Information processing method and device

Abstract

The embodiments of this specification provide an information processing method and device. A specific implementation of the information processing method includes: receiving comment information input by a user; analyzing the input comment information to determine prompt information to be provided to the user, the prompt information is used to prompt the user to adjust the comment information; and displaying the prompt information to the user.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of computer technology, and in particular, to information processing methods and devices. Background Art

[0002] With the rapid development of computer technology, users can now publish various comment information on the Internet, such as product evaluation information, reasons for product returns, public opinion information, article comment information, service content feedback information, etc. All of the above comment information can be used as user behavior data for statistical analysis by relevant institutions and personnel to portray user portraits, or to determine the highlights and / or problems of the comment object, etc.

[0003] At present, some user-submitted reviews are unconsciously described, and some information elements are usually missing, which makes the review information incomplete and cannot clearly reflect the user's true intention. When the review information with missing information is used for the above statistical analysis, less accurate analysis results are usually obtained. In addition, in real life, when the return reason with missing information is fed back to the seller, the seller generally cannot fully understand the buyer's return reason, which will cause communication between the buyer and the seller to be unsmooth.

[0004] Therefore, a feasible solution is needed to make it possible to obtain more comprehensive review information. Summary of the invention

[0005] The embodiments of this specification provide information processing methods and devices.

[0006] In a first aspect, an embodiment of the present specification provides an information processing method, the method comprising: receiving comment information input by a user; analyzing the input comment information to determine prompt information that needs to be provided to the user, the prompt information being used to prompt the user to adjust the comment information; and displaying the prompt information to the user.

[0007] In some embodiments, the method further includes: receiving supplementary comments provided by the user in response to the input comment information; generating comprehensive comment information based on the input comment information and the supplementary comments; and displaying the comprehensive comment information to the user.

[0008] In some embodiments, the comment object pointed to by the input comment information has a corresponding associated person; the executor of the method is the server; and after generating comprehensive comment information based on the input comment information and the supplementary comments, the method also includes: sending the comprehensive comment information to the client used by the associated person, so that the client displays the comprehensive comment information.

[0009] In some embodiments, the user is a buyer and the associated person is a seller; or, the user is a user who publishes public opinion information and the associated person is an administrator who manages public opinion information; or, the user is a user who makes comments on service content and the associated person is an administrator of feedback information on the service content.

[0010] In some embodiments, the analyzing the input comment information to determine the prompt information that needs to be provided to the user includes: analyzing the input comment information to obtain an analysis result, wherein the analysis result shows whether the input comment information is missing predetermined information elements; in response to the analysis result showing that the input comment information is missing predetermined information elements, determining the prompt information that needs to be provided to the user.

[0011] In some embodiments, the predetermined information element includes at least one of the following element items: attribute information of the review object, and description information of the attribute information of the review object.

[0012] In some embodiments, when the missing element item of the input comment information is the attribute information of the comment object, the prompt information is used to guide the user to supplement the attribute information of the comment object; when the missing element item of the input comment information is the description information of the attribute information of the comment object, the prompt information is used to guide the user to supplement the description information of the attribute information of the comment object.

[0013] In some embodiments, the review object is any one of the following: a product object, a public opinion object, or a service object.

[0014] In some embodiments, the receiving of supplementary comments provided by the user in response to the input comment information includes: receiving the supplementary comments input by the user in response to the prompt information; determining whether the number of prompts reaches a preset number, wherein the number of prompts is the number of times the prompt information is displayed to the user; if the number of prompts does not reach the preset number, analyzing the supplementary comments to obtain a supplementary comment analysis result; if the supplementary comment analysis result shows that the comment needs to be supplemented, determining the prompt information that needs to be supplemented to the comment; displaying the determined prompt information to the user so that the user can supplement the supplementary comments based on the prompt information; receiving the supplementary comments input by the user in response to the prompt information corresponding to the supplementary comments.

[0015] In some embodiments, the method further includes: if it is determined that the number of prompts reaches the preset number, or the supplementary comment analysis result shows that there is no need to supplement the comment, then executing the input comment information and the supplementary comment to generate comprehensive comment information.

[0016] In some embodiments, the input comment information is analyzed to obtain an analysis result, including: selecting comment information that matches the input comment information from a pre-collected set of comment information related to the comment object pointed to by the input comment information; determining whether the input comment information is missing the predetermined information element based on the part of speech of each word in the input comment information and the selected comment information; and generating an analysis result based on the determination result.

[0017] In some embodiments, the comment information in the comment information set is set with a corresponding click-through rate value; and selecting comment information that matches the input comment information from a pre-collected comment information set related to the comment object pointed to by the input comment information comprises: selecting comment information from the comment information set based on the click-through rate value corresponding to the comment information in the comment information set; and determining the selected comment information as comment information that matches the input comment information.

[0018] In some embodiments, generating an analysis result based on the determination result includes: when the determination result shows that the input comment information is not missing the predetermined information element, generating an analysis result including any one of the following items: the input comment information, a preset flag used to characterize that the predetermined information element is not missing.

[0019] In some embodiments, generating an analysis result based on the determination result includes: when the determination result shows that the input comment information is missing the predetermined information element, generating an analysis result including an information group, the information group including: the position of the missing element item in the input comment information, and a question phrase corresponding to the position of the missing element item; wherein the question phrase is a phrase used to ask questions about the missing element item at the position of the missing element item.

[0020] In some embodiments, the information group also includes at least one of the following: the position of the missing element item at the location of the missing element item in the comment information matching the input comment information, and the word at the position in the comment information matching the input comment information.

[0021] In some embodiments, the analysis results also show, in the case where a predetermined information element is missing, an information group related to the missing element item; and the determination of the prompt information that needs to be provided to the user includes: determining the prompt information that needs to be provided to the user based on the input comment information and the information group.

[0022] In some embodiments, determining the prompt information that needs to be provided to the user based on the input comment information and the information group includes: inputting the input comment information and the information group into a pre-trained prompt information generation model to obtain the prompt information, wherein the prompt information generation model is obtained by training a pointer network.

[0023] In some embodiments, generating comprehensive comment information based on the input comment information and the supplementary comments includes: inputting the input comment information and the supplementary comments into a pre-trained comment information generation model to obtain comprehensive comment information, wherein the comment information generation model is obtained by training a pointer network.

[0024] In some embodiments, the method is performed by a client; and after generating comprehensive comment information based on the input comment information and the supplementary comment, the method further includes: sending the comprehensive comment information to a server.

[0025] In some embodiments, the executor of the method is a server; and the receiving comment information input by the user includes: receiving comment information input by the user from the client; and the displaying the prompt information to the user includes: sending the prompt information to the client, so that the client displays the prompt information to the user.

[0026] In a second aspect, an embodiment of the present specification provides an information processing device, which includes: a receiving unit, configured to receive comment information input by a user; an analysis and determination unit, configured to analyze the input comment information and determine prompt information that needs to be provided to the user, wherein the prompt information is used to prompt the user to adjust the comment information; and an output unit, configured to display the prompt information to the user.

