Knowledge Graph-Based Public Opinion Information Generation Method
Through a knowledge graph-based method, the user activity and public opinion measurement values of each published network platform are analyzed, and the public opinion management tags are generated, which solves the problem of insufficient processing of platform characteristics in the generation of public opinion information, and realizes the depth and accurate analysis of public opinion data.
Patent Information
- Application Number
- CN202410512586.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-26
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-04-26
AI Technical Summary
The existing technology lacks characteristic processing for different platforms in the generation of public opinion information, resulting in incomplete understanding of the true situation of public opinion and user participation, which affects the in-depth analysis and evaluation of public opinion data.
Using a knowledge graph-based method, crawl the associated data of each published network platform through crawling technology, build product public opinion knowledge graphs, analyze user activity and public opinion measurement values, generate public opinion management tags, provide information prompts, and deeply explore the correlation of public opinion data. and trends.
The systematization and precise generation of public opinion data has been achieved, the efficiency and accuracy of public opinion data processing have been improved, and the degree of user participation and public opinion attention can be more accurately evaluated, and the depth and breadth of public opinion information management have been improved.
Smart Images

Figure CN118551094B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of public opinion information technology, and specifically to a method for generating public opinion information based on a knowledge graph. Background Art
[0002] With the rapid development of the Internet, public opinion monitoring and analysis has gradually become an important support tool for corporate development. However, the massive amount of network information and complex public opinion trends make public opinion monitoring and analysis face challenges. Therefore, public opinion information generation plays an increasingly important role in the field of public opinion monitoring and analysis, providing more intelligent and accurate public opinion data processing support for the improvement of various industries. In the generation of public opinion information, in-depth information mining and correlation analysis can be carried out to discover potential public opinion data and trends.
[0003] The prior art, such as the patent application with announcement number: CN111914087B, is a public opinion analysis method, including: collecting public opinion data, the public opinion data includes project publicity materials and public opinions under the project publicity materials, media reports on the project publicity materials, and public discussions on the project publicity materials; preprocessing the public opinion data to obtain structured public opinion information; generating public opinion objects based on the public opinion information, the public opinion objects include object identification, object category, spatiotemporal information, semantic information, emotional information, and relationship information; performing similarity matching between the public opinion objects and public opinion cases in the public opinion case library to obtain the most similar public opinion cases; and obtaining a public opinion control plan based on the analysis of the most similar public opinion cases.
[0004] Existing technology, such as the patent application with announcement number: CN112100474B, is a passenger service quality public opinion supervision system and method. The system mainly includes a public opinion monitoring scheme configuration module, a public opinion information overview module, a public opinion information aggregation module, a public opinion early warning module, a public opinion analysis module, a public opinion briefing module and a public opinion data monitoring module. This solution can obtain public opinion information from multiple public opinion information sources, and perform sentiment analysis on the obtained information text, classify the information into positive, neutral and negative, establish charts to visualize the public opinion information, generate and output public opinion information briefings, and at the same time, it can warn of sensitive public opinions in the public opinion information and push them to users. Users can directly issue work orders for certain public opinions.
[0005] Combining the above solutions, it is found that the current generation of public opinion information lacks data processing based on the characteristics of different platforms, resulting in an incomplete understanding of the actual public opinion situation and user engagement of products on different network platforms, resulting in a lack of deep understanding and in-depth analysis of public opinion data, thus affecting the comprehensive evaluation and effective processing of public opinion data. Summary of the invention
[0006] In view of the deficiencies in the prior art, the present invention provides a method for generating public opinion information based on a knowledge graph, which can effectively solve the problems involved in the above-mentioned background technology.
[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for generating public opinion information based on a knowledge graph, comprising a cloud computing center receiving a request for generating public opinion information, and counting the published network platforms of the product as each publishing network platform, and using crawler technology to capture the related data of each publishing network platform for analysis, so as to obtain a product public opinion management metering trigger value for each publishing network platform.
[0008] Collect public opinion information on products in the public domain of each publishing network platform, build a product public opinion knowledge graph for each publishing network platform, and analyze the product public opinion measurement value of each publishing network platform.
