Intelligent comment management method and device in social media, storage medium and computer equipment
By automatically identifying the emotional tendencies of comments and automatically responding with preset templates, the problem of inefficient social media comment management is solved, and efficient comment processing and brand image maintenance is achieved.
Patent Information
- Application Number
- CN202410791146.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-19
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-06-19
AI Technical Summary
The existing social media comment management tools lack the ability to intelligently analyze and automate reply generation, resulting in inefficiency and inability to effectively process large and complex comments, affecting the user experience and brand image.
通过自动识别评论的情感倾向,并使用预设模板进行自动回复,将评论分为智能处理组和暂缓处理组,智能处理组中的评论进行语义分析自动匹配回复模板,暂缓处理组中的评论由管理员手动处理。
It significantly reduces the time and cost of manually processing comments, improves management efficiency, responds to user comments in a timely manner, improves user experience and brand image, and ensures the accuracy and appropriateness of processing results.
Smart Images

Figure CN118916483A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of Internet technology, and in particular to a method and device, storage medium, and computer equipment for intelligent management of comments in social media. Background Art
[0002] With the rapid development of the Internet and social media, user interactions on websites, blogs, and social networks are mainly conducted in the form of comments. The number of these comments has exploded as the user base has expanded. Traditional comment management methods rely on manual review and reply, which is inefficient and unable to respond in a timely manner. In addition, malicious comments, spam, and inappropriate remarks increase the difficulty of management and affect user experience and brand image.
[0003] Current comment management tools usually provide basic functions such as keyword filtering, automatic tagging, and manual batch processing, but this still cannot meet the needs of managers to handle large and complex comments. In addition, existing tools lack the ability to perform intelligent analysis and automatically generate responses. Summary of the invention
[0004] In view of this, the embodiments of the present application provide a method and apparatus, storage medium, and computer equipment for intelligent management of comments in social media. By automatically identifying the emotional tendencies of comments and using preset templates for automatic replies, the system can significantly reduce the time and cost of manually processing comments, thereby improving management efficiency.
[0005] According to one aspect of the present application, a method for intelligent management of comments in social media is provided, the method comprising:
[0006] Based on the target published content of the target management account in the social media, obtaining at least one comment data corresponding to the target published content, and identifying the sentiment tendency of each of the comment data respectively, wherein the sentiment tendency is one of positive, negative and neutral;
[0007] Identify offensive comment data among the comment data with negative sentiment tendency, and use the remaining comment data as comment data to be replied, and divide the comment data to be replied into an intelligent processing group and a postponed processing group according to the corresponding sentiment tendency;
[0008] For the comment data to be replied in the intelligent processing group, semantic analysis is performed on the comment data to be replied, and based on the semantic analysis result, a target reply template matching the comment data to be replied is obtained from a preset automatic reply template set, and the target reply template is used to automatically reply to the comment data to be replied;
[0009] A first prompt message is sent to the management terminal of the target management account to prompt the processing of the to-be-replied comment data in the suspended processing group.
[0010] Optionally, the to-be-replied comment data are divided into an intelligent processing group and a postponed processing group according to corresponding emotional tendencies, including:
[0011] Divide the comments to be replied with positive or neutral sentiment into intelligent processing groups;
[0012] The review data with negative sentiment tendency are divided into the postponed processing group.
[0013] Optionally, semantic analysis is performed on the comment data to be replied to, and based on the result of the semantic analysis, a target reply template matching the comment data to be replied to is obtained from a preset automatic reply template set, including:
[0014] Identifying the comment intention of the comment data to be replied, wherein the comment intention is a statement or an inquiry;
[0015] For the comment data to be replied whose comment intention is statement, according to the emotional tendency corresponding to the comment data to be replied, a target reply template matching the comment data to be replied is obtained in a first preset automatic reply template set, wherein the preset automatic reply template set includes the first preset automatic reply template set, and the first preset automatic reply template set includes at least one predefined standard reply content for positive or neutral comment data;
[0016] For the comment data to be replied whose comment intention is to inquire, semantic recognition is performed on the comment data to be replied, and a query is made in the second preset automatic reply template set whether there is a preset question matching the semantic recognition result. If so, the reply template corresponding to the matching preset question is obtained in the second preset automatic reply template set as the target reply template for the comment data to be replied.
[0017] Optionally, after querying whether there is a preset question matching the semantic recognition result in the second preset automatic reply template set, the method further includes:
[0018] If it does not exist, a prompt word is generated based on the comment data to be replied, the target published content and the reply attribute information corresponding to the target management account, and a large language model is called to generate at least one reference reply information corresponding to the prompt word, and a second prompt information is generated based on the comment data to be replied and the reference reply information and sent to the management terminal.