[0027] In some embodiments, the device also includes a generating unit; the receiving unit is further configured to receive supplementary comments provided by the user regarding the input comment information; the generating unit is configured to generate comprehensive comment information based on the input comment information and the supplementary comments; and the output unit is further configured to display the comprehensive comment information to the user.

[0028] In a third aspect, an embodiment of the present specification provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed in a computer, the computer is caused to execute the method described in any implementation manner in the first aspect.

[0029] In a fourth aspect, an embodiment of the present specification provides a computing device, comprising a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method described in any implementation manner in the first aspect is implemented.

[0030] The information processing method and device provided by the above-mentioned embodiments of the present specification receive the comment information input by the user, and then analyze the input comment information to determine the prompt information to be provided to the user, the prompt information is used to prompt the user to adjust the comment information, and then display the prompt information to the user so that the user can adjust the input comment information based on the prompt information. In this way, intelligent guidance of the user can be achieved so as to obtain more comprehensive comment information. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the drawings required for use in the embodiments or the prior art description will be briefly introduced below. Obviously, the drawings described below are only some examples or embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without creative work, and this specification can also be applied to other similar scenarios based on the provided drawings. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0032] Figure 1a is an exemplary system architecture diagram to which some embodiments of the present specification may be applied;

[0033] Figure 1b is another exemplary system architecture diagram to which some embodiments of the present specification may be applied;

[0034] Figure 2 is a flow chart of an embodiment of an information processing method according to the present specification;

[0035] Figure 3a is a schematic diagram of an application scenario of the information processing method according to this specification;

[0036] Figure 3b This is a schematic diagram of the interface for posting a follow-up review;

[0037] Figure 3c It is a schematic diagram of the display effect of the prompt information;

[0038] Figure 4 is a flowchart of another embodiment of the information processing method according to the present specification;

[0039] Figure 5is a flow chart of a process for obtaining information of supplementary comments for comment information input by a user;

[0040] Figure 6a is another schematic diagram of an application scenario of the information processing method according to this specification;

[0041] Figure 6b This is another schematic diagram of the interface for posting a follow-up review;

[0042] Figure 6c It is another schematic diagram of the display effect of the prompt information;

[0043] Figure 6d This is another schematic diagram of the interface for posting a follow-up review;

[0044] Figure 6e It is a schematic diagram of the display effect of comprehensive evaluation information;

[0045] Figure 7 is another schematic diagram of an application scenario of the information processing method according to the present specification;

[0046] Figure 8 It is a schematic diagram of the structure of an information processing device according to this specification. DETAILED DESCRIPTION

[0047] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. The described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of this application.

[0048] It should be noted that, for the convenience of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments in this specification can be combined with each other.

[0049] As mentioned above, some user-submitted reviews are unconsciously described and usually lack some information elements, which makes the review information incomplete and unable to clearly reflect the user's true intention. When the review information with missing information is used for the above statistical analysis, less accurate analysis results are usually obtained. In addition, in real life, when the return reason with missing information is fed back to the seller, the seller generally cannot fully understand the buyer's return reason, which will cause communication between the buyer and the seller to be unsmooth.

[0050] Based on this, some embodiments of this specification disclose information processing methods. Specifically, Figure 1a , Figure 1b Exemplary system architecture diagrams applicable to these embodiments are respectively shown.

[0051] like Figure 1a As shown, it shows the client 101 and the user User1. Among them, User1 is the user to which the client 101 belongs.

[0052] In practice, the client 101 has a comment information publishing function, and can allow users to publish comment information. The client 101 can be embodied as a client application, a software module in the client application, or a terminal device, etc. When the client 101 is embodied as a client application, the client application can be various applications with a comment information publishing function, including but not limited to shopping applications, microblog applications, blog applications, audio and video applications, game applications, teaching applications, public opinion applications, etc. When the client 101 is embodied as a terminal device, at least one of the applications listed above can be installed on the terminal device. In addition, the terminal device can be various electronic devices, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, etc.

[0053] Specifically, the client 101 can receive the comment information input by User1, and when the comment information needs to be adjusted, obtain prompt information for prompting the user to adjust the comment information, and display the prompt information to User1, so that User1 can adjust the input comment information based on the prompt information. The client 101 can obtain the prompt information locally or remotely.

[0054] like Figure 1b As shown, it displays the client 101 and the user User1, and also displays the server 102. The server 102 is the background server of the client 101. The server 102 can be embodied as a cloud platform, a single server or a server cluster.

[0055] When the client 101 remotely obtains the prompt information, the client 101 may send the comment information input by User1 to the server 102 for analysis by the server 102. When the server 102 analyzes that the comment information needs to be adjusted, the server 102 may determine the prompt information for prompting the user to adjust the comment information, and return the prompt information to the client 101. The client 101 may display the prompt information to User1.

[0056] The specific implementation steps of the above method are described below in conjunction with specific embodiments.

[0057] See also Figure 2, which shows a process 200 of an embodiment of an information processing method. The execution subject of the information processing method can be a client (such as Figure 1a The client 101 in FIG. 101 may also be a server (such as Figure 1b The method comprises the following steps:

[0058] Step 201, receiving comment information input by a user;

[0059] Step 202, analyzing the comment information input by the user to determine prompt information that needs to be provided to the user, the prompt information is used to prompt the user to adjust the comment information;

[0060] Step 203: Display the prompt information to the user.

[0061] The steps 201-203 are described in detail below.

[0062] In step 201, when the execution subject is a client, the execution subject may directly receive the comment information input by the user. When the execution subject is a server, the execution subject may receive the comment information input by the user from the client.

[0063] The comment information may be in the form of audio or text. When the comment information input by the user is not text information, an existing conversion algorithm may be used to convert the comment information input by the user into text information. For example, when the comment information input by the user is audio information, an existing audio recognition algorithm may be used to convert the audio information into corresponding text information.

[0064] In practice, review information can be comments on purchased products in e-commerce scenarios, or reviews in other scenarios, such as movies. Furthermore, in addition to traditional evaluation information, the review can also be information submitted by users in after-sales scenarios, such as when requesting a return, the system prompts the user to complete the information.

[0065] The user may be, for example, a buyer, an article reviewer, an audio and video reviewer, a user who publishes public opinion information, or a user who makes comments on the service content, etc. The comment information input by the user may be comment information on the comment object. The comment object may be, for example, a product object, a public opinion object, a service object, an audio and video object, or an article object, etc.

[0066] Among them, commodity objects can be physical commodities or virtual commodities. Physical commodities can be various tangible commodities, such as but not limited to clothing commodities, shoes and hats commodities, electronic commodities, maternal and child commodities, food commodities, learning tools commodities, construction commodities, etc. Virtual commodities can be various intangible commodities, including but not limited to computer software, online games, virtual cloud disks, music and images, e-books, etc. The public opinion objects can be, for example, events, social managers, enterprises, individuals and other organizations. Service objects can be various types of services, including but not limited to software services, housekeeping services, educational services, etc. Article objects can be various articles on the Internet. Audio and video objects can include various types of audio, video, and a combination of audio and video.