[0009] According to the product public opinion management measurement trigger value of each publishing network platform and the product public opinion measurement value of each publishing network platform, the pre-managed publishing network platform is screened to generate public opinion management labels and provide information prompts.
[0010] Furthermore, the associated data of each publishing network platform is captured by crawler technology for analysis. The specific process is: the associated data of each publishing network platform is captured by crawler technology, wherein the associated data of each publishing network platform includes the average monthly active number of users of each publishing network platform, the cumulative number of registered users, the average daily growth rate of new users, and the average single usage time of users.
[0011] The monthly active number of reference users and the number of reference registered users stored in the cloud database are extracted and processed to obtain the first user activity index of each publishing network platform.
[0012] The reference daily average new user growth rate and the reference user's average single usage time stored in the cloud database are extracted and processed to obtain the user second activity index of each publishing network platform.
[0013] According to the first active index of users of each publishing network platform and the second active index of users of each publishing network platform, an active representation value of each publishing network platform is obtained through integrated analysis.
[0014] Furthermore, the product public opinion management metering trigger value of each publishing network platform is obtained, and the specific process is: according to the active representation value of each publishing network platform, and compared with the product public opinion management metering trigger value corresponding to each set active representation value interval, the product public opinion management metering trigger value of each publishing network platform is obtained.
[0015] Furthermore, the product public opinion management metering trigger value of each publishing network platform represents a quantitative result obtained by analyzing the associated data of each publishing network platform for analyzing the user activity level of each publishing network platform, and serves as an analysis basis for screening and obtaining pre-managed publishing network platforms for generating public opinion management labels.
[0016] Furthermore, the public opinion information on the public domain of the product in each publishing network platform is collected, and the specific data acquisition process is: the product release time point is used as the starting time point of the monitoring period, and according to the current collection time point, the monitoring period is divided into several monitoring time periods, and the public opinion information on the public domain of the product in each publishing network platform in each monitoring time period is collected.
[0017] Based on the public opinion information of the product in the public domain of each publishing network platform during each monitoring time period, the public opinion information includes the cumulative number of public opinion comment entries, the cumulative number of product interface views, the cumulative number of product interface shares, and the average browsing time per person on the product interface.
[0018] The public opinion indicator factors corresponding to the single public opinion comment terms stored in the cloud database and the number of reference product interface browsing times are extracted, and the first product public opinion indicator index of each publishing network platform is obtained after processing.
[0019] The public opinion indicator factors of a single shared product interface and the reference average browsing time of the product interface stored in the cloud database are extracted, and the second indicator index of product public opinion on each publishing network platform is obtained after processing.
[0020] All public opinion comment terms in the public domain of each publishing network platform during the monitoring period are extracted, and the public opinion fitting text of each publishing network platform is obtained by fitting after preprocessing. The public opinion fitting text of each publishing network platform is tagged with parts of speech using the natural language processing toolkit in Python. The word frequency and total word frequency of negative parts of speech in the public opinion text of each publishing network platform are counted, and the public opinion fitting text of each publishing network platform is processed with the public opinion reference negative fitting text in the cloud database through a text matching model. The semantic similarity between the public opinion fitting text of each publishing network platform and the public opinion reference negative fitting text is extracted, and the product public opinion third indicator index of each publishing network platform is obtained after processing.
[0021] Furthermore, the analysis of the product public opinion measurement values of each publishing network platform specifically includes the following steps: based on the first indicator index of product public opinion of each publishing network platform, the second indicator index of product public opinion of each publishing network platform, and the third indicator index of product public opinion of each publishing network platform, an integrated analysis is performed to obtain the product public opinion measurement values of each publishing network platform.
[0022] Furthermore, the product public opinion measurement value of each publishing network platform represents the quantitative results obtained by analyzing the public opinion information of the product in the public domain of each publishing network platform, which is used to analyze the user's public opinion attention level and the negative degree of public opinion on the product, and serves as the analysis basis for screening the pre-management publishing network platform to generate public opinion management labels.
[0023] Furthermore, the screening obtains the pre-managed publishing network platforms for generating public opinion management labels and providing information prompts. The specific screening prompt process is: comparing the product public opinion management meter trigger value of each publishing network platform with the product public opinion meter value of each publishing network platform. If the product public opinion meter value of a publishing network platform is higher than or equal to the product public opinion management meter trigger value of the publishing network platform, then the publishing network platform is marked as a pre-managed publishing network platform. The screening obtains the pre-managed publishing network platforms for generating public opinion management labels and providing information prompts.