[0019] Optionally, before generating the prompt words based on the comment data to be replied, the target published content and the reply attribute information corresponding to the target management account, the method further includes:
[0020] Acquire the historical published content of the target management account, the historical comment data of the historical published content, and the historical reply data of the historical comment data;
[0021] The reply attribute information of the target management account is generated through a large language model according to the historical published content, the historical comment data and the historical reply data, wherein the reply attribute information includes the reply language style of the target management account.
[0022] Optionally, after generating second prompt information based on the comment data to be replied and the reference reply information and sending it to the management terminal, the method further includes:
[0023] Receiving a confirmation instruction for any reference reply information or a modified confirmation instruction for any reference reply information, and determining a target reply template for the comment data to be replied to according to the confirmed reference reply information;
[0024] A summary analysis is performed based on the comment data to be replied, and the summary of the comment data to be replied is used as a new preset question, and the target reply template is used as a reply template for the new preset question, and added to the second preset automatic reply template set.
[0025] Optionally, based on target published content of a target management account in social media, obtaining at least one comment data corresponding to the target published content includes:
[0026] By calling the API of the social media platform, based on the target published content of the target management account in the social media, obtaining at least one comment data corresponding to the target published content;
[0027] The correlation between the comment data and the target published content is analyzed, and the comment data with a correlation lower than a preset threshold is deleted.
[0028] According to another aspect of the present application, a device for intelligently managing comments in social media is provided, the device comprising:
[0029] A data acquisition module is used to acquire at least one comment data corresponding to the target published content of the target management account in the social media, and respectively identify the sentiment tendency of each of the comment data, wherein the sentiment tendency is one of positive, negative and neutral;
[0030] A grouping module, used to identify offensive comment data among the comment data with negative sentiment tendency, and use the remaining comment data as the comment data to be replied, and divide the comment data to be replied into an intelligent processing group and a postponed processing group according to the corresponding sentiment tendency;
[0031] An intelligent reply module, for performing semantic analysis on the comment data to be replied in the intelligent processing group, obtaining a target reply template matching the comment data to be replied from a preset automatic reply template set based on the semantic analysis result, and automatically replying to the comment data to be replied using the target reply template;
[0032] The sending module is used to send a first prompt message to the management terminal of the target management account to prompt the processing of the to-be-replied comment data in the suspended processing group.
[0033] Optionally, the grouping module is further used to:
[0034] Divide the comments to be replied with positive or neutral sentiment into intelligent processing groups;
[0035] The review data with negative sentiment tendency are divided into the postponed processing group.
[0036] Optionally, the intelligent reply module is further used to:
[0037] Identifying the comment intention of the comment data to be replied, wherein the comment intention is a statement or an inquiry;
[0038] For the comment data to be replied whose comment intention is statement, according to the emotional tendency corresponding to the comment data to be replied, a target reply template matching the comment data to be replied is obtained in a first preset automatic reply template set, wherein the preset automatic reply template set includes the first preset automatic reply template set, and the first preset automatic reply template set includes at least one predefined standard reply content for positive or neutral comment data;
[0039] For the comment data to be replied whose comment intention is to inquire, semantic recognition is performed on the comment data to be replied, and a query is made in the second preset automatic reply template set whether there is a preset question matching the semantic recognition result. If so, the reply template corresponding to the matching preset question is obtained in the second preset automatic reply template set as the target reply template for the comment data to be replied.
[0040] Optionally, the intelligent reply module is further used to:
[0041] If it does not exist, a prompt word is generated based on the comment data to be replied, the target published content and the reply attribute information corresponding to the target management account, and a large language model is called to generate at least one reference reply information corresponding to the prompt word, and a second prompt information is generated based on the comment data to be replied and the reference reply information and sent to the management terminal.
[0042] Optionally, the intelligent reply module is further used to:
[0043] Acquire the historical published content of the target management account, the historical comment data of the historical published content, and the historical reply data of the historical comment data;
[0044] The reply attribute information of the target management account is generated through a large language model according to the historical published content, the historical comment data and the historical reply data, wherein the reply attribute information includes the reply language style of the target management account.
[0045] Optionally, the intelligent reply module is further used to:
[0046] Receiving a confirmation instruction for any reference reply information or a modified confirmation instruction for any reference reply information, and determining a target reply template for the comment data to be replied to according to the confirmed reference reply information;
[0047] A summary analysis is performed based on the comment data to be replied, and the summary of the comment data to be replied is used as a new preset question, and the target reply template is used as a reply template for the new preset question, and added to the second preset automatic reply template set.
[0048] Optionally, the data acquisition module is further used to:
[0049] By calling the API of the social media platform, based on the target published content of the target management account in the social media, obtaining at least one comment data corresponding to the target published content;
[0050] The correlation between the comment data and the target published content is analyzed, and the comment data with a correlation lower than a preset threshold is deleted.
[0051] According to another aspect of the present application, a storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned intelligent management method of comments in social media is implemented.