[0067] Specifically, when the review object is a product object, the review information input by the user may be evaluation information for the product or a reason for return, etc. When the review object is a service object, the review information input by the user may be feedback information on the service content, etc.

[0068] In step 202, the main purpose of analyzing the comment information input by the user is to determine whether the comment information is complete enough and whether it can reflect the real intention of the user.

[0069] In practice, since some users submit comments unconsciously, some information elements are usually missing, which makes the comments incomplete and cannot clearly reflect the user's true intention. Therefore, it is necessary to analyze the comments after receiving them.

[0070] In this embodiment, the execution entity may use various analysis methods to analyze the comment information input by the user.

[0071] For example, the execution entity may input the comment information input by the user into a pre-trained text analysis model to obtain the analysis result output by the text analysis model. The text analysis model may be trained based on any machine learning model for text analysis. The analysis result may indicate whether the comment information needs to be adjusted. Adjusting the comment information includes supplementing the comment information.

[0072] Optionally, the comment information input by the user can be analyzed based on predetermined information elements. The predetermined information elements are elements that can reflect the user's intention. When it is analyzed that the comment information input by the user lacks predetermined information elements, it can be determined that the comment information needs to be adjusted. In practice, when the comment information input by the user is comment information for the comment object, the predetermined information elements may include at least one of the following element items: attribute information of the comment object, and descriptive information for the attribute information of the comment object.

[0073] As an exemplary analysis method, the execution subject may pre-collect the attribute information of each attribute of the review object and the evaluation phrase set related to each attribute. The execution subject may first determine whether the review information input by the user includes the attribute information of the review object based on the stored attribute information of the review object. If not, it may be determined that the review information input by the user lacks the attribute information of the review object and the description information of the attribute information of the review object. If it is determined that the review information input by the user includes the attribute information of the review object, the phrase following the attribute information in the review information input by the user may be further matched with the evaluation phrase in the evaluation phrase set. If there is an evaluation phrase matching the phrase following the attribute information in the evaluation phrase set, it may be determined that the review information input by the user does not lack the predetermined information element. If there is no evaluation phrase matching the phrase following the attribute information in the evaluation phrase set, it may be determined that the review information input by the user lacks the description information of the attribute information of the review object.

[0074] In addition, after determining that the review information input by the user lacks predetermined information elements, the above-mentioned execution subject may also determine the question phrase corresponding to the element item based on the element item missing from the review information input by the user. The question phrase is a phrase for asking questions about the element item. For example, the above-mentioned execution subject may pre-store question phrases corresponding to each element item in the predetermined information elements. One element item may correspond to one or more question phrases. For example, a descriptive information element item for the attribute information of the review object may correspond to multiple question phrases. The multiple question phrases include question phrases related to positive reviews and question phrases related to negative reviews. The above-mentioned execution subject may determine the question phrase corresponding to the element item missing from the input review information based on the correspondence between the element item and the question phrase.

[0075] After the above-mentioned execution subject performs the above analysis on the comment information input by the user, if it is determined that the comment information input by the user does not lack the predetermined information elements, the above-mentioned execution subject may generate an analysis result including any of the following items: the comment information input by the user, and a preset mark used to characterize that the predetermined information elements are not missing. The preset mark can be a number, a letter, or a combination of the two, which is not specifically limited here. In addition, if it is determined that the comment information input by the user lacks the predetermined information elements, the above-mentioned execution subject may generate an analysis result including an information group. The information group may, for example, include the position of the missing element item in the comment information input by the user and the question phrase corresponding to the position of the missing element item.

[0076] Taking the evaluation business as an example, suppose the review object is clothes, and the review information input by the user is "the clothes are not good". After the above-mentioned execution subject adopts the above-mentioned analysis method, it can be determined that the review information input by the user lacks the attribute information of the clothes between "clothes" and "not good" and the specific problem description for the missing attribute information. Because when asking questions about the attribute information of the review object, the user may supplement the attribute information of the review object while supplementing the specific problem description for the attribute information, therefore, the question can be asked about the attribute information of the review object first. In this way, the above-mentioned execution subject can obtain the question phrase corresponding to the attribute information of the clothes, such as "where". The information group in the analysis result generated by the above-mentioned execution subject using the above-mentioned analysis method may include the location of the missing attribute information of the clothes in "the clothes are not good" and the question phrase "where".

[0077] Continuing with the evaluation business as an example, suppose the review object is clothing, and the review information input by the user is "the material of the clothing is not good." After the above-mentioned execution subject adopts the above-mentioned analysis method, it can be determined that the review information input by the user includes the attribute "material" of the clothing, and there is a lack of a specific problem description for the material between the attribute "material" and the evaluation word "not good." Since "not good" is an evaluation word used to characterize a bad review, the above-mentioned execution subject can obtain a question phrase corresponding to the material and related to the bad review, such as "what specific problem has occurred." Therefore, the information group in the analysis result generated by the above-mentioned analysis method may include the position where the specific problem description for the material is missing in "the material of the clothing is not good" and the question phrase "what specific problem has occurred."

[0078] Still taking the evaluation business as an example, suppose the review object is clothes, and the review information input by the user is "the clothes are made of good materials." After the above-mentioned execution subject adopts the above-mentioned analysis method, it can be determined that the review information input by the user includes the attribute "material" of the clothes, and there is a lack of specific highlights description of the materials between the attribute "material" and the evaluation word "good." Since "good" is an evaluation word used to characterize good reviews, the above-mentioned execution subject can obtain question phrases corresponding to the materials and related to the good reviews, such as "what is good about it specifically?" Therefore, the information group in the analysis result generated by the above-mentioned analysis method may include the position of the missing specific highlights description of the materials in "the clothes are made of good materials" and the question phrase "what is good about it specifically?"

[0079] Taking the public opinion business as an example, suppose the comment object is Company A, and the comment information input by the user is "The executive surnamed Li of Company A is not good." After the above-mentioned execution subject adopts the above-mentioned analysis method, it can be determined that the comment information input by the user includes the attribute "Executive surnamed Li" of Company A, and there is a lack of a specific problem description for Executive surnamed Li between the attribute "Executive surnamed Li" and the evaluation word "not good." Since "not good" is an evaluation word used to characterize a bad review, the above-mentioned execution subject can obtain question phrases corresponding to the executive and related to the bad review, such as "Where is the performance not good?" Therefore, the information group in the analysis result generated by the above-mentioned analysis method may include the position of the missing specific problem description for Executive surnamed Li in "Executive surnamed Li of Company A is not good" and the question phrase "Where is the performance not good?"

[0080] In step 202, in response to the analysis result indicating that the comment information needs to be adjusted, prompt information to be provided to the user may be determined. As an implementation, the prompt information may be a preset general prompt information, such as "Dear, please supplement the comment."