[0024] Furthermore, the specific analysis conditions of the activity characterization values of the publishing network platforms are as follows:
[0025] In the formula, ωi represents the active representation value of the i-th publishing network platform, represents the first user activity index of the i-th publishing network platform, represents the second active index of users of the i-th publishing network platform, γ1 represents the weight factor corresponding to the set first active index of users, γ2 represents the weight factor corresponding to the set second active index of users, e represents a natural constant, i represents the number of each publishing network platform, i=1,2,3,...,n, and n represents the number of publishing network platforms.
[0026] Furthermore, the specific analysis conditions for the product public opinion measurement values of each publishing network platform are as follows:
[0027]
[0028] Where Di represents the product public opinion measurement value of the i-th publishing network platform, represents the first indicator index of product public opinion on the ith publishing network platform, represents the second indicator index of product public opinion on the ith publishing network platform, represents the third indicator index of product public opinion of the i-th publishing network platform, γ3 represents the weight factor corresponding to the set first indicator index of product public opinion, γ4 represents the weight factor corresponding to the set second indicator index of product public opinion, and γ5 represents the weight factor corresponding to the set third indicator index of product public opinion. The present invention has the following beneficial effects:
[0029] (1) The present invention provides a method for generating public opinion information based on a knowledge graph. First, the method analyzes and obtains the product public opinion management measurement trigger value of each publishing network platform. Then, by constructing the product public opinion knowledge graph of each publishing network platform and analyzing the product public opinion measurement value of each publishing network platform, the development trend and influencing factors of the public opinion data can be more deeply understood. Finally, the pre-management publishing network platform is obtained to generate public opinion management labels and provide information prompts, so that the generation of public opinion information is more systematic and accurate, and the correlation and development trend of the product public opinion data of each publishing network platform can be deeply explored, thereby improving the processing efficiency of public opinion data.
[0030] (2) The present invention obtains the active characterization value of each publishing network platform through analysis, so as to quantitatively evaluate the user participation, which helps to more accurately grasp the user behavior characteristics and activity of different platforms, provides reliable basic data support for the generation of public opinion information, and improves the efficiency and accuracy of public opinion data management.
[0031] (3) The present invention obtains the product public opinion measurement values of each publishing network platform through analysis, and makes full use of the public opinion information of the products of each publishing network platform in the public domain of each publishing network platform, providing a more accurate and in-depth basis for the generation and analysis of public opinion information. It can comprehensively evaluate the degree of user interaction and user emotional tendencies based on public opinion information, provide data support for public opinion information management, and improve the depth, breadth and accuracy of public opinion data analysis.
[0032] (4) The present invention obtains the pre-management publishing network platform through screening to generate public opinion management tags and provide information prompts, which helps to improve the efficiency and practicality of public opinion information generation and can perform corresponding processing according to the public opinion analysis results in a more targeted manner.
[0033] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 It is a schematic diagram of the method flow of the present invention.
[0035] Figure 2 This is a knowledge graph of product public opinion data involved in an embodiment of the present invention. DETAILED DESCRIPTION
[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0037] In the description of the present invention, it should be understood that the terms "opening", "upper", "lower", "thickness", "top", "middle", "length", "inside", "all around" and the like indicating orientation or positional relationship are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the components or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.
[0038] See also Figure 1 As shown, an embodiment of the present invention provides a technical solution for a public opinion information generation method based on a knowledge graph: a public opinion information generation method based on a knowledge graph, comprising a cloud computing center receiving a public opinion information generation request, and counting each published network platform of the product and marking it as each publishing network platform, and using crawler technology to capture the related data of each publishing network platform for analysis, so as to obtain a product public opinion management metering trigger value for each publishing network platform.
[0039] Collect public opinion information on products in the public domain of each publishing network platform, build a product public opinion knowledge graph for each publishing network platform, and analyze the product public opinion measurement value of each publishing network platform.