[0052] According to another aspect of the present application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the above-mentioned intelligent management method for comments in social media when executing the program.
[0053] By means of the above technical scheme, the embodiment of the present application provides a method and device for intelligent management of comments in social media, a storage medium, and a computer device, collects the comment data of the target account, identifies its emotional tendency (positive, negative, neutral), and then identifies the offensive content in the negative comments, and divides the remaining comments into an intelligent processing group and a suspended processing group. The comments in the intelligent processing group automatically match the reply template and reply through semantic analysis, while the comments in the suspended processing group prompt the administrator to manually process them. The embodiment of the present application uses automated emotional tendency recognition and intelligent reply functions, and the system can significantly reduce the time and cost of manually processing comments, thereby improving the overall processing efficiency, helping to respond to the comments of social media users in a timely manner, improving the user's sense of participation and satisfaction, and helping to maintain the good image and reputation of the target account. By distinguishing between the intelligent processing group and the suspended processing group, the processing flow is made more flexible. Simple comments can be quickly processed by the system, while complex or sensitive comments are manually processed by the administrator to ensure the accuracy and appropriateness of the processing results. By identifying the offensive content in negative comments, the system can help administrators to promptly discover and deal with potential negative public opinion, reduce the negative impact on the target account, and maintain the brand image. In addition, the comment data collected and analyzed by the system can provide strong data support for the operation of the target account, helping administrators better understand user needs and market trends and make more informed decisions.
[0054] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0055] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0056] Figure 1 A schematic diagram of a process of an intelligent management method for comments in social media provided by an embodiment of the present application is shown;
[0057] Figure 2 A schematic diagram of a process of another intelligent management method for comments in social media provided by an embodiment of the present application is shown;
[0058] Figure 3 A schematic diagram of the structure of an intelligent management device for comments in social media provided in an embodiment of the present application is shown. DETAILED DESCRIPTION
[0059] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.
[0060] In this embodiment, a method for intelligent management of comments in social media is provided. Figure 1 As shown, the method includes:
[0061] Step 101, based on the target published content of the target management account in the social media, obtain at least one comment data corresponding to the target published content, and identify the emotional tendency of each of the comment data respectively, wherein the emotional tendency is one of positive, negative and neutral.
[0062] In this embodiment, a workflow of a comment management system for content posted by a target account in social media is described. Specifically, these steps include collecting relevant comment data of the target content, identifying sentiment tendency, classifying, and automating processing or prompting manual processing. In actual application scenarios, this workflow can be automated through pre-developed plug-ins. First, collect the comment data corresponding to a certain content posted by the target account on social media. Then, the system will perform sentiment tendency analysis on each comment and classify it as positive, negative or neutral.
[0063] Step 102, identifying the offensive comment data in the comment data with negative sentiment tendency, and taking the remaining comment data as the comment data to be replied, and dividing the comment data to be replied into an intelligent processing group and a suspended processing group according to the corresponding sentiment tendency.
[0064] In this embodiment, after identifying negative comments, the system will further identify whether these negative comments contain offensive content. For negative comments that are not offensive, they will be regarded as "comment data to be replied". These comment data to be replied will be further divided into "intelligent processing group" and "deferred processing group" according to their emotional tendency (negative or neutral).
[0065] Step 103, for the comment data to be replied in the intelligent processing group, perform semantic analysis on the comment data to be replied, obtain a target reply template matching the comment data to be replied from a preset automatic reply template set based on the semantic analysis result, and use the target reply template to automatically reply to the comment data to be replied.
[0066] In this embodiment, the system will perform a more in-depth semantic analysis on the comments to be replied in the intelligent processing group. Based on these analysis results, the system will search for templates matching the comments to be replied in the preset automatic reply template set, and automatically use these templates to reply to the comments.
[0067] Step 104: Send a first prompt message to the management terminal of the target management account to prompt the management terminal to process the pending comment data in the suspended processing group.
[0068] In this embodiment, finally, a prompt message can be sent to the management terminal responsible for managing the target account, informing the administrator that the data of the comments to be replied in the suspended processing group needs to be manually processed. The embodiment of the present application can significantly reduce the time and cost of manually processing comments by automatically identifying the emotional tendency of the comments and using preset templates for automatic replies, thereby improving management efficiency. Further segmentation of negative comments (i.e., identifying offensive comments) helps to ensure that the system only automatically processes comments that really need attention, while ensuring that administrators can give priority to those comments that may have a greater impact. The data of the comments to be replied is divided into intelligent processing groups and suspended processing groups, and different processing strategies are implemented based on these groups to ensure the flexibility and adaptability of the system. For simple comments that can be processed automatically, the system can process them immediately; for complex comments that require more thinking, the system will prompt the administrator to process them. Furthermore, by replying to user comments in a timely and appropriate manner, whether through automatic system replies or manual replies by administrators, the user experience can be effectively improved, and the user's favorability and loyalty to the target account can be enhanced.