[0081] As another implementation, in order to allow users to supplement their comments in a targeted manner, the prompt information to be provided to the user can be determined based on the comment information input by the user and the information group described above in the analysis result. When the missing element item of the comment information input by the user is the attribute information of the comment object, the prompt information can be used to guide the user to supplement the attribute information of the comment object. When the missing element item of the comment information input by the user is the description information of the attribute information of the comment object, the prompt information can be used to guide the user to supplement the description information of the attribute information of the comment object.

[0082] As an example, assuming that the comment information input by the user is "the clothes are not good", the information group includes the position of the missing attribute information of the clothes in "the clothes are not good" and the question phrase "where", and the position is used to point between "clothes" and "not good". The above execution subject can generate prompt information such as "where is the clothes not good?"

[0083] In order to reduce the user's aversion to the question asked and encourage the user to actively supplement the information, a more friendly prompt message can be generated, such as "Excuse me, what do you think is wrong with the clothes?" Here, the opening phrase used to reflect the intimacy, such as "Excuse me, what do you think" can be preset or randomly generated according to a specific algorithm, and this specification does not make any specific restrictions on this.

[0084] In addition, in order to emphasize that the prompt information is a question that needs to be answered by the user, a mark used to represent the question can be added at the beginning of the prompt information, such as "Q:" or "Question:". Among them, "Q" can be understood as the English abbreviation of "question". At this time, the prompt information generated for the comment information "The clothes are not good" input by the user can be, for example, "Q: What do you think is not good about the clothes?" The prompt information generated for the comment information "The executive surnamed Li of Company A is not good" input by the user can be, for example, "Q: What do you think is not good about the performance of the executive surnamed Li of Company A?"

[0085] In step 203, the execution entity may display prompt information to the user so that the user can adjust the comment information according to the prompt information. In practice, the user may adjust the comment information by submitting additional comments. Alternatively, the user may adjust the comment information by modifying the comment information.

[0086] It should be noted that when the execution subject is a client, the execution subject may directly display the determined prompt information to the user. When the execution subject is a server, the execution subject may return the determined prompt information to the client used by the user so that the client displays the prompt information to the user.

[0087] Continue to see Figure 3a , Figure 3a is a schematic diagram of an application scenario of the information processing method according to this embodiment. This application scenario shows content related to the evaluation service. In this application scenario, the comment information is called evaluation information. The analysis of the evaluation information and the determination of the prompt information are both performed on the server side.

[0088] Specifically, in this application scenario, the client is a shopping application, and the user to which the client belongs is a buyer who has purchased clothes through the shopping application. When the buyer is not satisfied with the clothes he has bought, he can input evaluation information about the clothes through the client, as shown in reference numeral 301. The evaluation information can be, for example, "the clothes are not good." Specifically, the buyer can, for example, Figure 3b Submit the evaluation information on the posting review interface shown. Figure 3bA schematic diagram of the interface for posting a follow-up review is shown. After receiving the evaluation information, the client can send the evaluation information to the server as shown by label 302. After receiving the evaluation information, the server can analyze the evaluation information using a predetermined analysis algorithm as shown by label 303 to obtain an analysis result, which shows that the evaluation information is missing attribute information of the clothes, as well as the location of the missing attribute information of the clothes in the evaluation information and the question phrase "where". Then, the server can generate a prompt message "Q: What do you think is wrong with the clothes?" based on the evaluation information and the location and question phrase in the analysis result as shown by label 304. Then, the server can send the prompt message to the client as shown by label 305. The client can display the prompt message to the buyer as shown by label 306, so that the buyer can supplement the attribute information of the clothes based on the evaluation information. Here, the display effect of the prompt message can be as shown in the following example. Figure 3c As shown. Among them, Figure 3c A schematic diagram showing the display effect of prompt information. Figure 3c On the interface shown, the user can input supplementary evaluation according to the prompt information to adjust the evaluation information.

[0089] The information processing method provided by the above-mentioned embodiment of the present specification receives the comment information input by the user, and then analyzes the input comment information to determine the prompt information to be provided to the user, the prompt information is used to prompt the user to adjust the comment information, and then displays the prompt information to the user so that the user can adjust the input comment information based on the prompt information. In this way, intelligent guidance of the user can be achieved so as to obtain more comprehensive comment information.

[0090] Further references Figure 4 , which shows a process 400 of another embodiment of the information processing method. The execution subject of the information processing method can be a client (such as Figure 1a The client 101 in FIG. 101 may also be a server (such as Figure 1b The method comprises the following steps:

[0091] Step 401, receiving comment information input by a user;

[0092] Step 402, analyzing the comment information input by the user to obtain an analysis result, wherein the analysis result shows whether the comment information is missing a predetermined information element, and if the predetermined information element is missing, an information group related to the missing element item;

[0093] Step 403, in response to the analysis result indicating that the review information is missing a predetermined information element, based on the review information input by the user and the information group in the analysis result, determining prompt information to be provided to the user, the prompt information being used to prompt the user to adjust the review information;

[0094] Step 404, displaying prompt information to the user;

[0095] Step 405, receiving a supplementary comment provided by the user in response to the input comment information;

[0096] Step 406, generating comprehensive comment information based on the comment information and supplementary comments input by the user;

[0097] Step 407, displaying the comprehensive review information to the user.

[0098] In this embodiment, for explanations of steps 401-404, please refer to Figure 2 The relevant descriptions in the corresponding embodiments are not repeated here.

[0099] In this embodiment, as an implementation, after the execution subject receives the supplementary comment input by the user in response to the prompt information displayed for the first time, in order to avoid disturbing the user again, regardless of whether the supplementary comment lacks the predetermined information element, the execution subject may execute step 406. In this case, step 405 may specifically include: receiving the supplementary comment input by the user in response to the prompt information displayed for the first time.

[0100] In some optional implementations of this embodiment, in order to obtain as comprehensive comment information as possible and avoid excessively disturbing the user, a prompting threshold may be set, and the prompting threshold may be 2 or 3. Preferably, the prompting threshold is set to 2, so that the user is prompted twice at most, which can avoid excessively disturbing the user. Therefore, in step 405, the execution subject may execute the following Figure 5 The information acquisition process shown in FIG. Figure 5 The information acquisition process 500 of the supplementary comments for the comment information input by the user is shown. It should be noted that, hereinafter, the prompt times threshold is referred to as the preset times.

[0101] like Figure 5 As shown, the information acquisition process 500 includes the following steps:

[0102] Step 501, receiving a supplementary comment input by a user in response to a prompt message;

[0103] Step 502, determining whether the number of prompts reaches a preset number, wherein the number of prompts is the number of times prompt information is displayed to the user;

[0104] Step 503, if the number of prompts does not reach the preset number, the supplementary comments are analyzed to obtain the supplementary comments analysis results;

[0105] Step 504, if the supplementary comment analysis result shows that the comment needs to be supplemented, then determine prompt information for supplementing the comment;

[0106] Step 505, displaying the determined prompt information to the user so that the user can supplement the supplementary comment based on the prompt information;

[0107] Step 506: receiving a supplementary comment input by the user in response to the prompt information corresponding to the supplementary comment.