[0040] According to the product public opinion management measurement trigger value of each publishing network platform and the product public opinion measurement value of each publishing network platform, the pre-managed publishing network platform is screened to generate public opinion management labels and provide information prompts.
[0041] In the embodiments, a cloud computing center refers to a location or facility that centrally manages and provides cloud computing services. A cloud computing center is usually equipped with a large number of servers, storage devices, and network devices, as well as related software and management tools, for hosting and running various cloud computing services and applications.
[0042] In the embodiment, the public opinion information generation request refers to a request made by a relevant person to the system to obtain public opinion information according to specific needs or goals.
[0043] In the embodiments, each publishing network platform refers to a website or application that publishes content or provides services on different network platforms. These platforms may include social media websites, video sharing platforms, blog platforms, e-commerce platforms, forums and communities, news portals, etc. Each publishing network platform has its own unique user groups, characteristics and usage habits, which are of great significance to product public opinion.
[0044] In the embodiment, crawler technology is a technology for automatically acquiring Internet information, usually by writing programs to simulate human browsing behavior, access web pages and extract required data.
[0045] In an embodiment, the generation of public opinion management labels is based on the product public opinion management metering trigger value of each publishing network platform, and each publishing network platform is evaluated. Based on the evaluation results, a corresponding public opinion management label is generated for each platform to mark its public opinion status, such as "high risk", "medium risk", "low risk", etc.
[0046] In the embodiment, the specific process of information prompting is to generate corresponding public opinion management tags for each platform and send early warning notifications to relevant responsible persons through emails, text messages, instant messaging tools, etc., to remind them to pay attention to and promptly handle the public opinion data of relevant network platforms.
[0047] In the implementation example, formulate corresponding public opinion response strategies and measures, including actively responding to user feedback, explaining product features and advantages, and fixing possible problems with the product.
[0048] Specifically, the related data of each publishing network platform is captured by crawler technology for analysis. The specific process is: the related data of each publishing network platform is captured by crawler technology, wherein the related data of each publishing network platform includes the average monthly active number of users of each publishing network platform, the cumulative number of registered users, the average daily growth rate of new users and the average single usage time of users.
[0049] The monthly active number of reference users and the number of reference registered users stored in the cloud database are extracted and processed to obtain the first user activity index of each publishing network platform.
[0050] It should be noted that the user first activity index of each publishing network platform is a quantitative result obtained by analyzing the average monthly active number of users and the cumulative number of registered users of each publishing network platform to measure the activity level of users of each publishing network platform and the user base of the platform. In the embodiment, the user first activity index of each publishing network platform can be obtained by comprehensive analysis based on the user's behavioral activities on the platform, such as publishing content, commenting, liking, sharing, etc. In this embodiment, it is obtained by the following calculation method, and the specific calculation method is as follows:
[0051]
[0052] In the formula, represents the first user activity index of the i-th publishing network platform, represents the average monthly active number of users on the ith publishing network platform, represents the cumulative number of registered users of the ith publishing network platform, ΔS1 represents the monthly reference active number of users, ΔS2 represents the minimum number of registered users, α1 represents the correction factor corresponding to the set monthly active number of users, α2 represents the correction factor corresponding to the set number of registered users, e represents a natural constant, i represents the number of each publishing network platform, i=1,2,3,...,n, and n represents the number of publishing network platforms.
[0053] The reference daily average new user growth rate and the reference user's average single usage time stored in the cloud database are extracted and processed to obtain the user second activity index of each publishing network platform.
[0054] It should be noted that the second active index of users of each publishing network platform is obtained by analyzing the data of the daily average growth rate of new users of each publishing network platform and the average single usage time of users, and is used to analyze the user growth trend of each publishing network platform and the quantitative results of the usage time of users in each publishing network platform. In the embodiment, the second active index of users of each publishing network platform can be obtained by analyzing the depth of interaction of users on the platform, including participating activities, joining communities or groups, and paying attention to topics. In this embodiment, it is obtained by the following calculation method, and the specific calculation method is as follows:
[0055]
[0056] In the formula, represents the second active index of users on the ith publishing network platform, represents the average daily growth rate of new users on the ith publishing network platform, It represents the average single usage time of users on the i-th publishing network platform, ΔB1 represents the reference daily average new user growth rate, ΔB2 represents the reference user's average single usage time, α3 represents the correction factor corresponding to the set daily average new user growth rate, and α4 represents the correction factor corresponding to the set user's average single usage time.