[0069] By applying the technical solution of this embodiment, the comment data of the target account is collected, its emotional tendency (positive, negative, neutral) is identified, and then the offensive content in the negative comments is identified, and the remaining comments are divided into an intelligent processing group and a suspended processing group. The comments in the intelligent processing group automatically match the reply template and reply through semantic analysis, while the comments in the suspended processing group prompt the administrator to manually process them. The embodiment of the present application uses automated emotional tendency recognition and intelligent reply functions, and the system can significantly reduce the time and cost of manually processing comments, thereby improving the overall processing efficiency, helping to respond to the comments of social media users in a timely manner, improving the user's sense of participation and satisfaction, and helping to maintain the good image and reputation of the target account. By distinguishing between the intelligent processing group and the suspended processing group, the processing flow is made more flexible. Simple comments can be quickly processed by the system, while complex or sensitive comments are manually processed by the administrator to ensure the accuracy and appropriateness of the processing results. By identifying the offensive content in negative comments, the system can help administrators to promptly discover and deal with potential negative public opinion, reduce the negative impact on the target account, and maintain the brand image. In addition, the comment data collected and analyzed by the system can provide strong data support for the operation of the target account, helping administrators better understand user needs and market trends and make more informed decisions.
[0070] Further, as a refinement and extension of the specific implementation of the above embodiment, in order to fully illustrate the specific implementation process of this embodiment, another intelligent management method for comments in social media is provided, such as Figure 2 As shown, the method includes:
[0071] Step 201, by calling the API of the social media platform, based on the target published content of the target management account in the social media, obtain at least one comment data corresponding to the target published content; analyze the correlation between the comment data and the target published content, and delete the comment data with a correlation lower than a preset threshold; and identify the emotional tendency of each comment data respectively, wherein the emotional tendency is one of positive, negative and neutral.
[0072] In this embodiment, after authorization by the social media platform, the API (application programming interface) provided by the social media platform can be used to obtain data. After calling the API, the system will retrieve the comment data corresponding to the content based on the given target management account and the target content published on the social media. After obtaining the comment data, the system will analyze the relevance between each comment and the target published content. This is usually achieved through natural language processing technology, such as keyword matching, semantic similarity calculation, etc. If the relevance of a comment to the target published content is lower than a preset threshold, then this comment will be regarded as an irrelevant or low-quality comment and deleted from the subsequent processing flow. It should be noted that the deletion here is only not processed in the intelligent comment management method provided in the embodiment of the present application, and the social media platform still retains these data. In addition, these offensive comments can also be sent to the management terminal of the target management account separately for viewing and processing. Further, for the remaining comments related to the target published content, the system will further identify the emotional tendency of each comment. Emotional tendencies are usually classified as positive, negative or neutral. The identification of emotional tendencies is also achieved through natural language processing technology, such as matching based on emotional dictionaries, prediction methods of machine learning models (such as deep learning models), etc. The embodiment of the present application can filter out comments that are irrelevant or of low quality to the target published content through correlation analysis, thereby improving the data quality of subsequent analysis and focusing only on comments related to the target published content, which helps managers to more accurately understand users' feedback and attitudes towards the content. The identification of emotional tendencies helps managers understand users' emotional attitudes towards the target published content, whether they are positive support, negative criticism or neutral attitude, so as to further formulate more effective content strategies and interaction strategies.
[0073] Step 202, identifying the offensive comment data among the comment data with negative sentiment tendency, and taking the remaining comment data as the comment data to be replied; classifying the comment data to be replied with positive or neutral sentiment tendency into the intelligent processing group; and classifying the comment data with negative sentiment tendency into the suspended processing group.
[0074] In this embodiment, the comment data after the emotional tendency is identified is further classified and processed. Among all the comment data with negative emotional tendency, the system will further identify offensive comments. These offensive comments may contain insults, abuse, malicious slander and other contents. Identifying offensive comments can be based on natural language processing technology and a predefined offensive vocabulary library or model to identify offensive remarks by matching or analyzing text content. After excluding the offensive comments, the remaining comments with negative, positive or neutral emotional tendencies are regarded as comment data to be replied. For the comment data to be replied with positive or neutral emotional tendencies, the system will divide them into intelligent processing groups. The comments in the intelligent processing group can be replied or processed by preset automatic processing strategies, such as using automatic reply templates, automatic likes or thanks, etc. For the comment data to be replied with negative emotional tendencies, the system will divide them into the suspended processing group. The comments in the suspended processing group need to be handled more carefully and meticulously because they may involve users' complaints, dissatisfaction or suggestions, etc., which need to be focused on. These comments can be further analyzed and replied by manual administrators. The embodiment of the present application identifies and processes offensive comments, ignores offensive comments, and avoids the faster spread of offensive comments. For comments with positive or neutral sentiment, the use of automated processing strategies can greatly reduce the workload of manual processing and improve processing efficiency. Classifying comments with negative sentiment into a suspended processing group helps administrators focus more on issues that need to be focused on, such as user complaints and suggestions.