[0108] It should be noted that the analysis method used for supplementary comments and the method for determining prompt information corresponding to supplementary comments are similar to the analysis method used for comment information input by the user and the method for determining prompt information corresponding to the comment information, and will not be repeated here.

[0109] In step 503, the supplementary comment analysis result specifically shows whether the supplementary comment input by the user lacks a predetermined information element. If the supplementary comment analysis result shows that the supplementary comment input by the user lacks a predetermined information element, it can be indicated that the comment needs to be supplemented. If the supplementary comment analysis result shows that the supplementary comment input by the user does not lack a predetermined information element, it can be indicated that the comment does not need to be supplemented.

[0110] Optionally, if it is determined in step 502 that the number of prompts reaches a preset number, or if it is determined in step 503 that the supplementary comment analysis result shows that there is no need to supplement the comment, the execution body may switch to executing step 406.

[0111] Optionally, after executing step 506 , the execution subject may proceed to execute step 502 .

[0112] In step 406, the execution entity may generate comprehensive comment information based on the comment information input by the user and all supplementary comments provided for the comment information input by the user. When the supplementary comments contain information elements missing from the comment information input by the user, the comprehensive comment information generated by the execution entity is more complete than the comment information input by the user and can better reflect the user's true intention.

[0113] For example, the comment information input by the user is "the material of the clothes is not good", and the supplementary comment is "there are burrs on the edges of the materials", and the comprehensive comment information generated by the above execution entity based on the comment information input by the user and the supplementary comment may be "there are burrs on the edges of the clothes".

[0114] It should be noted that the above-mentioned execution subject can use a pre-trained comment information generation model to generate comprehensive comment information based on the comment information input by the user and all supplementary comments provided for the comment information input by the user. The comment information generation model can be obtained by training a model for text generation such as a Seq2Seq model or a Pointer Network.

[0115] The Seq2Seq model is the abbreviation of the Sequence to Sequence model, also known as an encoder-decoder model, which can convert one sequence into another sequence. However, since the prediction output target size of the Seq2Seq model is fixed, problems such as unregistered words and repeated generated words may occur. The English name of unregistered words can be Out-of-Vocabulary, abbreviated as OOV.

[0116] Pointer Network can solve the above problems of Seq2Seq model. In PointerNetwork, it not only inherits the attention mechanism of Seq2Seq model, but also first performs weighted sum on the hidden state of Encoder layer to obtain context vector, and then performs softmax normalization calculation to obtain the probability table of each word as the generation probability of the word of Decoder at the current time t. Pointer Network also adds the Pointer algorithm part, which is used to copy the words in the source text. The specific principle can be: also through the attention mechanism, calculate the weight of each word in the source input sequence, and perform weighted sum to obtain the second context vector, perform softmax normalization calculation on the vector, obtain the probability distribution of each word in the input sequence, and add the probability distribution of the source sequence and the probability distribution of the vocabulary table obtained by Seq2Seq to obtain the total probability distribution of the word, which is used as the probability of generating each word.

[0117] Since Pointer Network not only has the ability to generate seq2seq, but also has the ability to copy the vocabulary in the source text, preferably, the comment information generation model in this specification can be obtained by training Pointer Network. Among them, for the PointerNetwork used in training, its network structure and parameters can be designed according to the actual needs of comprehensive comment information generation, and are not specifically limited here.

[0118] In step 407, the execution subject may display the generated comprehensive review information to the user. When the execution subject is a client, the execution subject may directly display the comprehensive review information to the user. When the execution subject is a server, the execution subject may send the comprehensive review information to the client used by the user, so that the client displays the comprehensive review information to the user.

[0119] Continue to see Figure 6a , Figure 6a Another schematic diagram of an application scenario of the information processing method of this embodiment is shown. In this application scenario, the threshold of the number of prompts is 2. The client is a shopping application, and the user is a buyer who has purchased clothes through the shopping application. In addition to Figure 3a In addition to the content shown, the following interactive scenarios are also included:

[0120] After the buyer sees the prompt message “Q: What do you think is wrong with the clothes?” displayed by the client, the buyer can input a supplementary evaluation for the evaluation information “the clothes are not good” through the client, such as “the material of the clothes is not good”, as shown in reference numeral 307. Specifically, the buyer can, for example, Figure 6b Submit additional comments on the post review interface shown. Figure 6b Another schematic diagram of the interface for posting follow-up comments is shown. After receiving the supplementary comments, the client can send the supplementary comments to the server as shown by label 308. After receiving the supplementary comments, the server can, in response to the number of prompts not reaching 2 times, use a predetermined analysis algorithm to analyze the supplementary comments to obtain the supplementary comments analysis results as shown by label 309. The analysis results show that the supplementary comments lack descriptive information on the materials of the clothes, as well as the location of the missing descriptive information on the materials of the clothes in the supplementary comments and the question phrase "What specific problems have occurred?" Then, the server can generate a prompt message "Q: What specific problems do you think have occurred with the materials?" based on the supplementary comments and the location and question phrase in the analysis result as shown by label 310. Then, the server can send the prompt message to the client as shown by label 311. The client can display the prompt message to the buyer as shown by label 312, so that the buyer continues to supplement the descriptive information on the materials of the clothes based on the evaluation information. Here, the display effect of the prompt message can be as shown in the following example. Figure 6c As shown. Among them, Figure 6c Another schematic diagram showing the display effect of prompt information.

[0121] When the buyer sees the prompt message "Q: What specific problems do you think there are with the material?" displayed by the client, the buyer can input a supplementary evaluation of the evaluation information "the clothes are not good" through the client again, such as "there are burrs on the corners of the material", as shown in reference numeral 313. Specifically, the buyer can, for example, Figure 6d Submit the supplementary review through the posting review interface shown. Figure 6d Another schematic diagram of the interface for posting a follow-up review is shown. After receiving the supplementary evaluation, the client can send the supplementary evaluation to the server as shown by label 314. After receiving the supplementary evaluation, the server can, in response to the number of prompts reaching 2, use the pre-trained comment information generation model to generate comprehensive evaluation information based on the evaluation information and the two input supplementary evaluations, such as "there are burrs on the edges of the clothing material", as shown by label 315, and send the comprehensive evaluation information to the client as shown by label 316. In response to receiving the comprehensive evaluation information, the client can display the comprehensive evaluation information as shown by label 317. Here, the display effect of the comprehensive evaluation information can be as shown in the following example: Figure 6e As shown. Among them, Figure 6e A schematic diagram showing the display effect of comprehensive evaluation information.

[0122] In some optional implementations of this embodiment, the comment information input by the user may have a corresponding associated person. The associated person may be any type of person. Specifically, when the user is a buyer, the associated person may be a seller. When the user is a user who publishes public opinion information, the associated person may be an administrator who manages public opinion information. When the user is a user who makes comments on service content, the associated person may be an administrator of the feedback information of the service content. When the user is an article reviewer, the associated person may be the author of the article. When the user is an audio and video reviewer, the associated person may be an audio and video producer.

