, can improve the efficiency of user evaluation keyword generation, help users broaden their evaluation ideas, improve the richness of product evaluation content, and make the generated product evaluation content more in line with the user's expectations.

[0115] In one embodiment, the method for displaying evaluation keywords for at least one dimension in step 202 can be implemented in at least one of the following ways:

[0116] 202-1. The client performs structured processing on the evaluation keywords according to the dimensions of the evaluation keywords, and then displays the structured evaluation keywords.

[0117] Specifically, refer to Figure 5 As shown in the scene diagram on the right, the client can structure review keywords according to multiple dimensions such as product attributes, product name, and language style, and then display the structured review keywords. In this embodiment, the dimensions for structuring review keywords can be selected and set according to actual conditions, and this embodiment does not impose specific restrictions. The above method, by structuring review keywords according to their dimensions and displaying the structured review keywords, can improve the overall cleanliness of the page and help users quickly select the review keywords they need, thereby improving the efficiency of review keyword selection.

[0118] 202-2. The client arranges the evaluation keywords for at least one dimension according to preset sorting rules and / or random sorting rules, and displays the sorted evaluation keywords.

[0119] Specifically, refer to Figure 8 As shown in the scene diagram on the right, the client can arrange multiple evaluation keywords according to preset rules, such as a three-row arrangement. Alternatively, it can randomly sort multiple evaluation keywords and then display the ordered and randomly sorted rows of keywords. In this embodiment, multiple evaluation keywords can also be sorted according to preset sorting rules, such as sorting by the dimension or number of characters. Furthermore, multiple evaluation keywords can be arranged in other ways besides row and column arrangements, such as by keyword selection frequency or displaying evaluation keywords as hot words. The arrangement and sorting methods of evaluation keywords can be designed according to actual conditions; this embodiment will not provide examples of each method and will not impose specific limitations. This embodiment, by arranging evaluation keywords in an ordered and / or random manner, can effectively improve the interest in selecting evaluation keywords, thereby helping to enhance the user's evaluation experience.

[0120] 202-3. The client responds to the sliding operation of the evaluation keywords and scrolls to display the evaluation keywords of at least one dimension.

[0121] Specifically, when there are a large number of review keywords, a single page may not be able to display all of them. In this scenario, users can swipe through the page to view and select from the hidden review keywords. This method effectively improves the convenience of selecting review keywords.

[0122] 202-4. The client displays evaluation keywords for at least one dimension according to the preset scrolling speed.

[0123] Specifically, in addition to passively responding to the swiping action of evaluation keywords, the client can also scroll the evaluation keywords at a preset speed. This eliminates the need for swiping, improving the display efficiency of evaluation keywords and enhancing the fun of the intelligent evaluation function.

[0124] In one embodiment, the method of sending product information and user review information to the server in step 103 can be implemented in at least one of the following ways:

[0125] 103-1. The client responds to the trigger operation of the smart rating entry and sends product information and user rating information to the server.

[0126] 103-2. In response to the triggering operation of the evaluation content generation control, the client sends the input and / or selected user evaluation keywords to the server.

[0127] Specifically, the client can obtain product information and user review information after the user triggers the smart review entry point, and send the product information and user review information to the server; and / or, it can wait for the user to trigger the review content generation control again after the user triggers the smart review entry point before obtaining product information and user review information and sending the product information and user review information to the server.

[0128] In the above embodiments, reference is made to Figure 5 As shown in the scenario diagram, after a user triggers the intelligent rating entry point, the client can send product information and user rating information to the server. The server then calls a large model to generate product rating content based on the product information and user rating information, and sends it to the client for display. Further, referring to... Figure 6 As shown in the scenario diagram, after the user selects and inputs evaluation keywords, the evaluation content generation control "Generate" can be triggered to send the input and selected user evaluation keywords (at this time, product information and user rating information can be sent again, or not) to the server. This allows the server to call the large model again, generate product evaluation content using product information, user rating information, and user evaluation keywords, and send it to the client for display.