[0075] Step 203, for the comment data to be replied in the intelligent processing group, perform semantic analysis on the comment data to be replied, obtain a target reply template matching the comment data to be replied from a preset automatic reply template set based on the semantic analysis result, and use the target reply template to automatically reply to the comment data to be replied.
[0076] In this embodiment, the system will perform a more in-depth semantic analysis on the comments to be replied in the intelligent processing group. Based on these analysis results, the system will search for templates matching the comments to be replied in the preset automatic reply template set, and automatically use these templates to reply to the comments.
[0077] In the embodiment of the present application, optionally, in step 203, semantic analysis is performed on the comment data to be replied, and based on the semantic analysis result, a target reply template matching the comment data to be replied is obtained from a preset automatic reply template set, including:
[0078] Step 203-1, identifying the comment intention of the comment data to be replied, wherein the comment intention is a statement or an inquiry;
[0079] Step 203-2, for the comment data to be replied whose comment intention is statement, according to the emotional tendency corresponding to the comment data to be replied, obtain a target reply template matching the comment data to be replied in a first preset automatic reply template set, wherein the preset automatic reply template set includes the first preset automatic reply template set, and the first preset automatic reply template set includes at least one predefined standard reply content for positive or neutral comment data;
[0080] Step 203-3, for the comment data to be replied whose comment intention is an inquiry, semantic recognition is performed on the comment data to be replied, and a query is made in the second preset automatic reply template set whether there is a preset question matching the semantic recognition result; if so, a reply template corresponding to the matching preset question is obtained in the second preset automatic reply template set as the target reply template for the comment data to be replied; if not, a prompt word is generated based on the comment data to be replied, the target published content and the reply attribute information corresponding to the target management account, a large language model is called to generate at least one reference reply information corresponding to the prompt word, and a second prompt information is generated based on the comment data to be replied and the reference reply information and sent to the management terminal.
[0081] In the above embodiment, the comment intention of the comment data to be replied is identified, that is, whether the comment is a statement (such as expression of opinion, expression of emotion) or an inquiry (such as asking questions, seeking information). For comments intended to be statements, a matching template is searched in the first preset automatic reply template set according to the emotional tendency of the comment (positive or neutral). This template set usually contains standardized reply content for positive or neutral comments. For comments intended to be inquiries, semantic recognition is first performed to clarify the specific content of the inquiry, and then the second preset automatic reply template set is queried to see whether there is a preset question that matches the semantic recognition result. If so, the reply template corresponding to the question is used as the target reply template. If there is no matching preset question, the system will generate a prompt word based on the reply data to be replied, the target published content, and the reply attribute information of the target management account. Then, the system calls a large language model (such as GPT, BERT, etc.) to generate at least one reference reply message based on this prompt word. Finally, the system generates a second prompt message based on the comment data to be replied and the reference reply information, and sends it to the management terminal to prompt the administrator to reply manually or refer to the automatically generated reply information for reply. In the embodiment of the present application, through the preset automatic reply template, the system can quickly automatically reply to a large number of positive or neutral comments, reducing the time of manual processing and improving the efficiency of reply. Timely automatic replies can enhance the user's sense of participation and satisfaction, especially for those simple or common inquiries, automatic replies can provide immediate help and feedback. In addition, for inquiry comments, the system first tries to find a match in the preset template set. If it cannot be found, it uses a large language model to generate reference reply information, providing administrators with more options and processing methods. By analyzing the data of comments to be replied and the actual replies of the administrators, the system can continuously optimize the automatic reply templates and algorithms to improve the accuracy and effectiveness of automatic replies.
[0082] In an embodiment of the present application, further, after step 203-3, it can also include: step 203-4, receiving a confirmation instruction for any reference reply information or a modified confirmation instruction for any reference reply information, and determining a target reply template for the comment data to be replied to based on the confirmed reference reply information; performing summary analysis based on the comment data to be replied to, and taking the summary of the comment data to be replied to as a new preset question, and taking the target reply template as a reply template for the new preset question, and adding it to the second preset automatic reply template set.