[0057] According to the first active index of users of each publishing network platform and the second active index of users of each publishing network platform, an active representation value of each publishing network platform is obtained through integrated analysis.
[0058] Specifically, the product public opinion management metering trigger value of each publishing network platform is obtained. The specific process is: according to the active representation value of each publishing network platform, and compared with the product public opinion management metering trigger value corresponding to each set active representation value interval, the product public opinion management metering trigger value of each publishing network platform is obtained.
[0059] Specifically, the product public opinion management metering trigger value of each publishing network platform represents the quantitative results obtained by analyzing the associated data of each publishing network platform, which is used to analyze the user activity level of each publishing network platform, and serves as the analysis basis for screening the pre-managed publishing network platforms to generate public opinion management labels.
[0060] Specifically, public opinion information on the public domain of the product in each publishing network platform is collected. The specific data acquisition process is: the product release time point is used as the starting time point of the monitoring period, and the monitoring period is divided according to the current collection time point, and the monitoring period is divided into several monitoring time periods, and the public opinion information on the public domain of the product in each publishing network platform in each monitoring time period is collected.
[0061] Based on the public opinion information of the product in the public domain of each publishing network platform during each monitoring time period, the public opinion information includes the cumulative number of public opinion comment entries, the cumulative number of product interface views, the cumulative number of product interface shares, and the average browsing time per person on the product interface.
[0062] The public opinion indicator factors corresponding to the single public opinion comment terms stored in the cloud database and the reference product interface browsing times are extracted, and the first product public opinion indicator index of each publishing network platform is obtained after processing.
[0063] It should be noted that the first indicator index of product public opinion on each publishing network platform is a quantitative result for analyzing the degree of user participation in public opinion on each publishing network platform obtained by performing data analysis and processing on the cumulative number of comment entries and the cumulative number of product interface views in the public domain of each publishing network platform in each monitoring time period. In the embodiment, not only social media analysis tools or indicators can be used to evaluate the popularity of public opinion information on social media, for example, social media indicators such as Klout scores and influence ratings can be used to evaluate the degree of attention and user participation of public opinion information on social media to products, and the first indicator index of product public opinion on each publishing network platform is obtained by analysis. In this embodiment, it is obtained by the following calculation method, and the specific calculation method is as follows:
[0064]
[0065] In the formula, represents the first indicator index of product public opinion on the ith publishing network platform, Li→k represents the cumulative number of public opinion comment items on the public domain of the product in the ith publishing network platform during the kth monitoring period, Qi→k represents the cumulative number of product interface views on the public domain of the product in the ith publishing network platform during the kth monitoring period, ΔQ represents the number of reference product interface views, β1 represents the public opinion indicator factor corresponding to a single public opinion comment item, β2 represents the corresponding correction factor for the set cumulative number of product interface views, k represents the number of each monitoring period, k=1,2,3,...,r, r represents the number of monitoring periods, and e represents a natural constant.
[0066] The public opinion indicator factors of a single shared product interface and the reference average browsing time of the product interface stored in the cloud database are extracted, and the second indicator index of product public opinion on each publishing network platform is obtained after processing.
[0067] It should be noted that the second indicator index of product public opinion on each publishing network platform is a quantitative result reflecting the public opinion dissemination effect of the product on each publishing network platform obtained by analyzing and processing the cumulative number of times the product interface of the product is shared in the public domain of each publishing network platform in each monitoring period and the average browsing time of the product interface. In the embodiment, not only the depth and quality of the content of the comments or discussions can be analyzed, such as whether it involves important characteristics, functions or influencing factors of the product, but also the second indicator index of product public opinion on each publishing network platform can be obtained by analysis. In this embodiment, it is obtained by the following calculation method, and the specific calculation method is as follows:
[0068]
[0069] In the formula, represents the second indicator index of product public opinion on the ith publishing network platform, Ci→k represents the cumulative sharing times of the product interface in the public domain of the product on the ith publishing network platform during the kth monitoring period, Ti→k represents the average browsing time per person of the product interface in the public domain of the product on the ith publishing network platform during the kth monitoring period, ΔT represents the reference average browsing time per person of the product interface, β4 represents the public opinion indicator factor of a single sharing of the product interface, and β5 represents the correction factor corresponding to the set average browsing time per person of the product interface.