[0129] In one embodiment, the method for displaying product review content in step 105 can be implemented in at least one of the following ways:

[0130] 105-1. The client displays the first product review content, which is generated based on product information and user rating information.

[0131] 105-2. The client displays the second product review content, which is generated based on product information, user rating information, and user review keywords.

[0132] In the above embodiments, reference is made to Figure 6 As shown in the scenario diagram, the client can display the first product review content generated from product information and user rating information on the client side. Figure 6 The review content below "I want to use the recommendations" in the middle, and the second product review content generated from product information, user rating information and user review keywords. Figure 6 (The evaluation content is shown in the dashed box). In this embodiment, referring to other scenario diagrams, the client may also display only the first product evaluation content or the second product evaluation content; this embodiment does not make specific limitations here. By displaying the first product evaluation content and / or the second product evaluation content, the client can increase the richness of the product evaluation content, thereby making it easier for users to select product evaluation content that is closer to their expectations, thus improving the user's interactive experience and evaluation experience.

[0133] In one embodiment, the method for displaying product review content in step 105 can also be implemented in the following ways:

[0134] 105-3. The client displays at least one product review.

[0135] Specifically, refer to Figures 3 to 10 As shown in the scenario diagram, the client can display at least one product review, allowing users to select one that best matches their expectations, thereby improving the user's interactive and review experience. In this embodiment, at least one product review may include at least one first product review and / or at least one second product review.

[0136] In this scenario, the above-mentioned method for generating product review content may also include at least one of the following steps:

[0137] 106. The client responds to any edit operation of product review content and displays the edited product review content.

[0138] Specifically, refer to Figure 9 As shown in the scenario illustration, users can edit any product review displayed on the client and view the edited review. This method allows edited product reviews to more closely reflect users' true thoughts and significantly reduces the workload and time spent manually writing reviews.

[0139] 107. When the client responds to the selection of any product review content, the selected product review content will be filled into the product review page.

[0140] Specifically, refer to Figure 4 , Figure 7 and Figure 10 As shown in the scenario diagram, when a user selects any product review, the client can automatically populate the product review page with the selected review. This effectively increases the freedom of selecting product review content and improves the efficiency of product review generation.

[0141] In one embodiment, the method for displaying product review content in step 105 can also be implemented in the following ways:

[0142] 105-4. The client displays at least one product review in at least one language style, wherein the product review content is allocated according to a preset language style occupancy ratio.

[0143] Specifically, the client can display product reviews in one or more language styles according to a preset proportion. This proportion can be fixed, adjusted based on user preferences, or dynamically changed based on user preferences; this embodiment does not impose specific limitations. For example, suppose the large model generates 17 product reviews in five language styles at once: 5 "more interesting," 5 "more concise," 5 "more poetic," 2 "more professional," and 2 "more user-friendly." In this way, users can view product reviews in multiple language styles and choose their preferred ones, thereby increasing the richness and interest of the product reviews and improving the user experience.

[0144] In one embodiment, the above-mentioned method for generating product review content may further include the following steps:

[0145] 301. The client displays at least one language style keyword.

[0146] 302. In response to the selection of any language style keyword, the client displays at least one product review corresponding to the selected language style keyword.

[0147] In the above embodiments, the client can display at least one language style keyword. When the user selects any language style keyword, the client can switch the product review content on the page to the product review content corresponding to the selected language style keyword. In this embodiment, the product review content displayed by the client can be product review content with a certain language style selected from existing product review content, or it can be product review content with a certain language style regenerated by the server based on the user's selected language style keyword, and then sent to the client for display. This embodiment does not make specific limitations. The above method can improve the generation efficiency of product review content by displaying product review content corresponding to the user's selected language style keyword, making it easier for users to select their favorite product review content and improving the user experience.

[0148] In one embodiment, the above-mentioned method for generating product review content may further include the following steps:

[0149] 401. The client displays evaluation keywords for at least one dimension.

[0150] 402. In response to the selection of any evaluation keyword, the client sends the selected target evaluation keyword to the server.