[0083] In this embodiment, when the administrator receives the reference reply information generated by the system, they can select one of the reference reply information as a reply, or modify a reference reply information and then confirm it. The system needs to receive these confirmation instructions or modified confirmation instructions. Once the administrator confirms a reference reply information (whether native or modified), the reply information is regarded as the target reply template of the comment data to be replied to for automatic reply. In addition, the confirmed comment data to be replied can also be summarized and analyzed to extract its key information or theme. Then, this summary will be added to the second preset automatic reply template set as a new preset question. At the same time, the reply template corresponding to the summary and confirmed by the administrator will also be added to the template set as a reply template for the new preset question. In the embodiment of the present application, through the administrator's confirmation or modification of the reference reply information, the system can learn a more accurate and appropriate reply method, thereby improving the quality of the automatic reply. The feedback of the administrator is an important source of system learning. By continuously receiving and processing these feedbacks, the system can gradually optimize its reply strategy to make it more in line with the expectations and needs of users. With the addition of new preset questions and corresponding reply templates, the system's automatic reply template set will continue to be enriched and improved, so that it can cover more user inquiries and question types. More accurate automatic replies and more comprehensive template sets mean that the system can better meet user needs and expectations, thereby improving user experience and satisfaction. In addition, as the system's learning ability improves and the automatic reply template set is enriched, the number of situations that administrators need to handle manually will gradually decrease, thereby reducing operating costs.
[0084] In an embodiment of the present application, optionally, before generating a prompt word based on the comment data to be replied, the target published content, and the reply attribute information corresponding to the target management account, the method also includes: obtaining historical published content of the target management account, historical comment data of the historical published content, and historical reply data of the historical comment data; generating reply attribute information of the target management account according to the historical published content, the historical comment data, and the historical reply data through a large language model, wherein the reply attribute information includes the reply language style of the target management account.
[0085] In this embodiment, before generating prompt words for the comment data to be replied, the system can analyze the reply attribute information of the target management account according to the historical data. Specifically, the system obtains the historical published content of the target management account. Then, the system collects the historical comment data corresponding to these historical published contents, which represent the user's feedback and opinions on the published content. At the same time, the system also obtains the historical reply data corresponding to the historical comment data, which are the replies of the target management account or administrator to the comments. The collected historical data (including published content, comment data and reply data) are further analyzed in depth using the large language model. The specific language style used by the target management account when replying to the comment is extracted, such as formal, friendly, humorous, etc. The generated reply attribute information (especially the reply language style) will serve as an important reference for the subsequent generation of prompt words. Of course, the reply attribute information can also be customized by the administrator. By considering the historical reply data and language style of the target management account, the embodiment of the present application can generate a reply that is more in line with the personality of the account and the user's expectations, thereby improving user satisfaction. Personalized replies help to enhance the emotional connection between users and target management accounts, making users feel valued and understood, thereby enhancing user experience. Based on the reference of historical data and reply attribute information, the system can generate appropriate reply content more quickly, reduce the need for manual intervention, and improve reply efficiency.
[0086] Step 204: Send a first prompt message to the management terminal of the target management account to prompt the management terminal to process the pending comment data in the suspended processing group.
[0087] In this embodiment, a prompt message may be sent to the management terminal responsible for managing the target account, informing the administrator that the pending comment data in the suspended processing group needs to be manually processed.
[0088] By applying the technical solution of this embodiment, the whole solution involves processing and analyzing the comment data of the target management account on the social media platform to realize the collection of automated replies and user feedback. Specifically, by calling the API of the social media platform, the comment data related to the target published content is obtained, the correlation between the comment data and the target published content is analyzed, and the comments with correlation lower than the preset threshold are deleted; the emotional tendency of each comment data is identified, including positive, negative and neutral, and the offensive comments in the comment data with negative emotional tendency are identified and processed, and the remaining comment data are divided into intelligent processing groups and suspended processing groups according to the emotional tendency; semantic analysis is performed on the comment data to be replied in the intelligent processing group, and based on the semantic analysis results, the target reply template is matched in the preset automatic reply template set, and an automatic reply is performed. In addition, when the template cannot be directly matched, the system generates reference reply information and sends it to the administrator for confirmation or modification. Further, after the administrator's confirmation or modification instructions are received, the system updates the preset automatic reply template set according to this information. The embodiments of the present application greatly reduce the workload of manually processing comment data and improve processing efficiency through automated processing and the use of preset templates. Timely and accurate automatic replies can enhance user participation and satisfaction, especially when dealing with common inquiries and problems. Through the feedback and confirmation of the administrator, the system can continuously optimize the automatic reply templates and strategies to improve the accuracy and relevance of the replies. With the addition of new preset questions and corresponding reply templates, the system can better respond to different types of user inquiries and questions. With the improvement of the system's automated processing capabilities and the reduction of manual processing requirements, the operating costs of the customer service team can be effectively reduced. In summary, this solution achieves efficient processing and analysis of comment data on social media platforms through a combination of automated processing and manual participation, improves user experience, reduces operating costs, and enhances the system's learning ability.