[0123] When the execution subject is a client, after executing step 406, the execution subject may send the generated comprehensive review information to the server. The server may send the comprehensive review information to the client used by the associated user, so that the client displays the comprehensive review information to the associated user. When the execution subject is a server, after executing step 406, the execution subject may send the comprehensive review information to the client used by the associated user, so that the client displays the comprehensive review information to the associated user.

[0124] Thus, the generated relatively comprehensive review information can be sent to the client of the corresponding associated person, so that the associated person can obtain relatively comprehensive review information and understand the content expressed by the user. In addition, the communication between the associated person and the user can be made smoother. For example, when the user is a buyer and the associated person is a seller, the communication between the seller and the buyer can be made smoother.

[0125] Continue to refer Figure 7 , Figure 7 is another schematic diagram of an application scenario of the information processing method according to this embodiment. Figure 6a The application scenario shown is similar. In this application scenario, the clothes evaluated by the buyer are sold by seller A. Figure 6a In addition to the content shown, the server can also send the comprehensive evaluation information "there are burrs on the corners of the clothing material" to the client used by seller A as shown in label 318, so that the client can display the comprehensive evaluation information as shown in label 319 for seller A to view. Therefore, in this application scenario, the server can send the comprehensive evaluation information to the client used by the buyer and the client used by seller A respectively, so that both the buyer and seller A can see the comprehensive evaluation information.

[0126] from Figure 4 It can be seen that Figure 2 Compared with the corresponding embodiment, the process 400 of the information processing method in this embodiment highlights the steps of receiving the supplementary comments provided by the user for the input comment information, generating comprehensive comment information based on the comment information and supplementary comments input by the user, and displaying the comprehensive comment information to the user. The scheme described in this embodiment allows the user to adjust the comment information by submitting supplementary comments. In addition, compared with the comment information input by the user, the comprehensive comment information is more comprehensive and can better reflect the user's true intention. Therefore, the scheme described in this embodiment can achieve the generation of more comprehensive comment information.

[0127] exist Figure 2 and Figure 4In some optional implementations of the information processing methods provided in the corresponding embodiments, in order to quickly and accurately locate the missing elements of the comment information input by the user, the above-mentioned execution subject can adopt the following first analysis method: select the comment information matching the comment information input by the user from the pre-collected comment information set related to the comment object pointed to by the comment information input by the user; analyze whether the comment information input by the user is missing the predetermined information element based on the part of speech of each word in the comment information input by the user and the selected comment information; generate the analysis result according to the determination result. It should be pointed out that by performing part of speech analysis on the comment information input by the user and the selected comment information, the correlation between the parts of speech can be analyzed, and according to the correlation between the parts of speech, the position of the missing part of the parts of speech of the two comment information can be quickly determined, and then the missing elements of the comment information input by the user can be quickly located.

[0128] As an example, the execution subject may calculate the similarity between the comment information input by the user and each comment information in the comment information set, select the comment information with the highest similarity to the comment information input by the user from the comment information set, and use the selected comment information as the comment information matching the comment information input by the user. It should be understood that the execution subject may use any text similarity calculation method to calculate the similarity of comment information, and this specification does not specifically limit this.

[0129] As another example, the comment information in the comment information set may be provided with a corresponding click-through rate value. The click-through rate value may be determined based on the historical click volume of the corresponding comment information. The above-mentioned execution subject may select comment information from the comment information set based on the click-through rate value corresponding to the comment information in the comment information set, for example, select the comment information corresponding to the maximum click-through rate value, and determine the selected comment information as the comment information that matches the comment information input by the user.

[0130] In addition, when the determination result shows that the comment information input by the user does not lack the predetermined information elements, the above-mentioned execution subject may generate an analysis result including any of the following items: the comment information input by the user, and a preset flag used to characterize that the predetermined information elements are not missing. When the determination result shows that the comment information input by the user lacks the predetermined information elements, the above-mentioned execution subject may generate an analysis result including an information group, and the information group may include: the position of the missing element item in the comment information input by the user, and the question phrase corresponding to the position of the missing element item. Optionally, the information group may also include at least one of the following items: the position of the missing element item in the comment information that matches the comment information input by the user, and the words at that position in the comment information that matches the comment information input by the user.

[0131] It should be noted that the above-mentioned execution entity may store a pre-trained question phrase prediction model, which is used to predict the question phrases corresponding to the element items. The model may be obtained by training a model such as Transformer. The Transformer model adopts an encoder-decoder architecture and can be used for text processing. Here, when the determination result shows that the comment information input by the user is missing a predetermined information element, the above-mentioned execution entity may use the model to predict the question phrase corresponding to the element item missing from the comment information input by the user.

[0132] It should be emphasized that when the above-mentioned execution entity adopts the above-mentioned first analysis method to analyze the input comment information, the above-mentioned execution entity can also adopt a method similar to the above-mentioned first analysis method to analyze the supplementary comments provided in response to the comment information input by the user.

[0133] exist Figure 2 and Figure 4 In some optional implementations of the information processing methods provided in the corresponding embodiments, in order to improve the language quality of the generated prompt information and make the part of speech collocation of the prompt information more accurate, the above-mentioned execution subject can use a pre-trained question generation model to generate prompt information. Specifically, the above-mentioned execution subject can input the comment information input by the user and the information group in the analysis result obtained by the above-mentioned first analysis method into the question generation model to obtain the prompt information. Among them, the question generation model can be obtained by training a model for text generation such as a Seq2Seq model or a Pointer Network.

[0134] Since the Seq2Seq model has the problems described above, preferably, the question generation model can be obtained by training the Pointer Network. Here, for the Pointer Network used in model training, its network structure and parameters can be designed according to the actual needs of prompt information generation, and are not specifically limited here.

[0135] Further references Figure 8 As an implementation of the methods shown in some of the above figures, this specification provides an embodiment of an information processing device, which is similar to Figure 2 Corresponding to the method embodiment shown, the device can be applied to a client (eg Figure 1a The client 101 shown in FIG. 101 may also be applied to a server (eg Figure 1b The server 102 is shown).

[0136] like Figure 8As shown, the information processing device 800 of this embodiment includes: a receiving unit 801, an analyzing and determining unit 802, and an output unit 803. The receiving unit 801 is configured to receive comment information input by a user; the analyzing and determining unit 802 is configured to analyze the comment information input by the user and determine prompt information to be provided to the user, the prompt information is used to prompt the user to adjust the comment information; the output unit 803 is configured to display the prompt information to the user.

[0137] In this embodiment, the specific processing of the receiving unit 801, the analyzing and determining unit 802, and the output unit 803 and the technical effects thereof can be referred to in Figure 2 The relevant descriptions of step 201, step 202 and step 203 in the corresponding embodiment are not repeated here.