[0151] 403. Upon receiving the target evaluation keywords, the server generates product evaluation content based on product information, user evaluation information, and the target evaluation keywords using a large model, and sends the updated product evaluation content to the client.

[0152] 404. The client receives and displays the updated product review content.

[0153] In the above embodiments, in addition to displaying product review content, the client can also display review keywords of at least one dimension. When a user selects any review keyword, the client can send the selected keyword to the server, so that the server can regenerate the product review content based on the user's selected keyword and send it to the client for display. In this embodiment, the client can further expand and refine the generated product review content based on the user's selected review keywords to improve the quality of the review content; it can also regenerate new product review content based on product information, user review information, and target review keywords for display. This embodiment does not specifically limit this. The above method, by supporting users to further select review keywords when viewing product review content, can further generate product review content that meets user expectations based on existing product review content, thereby improving the content quality of product review content and enhancing the user review experience.

[0154] In one embodiment, the above-mentioned method for generating product review content may further include the following steps:

[0155] 501. In response to a request to change the product review content, the client sends the request to the server.

[0156] 502. Upon receiving a request to change the product review content, the server regenerates the product review content based on the product information and user review information using a large model, and sends the updated product review content to the client.

[0157] 503. The client receives and displays the updated product review content.

[0158] In the above embodiments, when a user is not satisfied with the existing product reviews or wants to view more product reviews, they can initiate a request to change the product reviews in order to view more product reviews. For example, the user can click on... Figure 3 The "Change" control in the scene illustration on the right initiates a request to change the product review content. After the user initiates this request, the server can regenerate at least one new product review based on existing product and user review information, and then send this new review to the client for display. This method can improve the richness of product review content and increase user satisfaction with it.

[0159] In one embodiment, the method for generating product review content using a large model in step 104 can be implemented in at least one of the following ways:

[0160] 104-1. The server uses a large model to generate evaluation keywords for at least one dimension based on product information and user review information, and generates at least one product review based on the evaluation keywords for at least one dimension.

[0161] 104-2. Based on product information and user review information, the server matches at least one product review from the review database. The review database stores historical product reviews generated by the large model based on historical product information and historical user review information.

[0162] In the above embodiments, the server can generate product review content through two methods: online processing and offline processing of the large model. Step 104-1 describes the method for generating product review content online using the large model, and step 104-2 describes the method for generating product review content offline using the large model and matching it to the application. It is understood that the process of generating product review content based on product information and user review information is basically the same under both methods. However, the specific method by which the server obtains the product review content differs when the client requests it. Each method has its advantages and applicable scenarios. Online processing ensures the novelty of the product review content, while offline processing increases the generation speed and reduces the overhead cost of the large model. Therefore, in practical applications, one or both methods can be selected to obtain product review content depending on the actual situation. For example, the server can first match product review content in the review database based on product information and user review information. If no relevant product review content is found, the server can then call the large model to generate product review content in real time. Alternatively, the server can generate part of the product review content through online processing and return the other part of the product review content to the client through offline processing.

[0163] Specifically, the server generates product review content based on product information and user review information in the following way: After receiving product information and user review information sent by the client, the server can generate at least one dimension of review keywords based on the product information and user review information, and then generate at least one piece of product review content based on the at least one dimension of review keywords. In this embodiment, the dimension of the review keywords generated by the large model can be pre-set in the prompt words of the large model or limited to a certain range of selection by the large model. In addition, the review keywords can also be limited to selection from a pre-established keyword library, thereby limiting the dimension and selection range of the review keywords. In this embodiment, the review keywords are generated based on product information and user review information. Therefore, the review keywords can better reflect various characteristics of the product, such as name characteristics, attribute characteristics, packaging characteristics, etc. At the same time, the review keywords can also better reflect the user's experience with the product, such as taste experience, temperature experience, satisfaction experience, etc. On this basis, by generating product review content based on at least one dimension of review keywords, the generated product review content can always revolve around the review keywords, thereby effectively improving the stability and accuracy of the product review content, and thus improving the content quality of the product review content and user satisfaction with the product review content.