[0089] Further, as Figure 1 The specific implementation of the method, the embodiment of the present application provides a device for intelligent management of comments in social media, such as Figure 3 As shown, the device comprises:
[0090] A data acquisition module is used to acquire at least one comment data corresponding to the target published content of the target management account in the social media, and respectively identify the sentiment tendency of each of the comment data, wherein the sentiment tendency is one of positive, negative and neutral;
[0091] A grouping module, used to identify offensive comment data among the comment data with negative sentiment tendency, and use the remaining comment data as the comment data to be replied, and divide the comment data to be replied into an intelligent processing group and a postponed processing group according to the corresponding sentiment tendency;
[0092] An intelligent reply module, for performing semantic analysis on the comment data to be replied in the intelligent processing group, obtaining a target reply template matching the comment data to be replied from a preset automatic reply template set based on the semantic analysis result, and automatically replying to the comment data to be replied using the target reply template;
[0093] The sending module is used to send a first prompt message to the management terminal of the target management account to prompt the processing of the to-be-replied comment data in the suspended processing group.
[0094] Optionally, the grouping module is further used to:
[0095] Divide the comments to be replied with positive or neutral sentiment into intelligent processing groups;
[0096] The review data with negative sentiment tendency are divided into the postponed processing group.
[0097] Optionally, the intelligent reply module is further used to:
[0098] Identifying the comment intention of the comment data to be replied, wherein the comment intention is a statement or an inquiry;
[0099] For the comment data to be replied whose comment intention is statement, according to the emotional tendency corresponding to the comment data to be replied, a target reply template matching the comment data to be replied is obtained in a first preset automatic reply template set, wherein the preset automatic reply template set includes the first preset automatic reply template set, and the first preset automatic reply template set includes at least one predefined standard reply content for positive or neutral comment data;
[0100] For the comment data to be replied whose comment intention is to inquire, semantic recognition is performed on the comment data to be replied, and a query is made in the second preset automatic reply template set whether there is a preset question matching the semantic recognition result. If so, the reply template corresponding to the matching preset question is obtained in the second preset automatic reply template set as the target reply template for the comment data to be replied.
[0101] Optionally, the intelligent reply module is further used to:
[0102] If it does not exist, a prompt word is generated based on the comment data to be replied, the target published content and the reply attribute information corresponding to the target management account, and a large language model is called to generate at least one reference reply information corresponding to the prompt word, and a second prompt information is generated based on the comment data to be replied and the reference reply information and sent to the management terminal.
[0103] Optionally, the intelligent reply module is further used to:
[0104] Acquire the historical published content of the target management account, the historical comment data of the historical published content, and the historical reply data of the historical comment data;
[0105] The reply attribute information of the target management account is generated through a large language model according to the historical published content, the historical comment data and the historical reply data, wherein the reply attribute information includes the reply language style of the target management account.
[0106] Optionally, the intelligent reply module is further used to:
[0107] Receiving a confirmation instruction for any reference reply information or a modified confirmation instruction for any reference reply information, and determining a target reply template for the comment data to be replied to according to the confirmed reference reply information;
[0108] A summary analysis is performed based on the comment data to be replied, and the summary of the comment data to be replied is used as a new preset question, and the target reply template is used as a reply template for the new preset question, and added to the second preset automatic reply template set.
[0109] Optionally, the data acquisition module is further used to:
[0110] By calling the API of the social media platform, based on the target published content of the target management account in the social media, obtaining at least one comment data corresponding to the target published content;
[0111] The correlation between the comment data and the target published content is analyzed, and the comment data with a correlation lower than a preset threshold is deleted.
[0112] It should be noted that for other corresponding descriptions of the functional units involved in the intelligent management device for comments in social media provided in the embodiment of the present application, please refer to Figure 1 to Figure 2 The corresponding description in the method will not be repeated here.
[0113] The embodiment of the present application also provides a computer device, which can be a personal computer, a server, a network device, etc. The computer device includes a bus, a processor, a memory and a communication interface, and can also include an input and output interface and a display device. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store location information. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the steps in each method embodiment are implemented.
[0114] Those skilled in the art will appreciate that the structure of the above-mentioned computer device is only a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components, or combine certain components, or have a different arrangement of components.
[0115] In one embodiment, a computer-readable storage medium is provided. The computer-readable storage medium may be non-volatile or volatile, and stores a computer program thereon. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.
[0116] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.
[0117] 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.
[0118] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.
[0119] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0120] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. An intelligent management method for comments in social media, It is characterized in that The method comprises: Based on the target published content of the target management account in the social media, obtaining at least one comment data corresponding to the target published content, and identifying the sentiment tendency of each of the comment data respectively, wherein the sentiment tendency is one of positive, negative and neutral; Identify offensive comment data among the comment data with negative sentiment tendency, and use the remaining comment data as comment data to be replied, and divide the comment data to be replied into an intelligent processing group and a postponed processing group according to the corresponding sentiment tendency; For the comment data to be replied in the intelligent processing group, semantic analysis is performed on the comment data to be replied, and based on the semantic analysis result, a target reply template matching the comment data to be replied is obtained from a preset automatic reply template set, and the target reply template is used to automatically reply to the comment data to be replied; A first prompt message is sent to the management terminal of the target management account to prompt the processing of the to-be-replied comment data in the suspended processing group.
