International propagation effect accurate evaluation method and system based on large language model

Abstract

The invention discloses an international propagation effect accurate evaluation method and system based on a large language model, and relates to the technical field of data processing, and the method comprises the steps: S1, reading media data, and carrying out the sensitive word recognition and data cleaning of the media data; s2, constructing a propagation diffusion model, acquiring the cleaned media data in the step S1, acquiring propagation path analysis value information through a graph neural network and time sequence analysis, identifying abnormal nodes and feeding back the abnormal nodes to the propagation diffusion model; s3, acquiring propagation behaviors of the sensitive words in the step S1, classifying the propagation behaviors, and calculating influence degrees of various propagation behaviors; s4, performing parallel analysis on the data obtained in the step S2 and the step S3, obtaining various impact factors to realize data fitting, and calculating a propagation effect value; and S5, inputting a result obtained in the step S4 into the Qwen large model, generating an evaluation report, and performing visual output. According to the method, multi-dimensional propagation effect evaluation is realized by combining a plurality of modules with evaluation indexes, and suggestions and predictions for optimizing the propagation effect are given by integrating the accurate analysis capability of a large model.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for accurately evaluating international communication effects based on a large language model. Background Art

[0002] With the rapid development of the Internet and social media platforms, the speed and scope of information dissemination have greatly increased. Especially in the context of globalization, cross-border communication and the formation of international public opinion have become more complicated. The evaluation of the effectiveness of international communication not only needs to consider traditional communication paths and audience characteristics, but also needs to incorporate a wider range of data analysis and social network analysis methods. The application of big data technology and artificial intelligence (AI) algorithms has provided unprecedented opportunities in this field, which can deeply explore the inherent laws of information dissemination, help predict dissemination effects and optimize dissemination strategies. In this context, a precise evaluation system for dissemination effects based on large models has emerged, providing a more sophisticated and intelligent evaluation method for global communication.

[0003] Traditional international communication effect evaluation methods mainly rely on statistical data and manual analysis. The evaluation process often focuses on the quantitative results of a single communication channel, such as evaluating the influence of communication content and audience coverage through the analysis of indicators such as media reprint volume and reading volume. However, these methods have some limitations. On the one hand, they often cannot accurately capture the entire process of information dissemination and lack dynamic tracking of communication paths and nodes; on the other hand, communication behavior analysis mostly relies on manual experience or simple classification models, ignoring the complex relationship between communication content and communication channels, and cannot comprehensively evaluate the multi-dimensional factors in the communication process. In addition, existing technologies lack efficient data cleaning and processing methods, and are not sophisticated enough for the integration and processing of massive data. Due to the lack of support for deep learning and intelligent analysis, the accuracy and real-time performance of traditional methods are poor, and they cannot cope with the rapidly changing international communication environment.

[0004] Although existing technologies can evaluate the effect of communication to a certain extent, their shortcomings are that they cannot comprehensively evaluate the multi-dimensional information of communication paths, communication nodes, and communication behaviors, and there is no effective mechanism to analyze abnormal nodes and changes in communication intensity during the communication process. Existing methods often simplify the communication process into static statistical data, and cannot dynamically track the communication paths and key behavior nodes during the communication process, resulting in poor timeliness and accuracy in adjusting communication strategies. In addition, traditional technologies also have major limitations in processing large-scale data, lacking in-depth semantic understanding and behavior recognition. Summary of the invention

[0005] In response to the problems of the prior art, the purpose of the present invention is to provide a method and system for accurately evaluating the international communication effects based on a large language model. Through data cleaning, multi-dimensional communication path analysis, accurate communication behavior classification and quantitative analysis, and real-time evaluation of communication effects, it aims to enhance the predictive ability of complex communication phenomena, thereby more comprehensively and accurately evaluating the communication effects and providing data support for the optimization of communication strategies. It has significant technical innovation and practical value.

[0006] In order to achieve the above objectives, the present invention provides the following technical solutions:

[0007] The present invention provides a method for accurately evaluating the effect of international communication based on a large language model, comprising:

[0008] S1. Read media data, and perform sensitive word recognition and data cleaning on the media data;

[0009] S2. Build a propagation diffusion model, obtain the media data cleaned in step S1, obtain information such as propagation path value through graph neural network and time series analysis, identify abnormal nodes and feed them back to the propagation diffusion model;

[0010] S3, obtaining and classifying the propagation behaviors of the sensitive words in step S1, and calculating the influence of each type of propagation behavior;

[0011] S4, analyzing the data obtained in step S2 and step S3 in parallel, obtaining various influencing factors to achieve data fitting, and determining the communication effect value;

[0012] S5. Input the results obtained in step S4 into the Qwen large model, generate an automated report, and perform visual output. The system gives predictions and suggestions.