[0138] In some optional implementations of the present embodiment, the above-mentioned device 800 may also include a generating unit (not shown in the figure); the receiving unit 801 may also be configured to receive supplementary comments provided by the user for the input comment information; the generating unit may be configured to generate comprehensive comment information based on the comment information and supplementary comments input by the user; the output unit 803 may also be configured to display the comprehensive comment information to the user.

[0139] In some optional implementations of this embodiment, the comment object pointed to by the comment information input by the user may have a corresponding associated person. If the above device 800 is applied to the server, the output unit 803 may also be configured to, after the generation unit generates the comprehensive comment information, send the comprehensive comment information to the client used by the associated person, so that the client displays the comprehensive comment information.

[0140] In some optional implementations of this embodiment, the user may be a buyer, and the above-mentioned associated person may be a seller; or, the user may be a user who publishes public opinion information, and the above-mentioned associated person may be an administrator who manages public opinion information; or, the user may be a user who makes comments on the service content, and the above-mentioned associated person may be an administrator of the feedback information of the service content.

[0141] In some optional implementations of the present embodiment, the analysis and determination unit 802 may include: an analysis subunit (not shown in the figure), configured to analyze the comment information input by the user to obtain an analysis result, and the analysis result may show whether the comment information input by the user is missing predetermined information elements; a determination subunit (not shown in the figure), configured to determine the prompt information that needs to be provided to the user in response to the analysis result showing that the comment information input by the user is missing predetermined information elements.

[0142] In some optional implementations of this embodiment, the predetermined information element may include at least one of the following element items: attribute information of the review object, and description information for the attribute information of the review object.

[0143] In some optional implementations of the present embodiment, when the missing element item in the comment information input by the user is the attribute information of the comment object, the prompt information can be used to guide the user to supplement the attribute information of the comment object; when the missing element item in the comment information input by the user is the description information of the attribute information of the comment object, the prompt information is used to guide the user to supplement the description information of the attribute information of the comment object.

[0144] In some optional implementations of this embodiment, the comment object may be any one of the following: a product object, a public opinion object, or a service object.

[0145] In some optional implementations of the present embodiment, the receiving unit 801 may be further configured to: receive supplementary comments input by the user in response to prompt information; determine whether the number of prompts reaches a preset number, wherein the number of prompts refers to the number of times the prompt information is displayed to the user; if the number of prompts does not reach the preset number, analyze the supplementary comments to obtain a supplementary comment analysis result; if the supplementary comment analysis result shows that the comment needs to be supplemented, determine the prompt information that needs to be supplemented to the comment; display the determined prompt information to the user so that the user can supplement the supplementary comment based on the prompt information; receive the supplementary comments input by the user in response to the prompt information corresponding to the supplementary comment.

[0146] In some optional implementations of this embodiment, the receiving unit 801 can also be configured to: if it is determined that the number of prompts reaches the preset number, or the supplementary comment analysis results show that there is no need to supplement the comments, then the notification generation unit executes the above-mentioned comment information and supplementary comments based on the user input to generate comprehensive comment information.

[0147] In some optional implementations of the present embodiment, the analysis sub-unit may be further configured to: select comment information that matches the comment information input by the user from a pre-collected set of comment information related to the comment object pointed to by the comment information input by the user; determine whether the comment information input by the user is missing predetermined information elements based on the part of speech of each word in the comment information input by the user and the selected comment information; and generate an analysis result based on the determination result.

[0148] In some optional implementations of this embodiment, the comment information in the comment information set may be set with a corresponding click-through rate value; and the analysis subunit may be further configured to: select comment information from the comment information set based on the click-through rate value corresponding to the comment information in the comment information set; and determine the selected comment information as comment information that matches the comment information input by the user.

[0149] In some optional implementations of this embodiment, the analysis subunit may be further configured to: when the determination result shows that the comment information input by the user is not missing predetermined information elements, generate an analysis result including any one of the following items: the comment information input by the user, and a preset flag for characterizing that the predetermined information elements are not missing.

[0150] In some optional implementations of this embodiment, the analysis sub-unit may be further configured to: when the determination result shows that the comment information input by the user is missing a predetermined information element, generate an analysis result including an information group, the information group including: the position of the missing element item in the comment information input by the user, and a question phrase corresponding to the position of the missing element item; wherein the question phrase is a phrase used to ask a question about the missing element item at the position of the missing element item.

[0151] In some optional implementations of the present embodiment, the above-mentioned information group may also include at least one of the following items: the position of the missing element item at the position of the missing element item in the comment information matching the comment information input by the user, and the word at that position in the comment information matching the comment information input by the user.

[0152] In some optional implementations of this embodiment, the analysis results may also show, in the case where a predetermined information element is missing, an information group related to the missing element item; and the determination subunit may be further configured to: determine the prompt information that needs to be provided to the user based on the comment information input by the user and the above-mentioned information group.

[0153] In some optional implementations of this embodiment, the determination subunit may be further configured to: input the comment information input by the user and the above-mentioned information group into a pre-trained prompt information generation model to obtain prompt information, wherein the prompt information generation model may be obtained by training a pointer network.

[0154] In some optional implementations of this embodiment, the generation unit may be further configured to: input the comment information and supplementary comments input by the user into a pre-trained comment information generation model to obtain comprehensive comment information, wherein the comment information generation model may be obtained by training a pointer network.

[0155] In some optional implementations of this embodiment, when the above-mentioned device 800 is applied to a client, the output unit 803 may also be configured to: send the comprehensive comment information to the server.

[0156] In some optional implementations of this embodiment, when the above-mentioned device 800 is applied to the server, the receiving unit 801 can be further configured to: receive comment information input by the user from the client; and the output unit 803 can be further configured to: send prompt information to the client, so that the client displays the prompt information to the user.

[0157] The information processing device provided in this embodiment receives the comment information input by the user through the receiving unit, and then analyzes the comment information input by the user through the analysis and determination unit to determine the prompt information that needs to be provided to the user, the prompt information is used to prompt the user to adjust the comment information, and then the prompt information is displayed to the user through the output unit, so that the user can adjust the input comment information based on the prompt information. In this way, intelligent guidance of the user can be achieved to obtain more comprehensive comment information.

[0158] The embodiments of this specification also provide a computer-readable storage medium on which a computer program is stored. When the computer program is executed in a computer, the computer is caused to execute the solutions respectively shown in the above method embodiments.

[0159] The embodiments of the present specification also provide a computing device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the solutions shown in the above method embodiments are implemented.

[0160] The embodiments of this specification also provide a computer program product, which, when executed on a data processing device, enables the data processing device to implement the solutions shown in the above method embodiments.

[0161] Those skilled in the art should be aware that in one or more of the above examples, the functions described in the multiple embodiments disclosed in this specification can be implemented by hardware, software, firmware, or any combination thereof. When implemented by software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium.