[0070] All public opinion comment terms in the public domain of each publishing network platform during the monitoring period are extracted, and the public opinion fitting text of each publishing network platform is obtained by fitting after preprocessing. The public opinion fitting text of each publishing network platform is tagged with parts of speech using the natural language processing toolkit in Python. The word frequency and total word frequency of negative parts of speech in the public opinion text of each publishing network platform are counted, and the public opinion fitting text of each publishing network platform is processed with the public opinion reference negative fitting text in the cloud database through a text matching model. The semantic similarity between the public opinion fitting text of each publishing network platform and the public opinion reference negative fitting text is extracted, and the product public opinion third indicator index of each publishing network platform is obtained after processing.
[0071] It should be noted that the third indicator index of product public opinion on each publishing network platform is obtained by performing data analysis on the word frequency, total word frequency of negative parts of speech in the public opinion texts of each publishing network platform, and the semantic similarity between the public opinion fitting text and the public opinion reference negative fitting text, and the quantitative result used to analyze the negative trend degree of the public opinion comment terms of the product on each publishing network platform. In the embodiment, not only deep learning technology, such as recurrent neural network (RNN) or convolutional neural network (CNN), can be used to perform sentiment classification on the public opinion text, and then the probability or score of the negative sentiment category is calculated to analyze and obtain the third indicator index of product public opinion on each publishing network platform. In this embodiment, it is obtained by the following calculation method, and the specific calculation method is as follows:
[0072]
[0073] In the formula, represents the third indicator index of product public opinion on the ith publishing network platform, Hi represents the word frequency of negative parts of speech in the public opinion text of the ith publishing network platform, ΔHi represents the total word frequency in the public opinion text of the ith publishing network platform, Pi represents the semantic similarity between the public opinion fitting text of the ith publishing network platform and the public opinion reference negative fitting text, ΔP represents the set defined semantic similarity, ε1 represents the correction factor corresponding to the set word frequency of negative parts of speech, and ε2 represents the correction factor corresponding to the set semantic similarity.
[0074] In this implementation scheme, the frequency of negative parts of speech includes words expressing negative emotions such as "bad", "awful", "unsatisfactory", "no", "none", "no", "disgusting", "hate", "angry", "sad", "disappointed", etc. The higher the frequency of these words in the text, the more negative emotions are expressed in the text.
[0075] In this implementation scheme, the total word frequency in the public opinion texts of each publishing network platform refers to the number of times each word appears in the text. The number of times all words appear is added up to obtain the total word frequency of the public opinion text.
[0076] In this implementation, the natural language processing toolkits in Python refer to NLTK (Natural Language Tool kit) and spaCy. NLTK is a popular natural language processing library that provides various functions, including part-of-speech tagging, named entity recognition, syntactic analysis, etc. spaCy is another widely used natural language processing toolkit, which has higher performance and a simpler API, and supports functions such as part-of-speech tagging, named entity recognition, syntactic analysis, etc. Both toolkits are common choices for natural language processing tasks in Python.
[0077] like Figure 2 As shown, it is a knowledge graph of product public opinion data provided by an embodiment of the present invention, wherein the nodes of the knowledge graph include products, publishing network platforms 1 to n, and m public opinion comment terms for each publishing network platform, wherein n represents the number of publishing network platforms, and m represents the number of public opinion comment terms.
[0078] Specifically, we analyze the product public opinion measurement values of each publishing network platform. The specific analysis and acquisition process is as follows:
[0079] Based on the first indicator index of product public opinion of each publishing network platform, the second indicator index of product public opinion of each publishing network platform, and the third indicator index of product public opinion of each publishing network platform, an integrated analysis is performed to obtain the product public opinion measurement value of each publishing network platform.
[0080] Specifically, the product public opinion measurement value of each publishing network platform represents the quantitative results obtained by analyzing the public opinion information of the product in the public domain of each publishing network platform, which is used to analyze the degree of users' public opinion attention to the product and the degree of negative public opinion. It also serves as the analysis basis for screening the pre-management publishing network platform to generate public opinion management labels.