[0164] In one embodiment, the method for generating product review content by evaluation keywords in step 104-1 can also be implemented in the following way: the server generates at least one product review content based on at least one language style keyword and at least one dimension of evaluation keywords.

[0165] In the above embodiments, when the server generates product review content using evaluation keywords, it can also add language style keywords to the evaluation keywords, thereby generating product review content with at least one language style. In this embodiment, language style keywords can be pre-set in the prompts of the large model, or they can be selected by the user in the client and sent to the server; this embodiment does not impose specific limitations. By generating product review content based on language style keywords and evaluation keywords, the server can facilitate users in choosing their preferred product review content from multiple language styles, thereby improving the richness and interest of the product review content and ultimately enhancing the user experience.

[0166] In one embodiment, the above-mentioned method for generating product review content further includes at least one of the following steps:

[0167] 601. Based on product information, the server matches at least one dimension of evaluation keywords in the keyword database and sends the evaluation keywords to the client.

[0168] 602. The server generates evaluation keywords for at least one dimension based on product information and user review information using a large model, and sends the evaluation keywords to the client.

[0169] In the above embodiments, the server can generate at least one dimension of evaluation keywords through keyword matching and / or large model generation, and send the evaluation keywords to the client for display. In this embodiment, the server can obtain at least one dimension of evaluation keywords by matching various features such as product name, ingredients, and flavor in a keyword database, or it can generate at least one dimension of evaluation keywords through a large model, or it can use appropriate prompts to search for at least one dimension of evaluation keywords in a keyword database using a large model. This embodiment does not impose specific limitations. By generating at least one dimension of evaluation keywords, users can more easily select suitable keywords, thereby improving the efficiency of user evaluation keyword generation, helping users broaden their evaluation ideas, increasing the richness of product evaluation content, and making the generated product evaluation content more in line with user expectations.

[0170] In one embodiment, the training method for the large model in step 104 can be implemented through the following steps: the server obtains multiple historical user reviews corresponding to stores of multiple product categories; extracts review keywords from the historical user reviews and sets at least one dimension of review tags for the review keywords; and pre-trains the original large model based on the review keywords and the review tags corresponding to the review keywords to obtain the trained large model.

[0171] In the above embodiments, the server can train a general large model using historical user reviews from stores across multiple product categories. This allows the trained model to better understand product and user reviews, generating more authentic product reviews. Taking milk tea reviews as an example, the server can collect reviews from multiple milk tea products and extract various review keywords, such as "very delicious," "would buy again," "top-tier milk tea," "milk jelly," "sago," "freshly hand-cut seasonal mango," "strong tea flavor," "green tea base," "sweet even without sugar," "hot," and "smooth and chewy." Then, based on the characteristics of each review keyword, it can assign review tags for each keyword across various dimensions, such as "overall rating," "ingredients," "tea base," "sweetness," "temperature," "taste," and "service." Finally, the server trains the large model using these review keywords and their corresponding tags, enabling the model to output product reviews centered around this product category.

[0172] Furthermore, after training the large model with historical user reviews of milk tea products, when the server receives product information for "Green Tea Milk Tea with Milk Pudding and Sago," and user reviews for "Overall order satisfaction 5 / 5, packaging 5 / 5, Green Tea Milk Tea tastes amazing, portion size is amazing, and packaging quality is amazing," the large model can output the following product review: "I bought standard sweetness + Green Tea Milk Tea + Milk Pudding + Sago from this store. I'm extremely satisfied with their service! First, their packaging is really amazing; each item is individually packaged, which is very convenient. Second, the taste is amazing; the mango is freshly cut. Most importantly, the portion size is also amazing; I love it! In short, it's delicious and I recommend everyone try it!" In this embodiment, the above method, by structuring the historical user reviews of multiple product categories, facilitates the understanding of the specific meaning of product information and user review information by the large model. It can return descriptive phrases such as "delicious" based on review keywords of different dimensions, rather than simply restating product content and user review content. This significantly improves the quality of the generated product review content and enhances its authenticity and stability.