[0013] The preferred solution, step S1, specifically comprises the following steps:

[0014] S11: Build a data collection module to obtain text data such as posts, news, and comments from foreign mainstream social media, news websites, and video websites to generate original data sets;

[0015] S12: Use the Qwen large model to identify sensitive words in the original data set obtained in step S11, and filter out words and sentences that are irrelevant to the analysis target;

[0016] S13: constructing a semantic model based on the vector relationship between sensitive words, and using vector decomposition technology to analyze the association relationship between the sensitive words in step S12;

[0017] S14: Use the Qwen big model to perform semantic analysis during data cleaning, remove noise information, merge similar sentences, and ensure the accuracy and quality of the data.

[0018] In step S12, compared with the traditional rule matching method, it can more accurately identify potential sensitive content according to the context, filter out words and sentences irrelevant to the analysis target, and thus refine the data content.

[0019] In step S13, a semantic model is constructed based on the vector relationship between sensitive words. After the identified sensitive words are processed by keyword recognition, multiple subject words are determined and weighted according to the importance coefficients of these subject words. The weights of these subject words reflect their relative importance in media communication and are comprehensively evaluated by multi-dimensional factors such as frequency of occurrence and user interaction.

[0020] The preferred solution, step S2, specifically comprises the following steps:

[0021] S21: The sensitive words and topic word feature elements weighted in step S1 are regarded as propagation nodes;

[0022] S22: Use the Qwen large model to process the text content of the communication nodes, convert these elements into time series data, and use graph neural network (GNN) and time series analysis to capture the communication relationship and communication intensity between different nodes, calculate the communication path value, and evaluate the overall effect of the communication activities, including the content’s communication breadth, depth, and interaction level;

[0023] S23: Further improve the propagation path analysis by analyzing and finding abnormal nodes, conduct detailed analysis of the node status, identify abnormal nodes and take corresponding corrective measures;

[0024] S24: Based on the propagation relationship, propagation strength, and propagation path value obtained in step S22 and step S23, a directed graph of the network relationship between each sensitive word node and each information is drawn, and then combined with the anomaly analysis function, a propagation diffusion model is constructed.

[0025] The preferred solution, step S3, specifically comprises the following steps:

[0026] S31: Based on the quantitative data during the dissemination process, the weighted sum of the number of different types of reposts is calculated each time a sensitive word appears, providing a quantitative indicator for the analysis of dissemination behavior and analyzing the impact of its behavioral characteristics on the dissemination effect;

[0027] S32: Combined with the deep learning capabilities of the Qwen large model, it is able to identify highly influential user nodes and their communication preferences through sentiment analysis and user behavior prediction, thereby analyzing the behavior patterns of different user groups to determine the communication behavior of sensitive words and classify them;

[0028] S33: According to the importance of the communication behavior in step S32, the behavior is divided into core behavior and sub-core behavior, and the influence of different communication behaviors on the communication effect of sensitive words is considered, and the influence of each type of communication behavior is calculated.

[0029] The preferred solution, step S4, specifically comprises the following steps:

[0030] S41: Determine the dissemination effect of sensitive words and make corrections to them. The average growth of the reprint volume of words with high sensitivity and low sensitivity is different. The reprint volume is corrected according to the sensitivity so that the dissemination effects of different sensitivities have the same dimension.

[0031] S42: fitting the reprint volume of sensitive words through a quadratic function, and using an exponential or logarithmic function to fit the reading volume and reprint volume;

[0032] S43: Further quantify the dissemination effect and calculate the correlation between the frequency of reprinting and the amount of reprinting and reading;

[0033] S44: The system combines all the corrected communication behaviors, communication paths, reprint frequencies, and relevances, and obtains the communication effect value through the fitting results;

[0034] When analyzing the communication results, if the communication results meet expectations, the next step of the output module will be entered; if the communication results do not meet expectations, it is necessary to construct an evaluation function to provide feedback and optimize the communication effect; the evaluation function can be comprehensively evaluated through factors such as the communication effect and influencing factor of sensitive words, and adjusted according to the feedback results.

[0035] The preferred solution, step S5, specifically comprises the following steps:

[0036] The Qwen big model is used to generate automated reports that explain the key factors, influence nodes and communication trends in the communication process. It is combined with the Tableau data visualization tool to generate interactive charts to help users understand the communication effect and support multi-dimensional analysis.

[0037] The present invention also provides an accurate evaluation system for international communication effects based on a large language model, comprising a data processing module, a communication path analysis module, a communication behavior analysis module, a communication effect evaluation module and an output module;

[0038] The data processing module is used to: read media data, identify sensitive words and clean the media data, and determine the cleaned media data;

[0039] The propagation path analysis module is used to: build a propagation diffusion model, obtain cleaned media data, obtain information such as propagation path values ​​through time series analysis, identify abnormal nodes and feed them back to the model;

[0040] The communication behavior analysis module is used to: determine and classify the communication behavior of sensitive words, and calculate the influence of various communication behaviors;

[0041] The communication effect evaluation module is used to: analyze the obtained data in parallel, obtain various influencing factors to achieve data fitting, and determine the communication effect value;

[0042] The output module is used to generate automated reports and provide visual output, and the system provides predictions and suggestions.