[0081] Specifically, the pre-managed publishing network platforms are screened out to generate public opinion management labels, and information prompts are given. The specific screening and prompting process is: compare the product public opinion management meter trigger value of each publishing network platform with the product public opinion meter value of each publishing network platform. If the product public opinion meter value of a publishing network platform is higher than or equal to the product public opinion management meter trigger value of the publishing network platform, then the publishing network platform is marked as a pre-managed publishing network platform. In this way, the pre-managed publishing network platforms are screened out to generate public opinion management labels, and information prompts are given.
[0082] Specifically, the activity representation value of each publishing network platform can be obtained not only by pattern recognition and analysis using historical data, but also by the following calculation method. The specific calculation method is as follows:
[0083]
[0084] In the formula, ωi represents the active representation value of the i-th publishing network platform, represents the first user activity index of the i-th publishing network platform, represents the second active index of users of the i-th publishing network platform, γ1 represents the weight factor corresponding to the set first active index of users, γ2 represents the weight factor corresponding to the set second active index of users, e represents a natural constant, i represents the number of each publishing network platform, i=1,2,3,...,n, and n represents the number of publishing network platforms.
[0085] Specifically, the product public opinion measurement values of each publishing network platform can be obtained not only by using historical data for pattern recognition and analysis, but also by the following calculation method. The specific calculation method is as follows:
[0086]
[0087] Where Di represents the product public opinion measurement value of the i-th publishing network platform, represents the first indicator index of product public opinion on the ith publishing network platform, represents the second indicator index of product public opinion on the ith publishing network platform, represents the third indicator index of product public opinion on the i-th publishing network platform, γ3 represents the weight factor corresponding to the set first indicator index of product public opinion, γ4 represents the weight factor corresponding to the set second indicator index of product public opinion, and γ5 represents the weight factor corresponding to the set third indicator index of product public opinion.
[0088] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0089] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.
Claims
1. The method for generating public opinion information based on knowledge graph is characterized by: include: The cloud computing center receives the request for generating public opinion information, and counts the published network platforms of the product and marks them as the published network platforms. It uses crawler technology to capture the related data of each published network platform for analysis, and obtains the product public opinion management measurement trigger value of each published network platform; Collect public opinion information on the products in the public domain of each publishing network platform, build a product public opinion knowledge graph of each publishing network platform, and analyze the product public opinion measurement value of each publishing network platform; According to the product public opinion management measurement trigger value of each publishing network platform and the product public opinion measurement value of each publishing network platform, the pre-managed publishing network platform is screened to generate public opinion management tags and provide information prompts; The crawler technology is used to capture the relevant data of each publishing network platform for analysis. The specific process is as follows: The related data of each publishing network platform are captured by crawler technology, including the average monthly active number of users, the cumulative number of registered users, the average daily growth rate of new users, and the average single usage time of users of each publishing network platform; Extract the reference user monthly active number and reference registered user number stored in the cloud database, and obtain the user first active index of each publishing network platform after processing; Extract the reference daily average new user growth rate and the reference user's average single usage time stored in the cloud database, and obtain the user second activity index of each publishing network platform after processing; According to the first active index of users of each publishing network platform and the second active index of users of each publishing network platform, the active representation value of each publishing network platform is obtained by integrated analysis; The specific process of obtaining the product public opinion management measurement trigger value of each publishing network platform is as follows: According to the active representation value of each publishing network platform, the product public opinion management measurement trigger value corresponding to each set active representation value interval is compared to obtain the product public opinion management measurement trigger value of each publishing network platform; The specific analysis conditions of the activity characterization values of each publishing network platform are as follows: In the formula, ωi represents the active representation value of the i-th publishing network platform, represents the first user activity index of the i-th publishing network platform, represents the second active index of users of the i-th publishing network platform, γ1 represents the weight factor corresponding to the set first active index of users, γ2 represents the weight factor corresponding to the set second active index of users, e represents a natural constant, i represents the number of each publishing network platform, i=1,2,3,...,n, and n represents the number of publishing network platforms.