[0173] In one embodiment, the training method for the large model in step 104 can also be implemented through the following steps: the server obtains multiple historical user reviews in various language styles, wherein each historical user review is assigned a corresponding language style label; based on the historical review content and the corresponding language style labels, the original large model is pre-trained to obtain the trained large model.

[0174] In the above embodiments, the server can acquire user historical review content in multiple language styles and retrain a general or pre-trained large model to enable the trained large model to generate product review content in multiple language styles. For example, the server can acquire user historical review content in multiple language styles and set corresponding language style tags for each user historical review content. For example, it can set various language style tags such as "interesting style," "freehand style," "refined style," "professional style," and "friendly style" for different user historical review content. During training, 10-20 user historical review contents in the same language style can be provided to the large model as templates for product review content, allowing the large model to learn and imitate the text style, ultimately enabling the large model to output product review content in five different language styles.

[0175] For example, when the template for "freehand style" is "People come and go on the river, but they all love the beauty of the perch," the product review content output by the large model in the "freehand style" could be "Orange pumpkin soft European bread, conquering the taste buds with its domineering flavor," or "Dudu cherry blossom and white peach sweetness, the taste is absolutely amazing," etc. When the template for "refined style" is "I really like this restaurant's food; the taste is authentic, the texture is great, it's the taste I like, very satisfied, highly recommended!", the product review content output by the large model in the "refined style" could be "Delicious and diverse, generous portions, exquisite packaging, five-star rating!" etc. This method, by training the large model to generate product review content in multiple language styles, can improve the large model's understanding of various language styles, thus facilitating the accurate generation of product review content in multiple different language styles. This improves the richness and authenticity of product review content, thereby enhancing the user experience.

[0176] Furthermore, as a refinement and extension of the specific implementation methods of the above embodiments, and to fully illustrate the implementation process of this embodiment, a method for generating product evaluation content is provided, such as... Figure 2As shown, the method includes the following steps: After receiving the purchased goods, the user can initiate a product review request. During the product review process, if the user clicks the smart review entry on the product review page, they can enter the smart review mode and view the smart review page. At this time, the client can send product information and user rating information to the server, so that the server calls the large model to generate multiple first product review contents based on the product information and user rating information and provide feedback to the user. Furthermore, the user can also enter and / or select multiple review keywords as user review keywords on the smart review page. The client can send the user's entered and / or selected user review keywords to the server, so that the server again calls the large model to generate second product review contents based on the product information, user rating information, and user review keywords and provide feedback to the user. The user can edit and refine the first and / or second product review contents displayed on the client. When the user confirms a certain product review content, they can select that product review content. At this time, the selected product review content can be automatically filled into the product review page. In this way, the user only needs to click submit review to complete the entire product review process. By using the methods described above, we can effectively improve the efficiency of product reviews, reduce the workload of users in reviewing products, improve the quality of product review content, and enhance the user experience.

[0177] In the above embodiments, the interaction method between the client and the server can be adjusted according to the actual form of the product. For example, Figures 3 to 10 Three interaction methods were demonstrated. Among them, in Figure 3 and Figure 4 In the illustrated scenario, the client can display multiple product reviews, and the user can select any one to populate the product review page. Alternatively, they can click the "Change" control to view and select from more product reviews. Figure 5 , Figure 6 and Figure 7 In the illustrated scenario, the client can display multiple first-order product reviews, and can also display second-order product reviews based on user-inputted or selected review keywords. Similarly, the user can select any product review to populate the product review page, or click the "Change" control to view and select from more product reviews. Figure 8 , Figure 9 and Figure 10In the illustrated scenario, the client can display second product reviews based on user-inputted or selected review keywords. Similarly, users can select any product review to fill in the product review page, or click the "Change" control to view more product reviews for selection. It is understood that the above three product forms and interaction methods are merely examples and not intended to limit the specific implementation of this embodiment. In practical applications, the required product form and interaction method can be selected according to the actual needs of the product to implement the methods described in any of the above embodiments.