2. The method according to claim 1, It is characterized in that Dividing the to-be-replied comment data into an intelligent processing group and a postponed processing group according to corresponding sentiment tendencies, including: Divide the comments to be replied with positive or neutral sentiment into intelligent processing groups; The review data with negative sentiment tendency are divided into the postponed processing group.
3. The method according to claim 2, It is characterized in that Performing semantic analysis on the comment data to be replied to, and acquiring a target reply template matching the comment data to be replied to from a preset automatic reply template set based on the semantic analysis result, including: Identifying the comment intention of the comment data to be replied, wherein the comment intention is a statement or an inquiry; For the comment data to be replied whose comment intention is statement, according to the emotional tendency corresponding to the comment data to be replied, a target reply template matching the comment data to be replied is obtained in a first preset automatic reply template set, wherein the preset automatic reply template set includes the first preset automatic reply template set, and the first preset automatic reply template set includes at least one predefined standard reply content for positive or neutral comment data; For the comment data to be replied whose comment intention is to inquire, semantic recognition is performed on the comment data to be replied, and a query is made in the second preset automatic reply template set whether there is a preset question matching the semantic recognition result. If so, the reply template corresponding to the matching preset question is obtained in the second preset automatic reply template set as the target reply template for the comment data to be replied.
4. The method according to claim 3, It is characterized in that After querying in the second preset automatic reply template set whether there is a preset question matching the semantic recognition result, the method further includes: If it does not exist, a prompt word is generated based on the comment data to be replied, the target published content and the reply attribute information corresponding to the target management account, and a large language model is called to generate at least one reference reply information corresponding to the prompt word, and a second prompt information is generated based on the comment data to be replied and the reference reply information and sent to the management terminal.
5. The method according to claim 4, It is characterized in that Before generating prompt words based on the comment data to be replied, the target published content, and the reply attribute information corresponding to the target management account, the method further includes: Acquire the historical published content of the target management account, the historical comment data of the historical published content, and the historical reply data of the historical comment data; The reply attribute information of the target management account is generated through a large language model according to the historical published content, the historical comment data and the historical reply data, wherein the reply attribute information includes the reply language style of the target management account.
6. The method according to claim 4, It is characterized in that After generating second prompt information based on the comment data to be replied and the reference reply information and sending the second prompt information to the management terminal, the method further includes: Receiving a confirmation instruction for any reference reply information or a modified confirmation instruction for any reference reply information, and determining a target reply template for the comment data to be replied to according to the confirmed reference reply information; A summary analysis is performed based on the comment data to be replied, and the summary of the comment data to be replied is used as a new preset question, and the target reply template is used as a reply template for the new preset question, and added to the second preset automatic reply template set.
7. The method according to any one of claims 1 to 6, It is characterized in that Based on target published content of a target management account in social media, obtaining at least one comment data corresponding to the target published content includes: By calling the API of the social media platform, based on the target published content of the target management account in the social media, obtaining at least one comment data corresponding to the target published content; The correlation between the comment data and the target published content is analyzed, and the comment data with a correlation lower than a preset threshold is deleted.
8. An intelligent management device for comments in social media, It is characterized in that The device comprises: A data acquisition module is used to acquire at least one comment data corresponding to the target published content of the target management account in the social media, and respectively identify the sentiment tendency of each of the comment data, wherein the sentiment tendency is one of positive, negative and neutral; A grouping module, used to identify offensive comment data among the comment data with negative sentiment tendency, and use the remaining comment data as the comment data to be replied, and divide the comment data to be replied into an intelligent processing group and a postponed processing group according to the corresponding sentiment tendency; An intelligent reply module, for performing semantic analysis on the comment data to be replied in the intelligent processing group, obtaining a target reply template matching the comment data to be replied from a preset automatic reply template set based on the semantic analysis result, and automatically replying to the comment data to be replied using the target reply template; The sending module is used to send a first prompt message to the management terminal of the target management account to prompt the processing of the to-be-replied comment data in the suspended processing group.
9. A storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.
10. A computer device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, It is characterized in that When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.
Citation Information
Patent Citations
Emotion analysis method, device and storage medium for public opinion comment information
CN109271512A
Reply information recommendation method and electronic equipment
CN110263247A
Intelligent voice verbal skill service system and device
CN111131637A
Online diagnosis and treatment data processing method and device, computer equipment and storage medium
CN112037904A
Intelligent question and answer method, device, computer equipment and computer readable medium
CN113010632A
Cited By
Method and device for improving comment quality, equipment, medium and program product
CN119670703A