2. The method for generating public opinion information based on knowledge graph according to claim 1, characterized in that: The product public opinion management metering trigger value of each publishing network platform represents the quantitative result obtained by analyzing the associated data of each publishing network platform, which is used to analyze the user activity level of each publishing network platform, and serves as the analysis basis for screening the pre-managed publishing network platform to generate public opinion management tags.
3. The method for generating public opinion information based on knowledge graph according to claim 1 is characterized in that: The public opinion information of the product in the public domain of each publishing network platform is collected, and the specific data acquisition process is as follows: The product release time point is taken as the starting time point of the monitoring cycle, and the monitoring cycle is divided according to the current collection time point, and the monitoring cycle is divided into several monitoring time periods, and the public opinion information of the product in the public domain of each release network platform in each monitoring time period is collected; Based on the public opinion information of the product in the public domain of each publishing network platform during each monitoring period, the public opinion information includes the cumulative number of public opinion comments, the cumulative number of product interface views, the cumulative number of product interface shares, and the average browsing time of the product interface; The public opinion indicator factors corresponding to the single public opinion comment terms stored in the cloud database and the number of reference product interface browsing times are extracted, and the first product public opinion indicator index of each publishing network platform is obtained after processing; The public opinion indicator factors of a single shared product interface and the average browsing time of the product interface stored in the cloud database are extracted, and the second indicator index of product public opinion on each publishing network platform is obtained after processing; All public opinion comment terms in the public domain of each publishing network platform during the monitoring period are extracted, and the public opinion fitting text of each publishing network platform is obtained by fitting after preprocessing. The public opinion fitting text of each publishing network platform is tagged with parts of speech using the natural language processing toolkit in Python. The word frequency and total word frequency of negative parts of speech in the public opinion text of each publishing network platform are counted, and the public opinion fitting text of each publishing network platform is processed with the public opinion reference negative fitting text in the cloud database through a text matching model. The semantic similarity between the public opinion fitting text of each publishing network platform and the public opinion reference negative fitting text is extracted, and the product public opinion third indicator index of each publishing network platform is obtained after processing.
4. The method for generating public opinion information based on knowledge graph according to claim 3 is characterized in that: The analysis of the product public opinion measurement values of each publishing network platform, the specific analysis and acquisition process is as follows: Based on the first indicator index of product public opinion of each publishing network platform, the second indicator index of product public opinion of each publishing network platform, and the third indicator index of product public opinion of each publishing network platform, an integrated analysis is performed to obtain the product public opinion measurement value of each publishing network platform.
5. The method for generating public opinion information based on knowledge graph according to claim 1, characterized in that: The product public opinion measurement value of each publishing network platform represents the quantitative results obtained by analyzing the public opinion information of the product in the public domain of each publishing network platform, which is used to analyze the user's public opinion attention level and the negative degree of public opinion on the product, and serves as the analysis basis for screening the pre-management publishing network platform to generate public opinion management labels.
6. The method for generating public opinion information based on knowledge graph according to claim 1, characterized in that: The screening results in the pre-management publishing network platform generating public opinion management tags and providing information prompts. The specific screening prompt process is as follows: The product public opinion management metering trigger values of each publishing network platform are compared with the product public opinion metering values of each publishing network platform. If the product public opinion metering value of a publishing network platform is higher than or equal to the product public opinion management metering trigger value of the publishing network platform, the publishing network platform will be marked as a pre-managed publishing network platform. The pre-managed publishing network platforms are screened out for the generation of public opinion management labels and information prompts are provided.
7. The method for generating public opinion information based on knowledge graph according to claim 3 is characterized in that: The specific analysis conditions for the product public opinion measurement values of each publishing network platform are as follows: Where Di represents the product public opinion measurement value of the i-th publishing network platform, represents the first indicator index of product public opinion on the ith publishing network platform, represents the second indicator index of product public opinion on the ith publishing network platform, represents the third indicator index of product public opinion on the i-th publishing network platform, γ3 represents the weight factor corresponding to the set first indicator index of product public opinion, γ4 represents the weight factor corresponding to the set second indicator index of product public opinion, and γ5 represents the weight factor corresponding to the set third indicator index of product public opinion.