[0178] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the labels corresponding to each step in the above embodiments are only for identification purposes and are not intended to limit the order of execution of the steps. The order of execution of the steps in each embodiment can be set according to the actual situation.

[0179] Furthermore, as Figures 1 to 10 The specific implementation of the method shown in this embodiment provides a product review content generation device, such as... Figure 11 As shown, the device includes: an evaluation entry display module 71, an evaluation information acquisition module 72, and an evaluation content display module 73, wherein:

[0180] The evaluation entry display module 71 can be used to respond to a product evaluation request, display a product evaluation page, and display a smart evaluation entry on the product evaluation page;

[0181] The evaluation information acquisition module 72 can be used to respond to the trigger operation of the intelligent evaluation entry, acquire the product information and user evaluation information of the product to be evaluated, and send the product information and the user evaluation information, wherein the user evaluation information includes user rating information and / or user evaluation keywords;

[0182] The evaluation content display module 73 can be used to receive and display the product evaluation content, wherein the product evaluation content is generated based on the product information and the user evaluation information through a large model.

[0183] In specific application scenarios, each module in the product review content generation device can be used to execute the product review content generation method described in any of the above embodiments. Based on this, the specific functions, implementation processes, and technical effects of each module in the product review content generation device will not be elaborated here.

[0184] Furthermore, as Figures 1 to 10The specific implementation of the method shown in this embodiment provides a product review content generation device, such as... Figure 12 As shown, the device includes: an evaluation content generation module 81 and an evaluation content sending module 82, wherein:

[0185] The evaluation content generation module 81 can be used to respond to receiving product information and user evaluation information of the product to be evaluated, and generate product evaluation content based on the product information and the user evaluation information through a large model. The product information and the user evaluation information are received after the intelligent evaluation entry on the product evaluation page of the client is triggered. The user evaluation information includes user rating information and / or user evaluation keywords.

[0186] The evaluation content sending module 82 can be used to send the product evaluation content to the client so that the client can receive and display the product evaluation content.

[0187] In specific application scenarios, each module in the product review content generation device can be used to execute the product review content generation method described in any of the above embodiments. Based on this, the specific functions, implementation processes, and technical effects of each module in the product review content generation device will not be elaborated here.

[0188] Based on the above, Figures 1 to 10 Accordingly, this embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the above-described method. Figures 1 to 10 The method for generating product review content is shown.

[0189] Based on this understanding, the technical solution of this application can be embodied in the form of a software product. The software product to be identified can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, or portable hard drive), including several instructions to cause a computer device (such as a personal computer, server, or network device) to execute the methods described in the various implementation scenarios of this application.

[0190] Based on the above, Figures 1 to 10 The method shown, and Figure 11 and Figure 12 The illustrated embodiment of the product review content generation device, in order to achieve the above objectives, such as... Figure 13 As shown, this embodiment also provides a computer device for generating product review content, which can be a personal computer, server, smartphone, tablet computer, smartwatch, or other network device, etc. The computer device includes a storage medium and a processor; the storage medium is used to store computer programs and an operating system; the processor is used to execute the computer program to achieve the above-mentioned... Figures 1 to 10 The method shown.

[0191] Optionally, the computer device may also include internal memory, a communication interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, a display screen, and input devices such as a keyboard. The communication interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0192] Those skilled in the art will understand that the computer device structure for recognizing operational actions provided in this embodiment does not constitute a limitation on the computer device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0193] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the aforementioned computer hardware and the software resources to be identified, supporting the operation of information processing programs and other software and / or programs to be identified. The network communication module is used to enable communication between the various components within the storage medium, as well as communication with other hardware and software in the information processing computer device.

[0194] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented using software plus necessary general-purpose hardware platforms, or it can be implemented through hardware. The application client responds to product review requests. Compared with existing technologies, the above method can improve the efficiency and quality of product review content generation, reduce the workload and time cost for users to write reviews, help increase the proportion of product reviews, improve the user experience of writing reviews, and motivate users to continuously output high-quality review content.