[0162] In some cases, the actions or steps described in the claims may be performed in a different order than in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0163] The specific implementation methods described above further illustrate in detail the purposes, technical solutions and beneficial effects of the multiple embodiments disclosed in this specification. It should be understood that the above description is only the specific implementation methods of the multiple embodiments disclosed in this specification, and is not used to limit the protection scope of the multiple embodiments disclosed in this specification. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solutions of the multiple embodiments disclosed in this specification should be included in the protection scope of the multiple embodiments disclosed in this specification.

Claims

1. An information processing method, include: Receive comment information entered by the user; Analyze the input review information to obtain an analysis result, which shows whether the input review information lacks a predetermined information element; In response to the analysis result showing that the input review information lacks a predetermined information element, determining prompt information to be provided to the user, the prompt information being used to prompt the user to adjust the review information; Displaying the prompt information to the user; Among them, obtaining the analysis result includes: selecting comment information that matches the input comment information from a pre-collected set of comment information related to the comment object pointed to by the input comment information; determining whether the input comment information is missing the predetermined information element based on the part of speech of each word in the input comment information and the selected comment information; and generating the analysis result based on the determination result.

2. The method according to claim 1, in, The method further comprises: receiving a supplementary comment provided by the user in response to the input comment information; generating comprehensive review information based on the input review information and the supplementary review; The comprehensive review information is displayed to the user.

3. The method according to claim 2, in, The comment object pointed to by the input comment information has a corresponding associated person; the execution subject of the method is the server; as well as After generating comprehensive comment information based on the input comment information and the supplementary comment, the method further includes: The comprehensive review information is sent to a client used by the associated person, so that the client displays the comprehensive review information.

4. The method according to claim 3, in, The user is a buyer and the associated person is a seller; or, The user is a user who publishes public opinion information, and the associated person is an administrator who manages public opinion information; or The user is a user who has made comments on the service content, and the associated person is an administrator of the feedback information of the service content.

5. The method according to claim 1, in, The predetermined information element includes at least one of the following element items: attribute information of the review object, and description information of the attribute information of the review object.

6. The method according to claim 5, in, When the missing element item of the input review information is the attribute information of the review object, the prompt information is used to guide the user to supplement the attribute information of the review object; When the missing element item of the input review information is the description information of the attribute information of the review object, the prompt information is used to guide the user to supplement the description information of the attribute information of the review object.

7. The method according to claim 5, in, The review object is any one of the following: product object, public opinion object, service object.

8. The method according to claim 2, in, The receiving a supplementary comment provided by the user in response to the input comment information includes: receiving a supplementary comment input by the user in response to the prompt information; Determine whether the number of prompts reaches a preset number, where the number of prompts is the number of times the prompt information is shown to the user; If the number of prompts does not reach the preset number, analyze the supplementary comment to obtain a supplementary comment analysis result; If the supplementary comment analysis result shows that the comment needs to be supplemented, determine the prompt information that needs to be supplemented for the comment; Show the determined prompt information to the user so that the user can supplement the supplementary comment based on the prompt information; Receive the supplementary comment input by the user in response to the prompt information corresponding to the supplementary comment.

9. The method according to claim 8, wherein, the method further includes: If it is determined that the number of prompts reaches the preset number, or the supplementary comment analysis result shows that the comment does not need to be supplemented, then execute generating comprehensive comment information based on the input comment information and the supplementary comment.

10. The method according to claim 1, wherein, the comment information in the comment information set is set with a corresponding click-through rate value; and selecting the comment information that matches the input comment information from the comment information set collected in advance related to the comment object pointed to by the input comment information includes: selecting comment information from the comment information set based on the click-through rate value corresponding to the comment information in the comment information set; determining the selected comment information as the comment information that matches the input comment information.

11. The method according to claim 1, wherein, generating an analysis result according to the determination result includes: When the determination result shows that the input comment information does not lack the predetermined information elements, generating an analysis result including any one of the following: the input comment information, a preset flag for indicating that the predetermined information elements are not lacking.

12. The method according to claim 1, wherein, generating an analysis result according to the determination result includes: When the determination result shows that the input comment information lacks the predetermined information elements, generating an analysis result including an information group, the information group including: the position of the missing element item in the input comment information, a question phrase corresponding to the position of the missing element item; wherein, the question phrase is a phrase for asking questions about the element item missing at the position of the missing element item.

13. The method according to claim 12, wherein, the information group further includes at least one of the following: the position of the element item missing at the position of the missing element item in the comment information that matches the input comment information, the word at the position in the comment information that matches the input comment information and is at the position.

14. The method according to claim 1, wherein, the analysis result also shows an information group related to the missing element item in the case of lacking the predetermined information elements; and determining the prompt information that needs to be provided to the user includes: determining the prompt information that needs to be provided to the user according to the input comment information and the information group.

15. The method according to claim 14, wherein, determining the hint information to be provided to the user according to the input comment information and the information group includes: inputting the input comment information and the information group into a pre-trained hint information generation model to obtain the hint information, wherein the hint information generation model is obtained by training a pointer network.

16. The method according to claim 2, wherein, generating the comprehensive comment information based on the input comment information and the supplementary comment includes: inputting the input comment information and the supplementary comment into a pre-trained comment information generation model to obtain the comprehensive comment information, wherein the comment information generation model is obtained by training a pointer network.

17. The method according to claim 2, wherein, the execution subject of the method is the client; and after generating the comprehensive comment information based on the input comment information and the supplementary comment, the method further includes: sending the comprehensive comment information to the server.

18. The method according to claim 1, wherein, the execution subject of the method is the server; and receiving the comment information input by the user includes: receiving the comment information input by the user from the client; and displaying the hint information to the user includes: sending the hint information to the client so that the client displays the hint information to the user.

19. An information processing device, comprising: a receiving unit configured to receive the comment information input by the user; an analysis and determination unit configured to analyze the input comment information to obtain an analysis result, which shows whether the input comment information lacks a predetermined information element; in response to the analysis result showing that the input comment information lacks a predetermined information element, determining the hint information to be provided to the user, where the hint information is used to prompt the user to adjust the comment information; an output unit configured to display the hint information to the user; wherein the analysis and determination unit is specifically configured to: select the comment information matching the input comment information from a pre-collected set of comment information related to the comment object pointed to by the input comment information; determine whether the input comment information lacks the predetermined information element based on the part of speech of each word in both the input comment information and the selected comment information; and generate an analysis result according to the determination result.

20. The device according to claim 19, wherein, the device further includes a generation unit; the receiving unit is further configured to receive the supplementary comment provided by the user for the input comment information; the generation unit is configured to generate comprehensive comment information based on the input comment information and the supplementary comment; the output unit is further configured to display the comprehensive comment information to the user.

21. A computer-readable storage medium, on which a computer program is stored, wherein, When the computer program is executed in a computer, the computer is caused to execute the method according to any one of claims 1-18.

22. A computing device, comprising a memory and a processor, wherein, an executable code is stored in the memory, and when the processor executes the executable code, the method according to any one of claims 1-18 is implemented.

Images