[0195] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing this application. Those skilled in the art will understand that the modules in the apparatus of the embodiment can be distributed within the apparatus of the embodiment as described, or can be modified to be located in one or more apparatuses different from this embodiment. The modules of the above-described embodiment can be combined into one module, or further divided into multiple sub-modules.

[0196] The serial numbers in this application are for descriptive purposes only and do not represent the superiority or inferiority of any particular implementation scenario. The above disclosures are merely a few specific implementation scenarios of this application; however, this application is not limited thereto, and any variations conceived by those skilled in the art should fall within the protection scope of this application.

Claims

1. A method for generating product review content, characterized in that, The method includes: The client responds to the product review request by displaying the product review page, and the smart review entry is displayed on the product review page; The client responds to the trigger operation of the intelligent evaluation entry to obtain product information and user evaluation information of the product to be evaluated, wherein the user evaluation information includes user rating information and / or user evaluation keywords; The client sends the product information and the user review information to the server; The server responds to receiving product information and user review information of the product to be evaluated, generates product review content based on the product information and user review information through a large model, and sends the product review content to the client; The client receives and displays the product review content.

2. The method according to claim 1, characterized in that, The client displays a smart review entry on the product review page, including: The client displays the smart review entry at at least one preset location on the product review page through at least one preset component, wherein the preset component is displayed in a static and / or dynamic manner, and the preset component consists of at least one element among text, image, and icon.

3. The method according to claim 1, characterized in that, The client displays a smart review entry on the product review page, including: In response to input of evaluation content, the client displays a keyboard component and the smart evaluation entry is displayed at a preset position within a preset range of the keyboard component. The smart evaluation entry is removed when the keyboard component is hidden.

4. The method according to claim 1, characterized in that, The user rating information includes first rating information and / or second rating information; then the client obtains the user rating information, including: The client obtains the user's first rating information for the order to be evaluated and / or the user's second rating information for at least one item in the order to be evaluated, wherein the user rating information includes at least one of the following: rating value information, rating component operation information, like component operation information, and dislike component operation information.

5. A method for generating product review content, characterized in that, The method includes: In response to a product review request, the product review page is displayed, and the smart review entry is shown on the product review page; In response to the triggering operation of the intelligent evaluation entry, the product information and user evaluation information of the product to be evaluated are obtained, and the product information and user evaluation information are sent, wherein the user evaluation information includes user rating information and / or user evaluation keywords; Receive and display the product review content, wherein the product review content is generated based on the product information and the user review information through a large model.

6. A method for generating product review content, characterized in that, The method includes: In response to receiving product information and user review information of the product to be reviewed, product review content is generated through a large model based on the product information and the user review information. The product information and the user review information are received after being triggered by the smart review entry on the product review page of the client. The user review information includes user rating information and / or user review keywords. The product review content is sent to the client so that the client can receive and display the product review content.

7. A product review content generation device, characterized in that, The device includes: The evaluation entry display module is used to respond to product evaluation requests, display the product evaluation page, and display the intelligent evaluation entry on the product evaluation page; The evaluation information acquisition module is used to respond to the trigger operation of the intelligent evaluation entry, acquire the product information and user evaluation information of the product to be evaluated, and send the product information and the user evaluation information, wherein the user evaluation information includes user rating information and / or user evaluation keywords; The evaluation content display module is used to receive and display the product evaluation content, wherein the product evaluation content is generated based on the product information and the user evaluation information through a large model.

8. A product review content generation device, characterized in that, The device includes: The evaluation content generation module is used to respond to receiving product information and user evaluation information of the product to be evaluated, and to generate product evaluation content based on the product information and the user evaluation information through a large model. The product information and the user evaluation information are received after being triggered by the smart evaluation entry on the product evaluation page of the client. The user evaluation information includes user rating information and / or user evaluation keywords. The evaluation content sending module is used to send the product evaluation content to the client so that the client can receive and display the product evaluation content.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images