[0043] The innovation of this invention lies in: utilizing the deep learning and natural language processing capabilities of the Qwen large model to improve the intelligence level of the system, deeply explore the complex patterns in the communication data, and thus provide a more accurate and personalized communication effect evaluation.

[0044] Based on historical communication data, the present invention uses the Qwen large model to predict the effect. The system gives suggestions for optimizing the communication effect and can dynamically adjust the evaluation model to cope with real-time changes in the communication process. Through multi-dimensional analysis and evaluation of the communication process, the problems of inaccurate communication effect evaluation and incomplete analysis in the prior art are overcome, and the transparency and evaluation accuracy of the media communication process can be effectively improved.

[0045] The beneficial technical effects of the present invention are:

[0046] (1) By integrating the precise analysis capabilities of a large model, combined with multiple modules such as sensitive word identification, communication path analysis, communication behavior analysis, and communication effect evaluation, a comprehensive analysis of various factors in the communication process is conducted. By using time series analysis and data fitting technology, key nodes in the communication path can be more accurately identified, and the communication effect can be calculated, thereby providing more accurate communication effect evaluation results. This improvement in accuracy provides a scientific basis for the optimization of communication strategies and the adjustment of media content.

[0047] (2) Through the propagation path analysis module, the system can finely track the propagation trajectory of sensitive words, establish a propagation diffusion model based on time series, and deeply explore the inherent laws of propagation. The propagation behavior analysis module helps to accurately identify the key influencing factors, propagation behavior patterns and propagation channels in the propagation by classifying and quantitatively analyzing the propagation behaviors of each node in the propagation path, providing guidance for further optimizing the propagation strategy. This process ensures an in-depth understanding of the propagation path and behavior, avoiding the one-sidedness of relying solely on traditional statistical methods.

[0048] (3) The communication effect evaluation module combines the fitting algorithm of the reprint volume and the reading volume, and realizes the dynamic correction of the communication effect by analyzing the correlation between the communication frequency and the communication volume. Through real-time analysis and feedback, the system can adjust the communication strategy in time according to the evaluation results, making the communication process more efficient and accurate. This is of great significance for international communication, especially cross-cultural communication and cross-language communication, and can effectively enhance the global influence of the communication content.

[0049] (4) The output module combines the communication path and effect evaluation results to provide intuitive and clear visual output. This visualization not only makes it easier for users to understand the complex information in the communication process, but also makes the evaluation of communication effects more transparent and easy to monitor. Through clear communication path diagrams and effect score diagrams, decision makers can quickly grasp the implementation of communication strategies, make targeted adjustments, and thus improve communication efficiency.

[0050] (5) With the increasing complexity of the information dissemination environment, especially cross-border and cross-cultural international communication, it faces interference and challenges from multiple factors such as language, culture, and region. This solution helps users cope with the uncertainty and complexity in international communication by accurately identifying and analyzing sensitive words, dissemination channels, and node characteristics. The system can flexibly adjust the evaluation method of dissemination behavior, adapt to the dissemination characteristics of different countries and regions, improve the ability to evaluate the effect of cross-border communication, and enhance the ability to respond to global public opinion. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a flow chart of a method for accurately evaluating the effect of international communication based on a large language model according to the present invention;

[0052] Figure 2 A flowchart for sensitive word identification and keyword weighting in the data processing module;

[0053] Figure 3 Flowchart for propagation path analysis;

[0054] Figure 4 Flowchart for communication effectiveness evaluation. DETAILED DESCRIPTION

[0055] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0056] Example

[0057] like Figure 1 As shown in the figure, a precise evaluation method of international communication effect based on a large language model includes:

[0058] Step 1, data processing module: read media data, and perform sensitive word recognition and data cleaning on the media data;

[0059] S11: Collect text data such as posts, news and comments from foreign mainstream social media, news websites and video websites, build a data collection module, and generate original data sets;

[0060] In this solution, the main task of the data processing module is to extract key sensitive words and subject words from media data, and provide reliable data support for subsequent communication path analysis and effect evaluation through semantic analysis and cleaning operations; Figure 2 A flowchart for sensitive word identification and keyword weighting in the data processing module;

[0061] S12: First, the Qwen large model in the text recognition unit identifies sensitive words in the data set obtained in S11, filters out words and sentences that are irrelevant to the analysis target, and thus refines the data content; specifically, after the identified sensitive words are processed by keyword recognition, multiple subject words will be determined, and these subject words will be weighted according to their importance coefficients; the weights of these subject words reflect their relative importance in media communication, which may be comprehensively evaluated by multi-dimensional factors such as frequency of occurrence and user interaction;

[0062] S13: In the process of identifying sensitive words and subject words, the semantic analysis unit is responsible for analyzing the association between these sensitive words. To achieve this goal, vector decomposition technology is used to build a semantic model based on the vector relationship between sensitive words.

[0063] This process can be described by the following formula:

[0064] A=W*S

[0065] Among them, A is the set of sensitive word entities, W is the sensitive word vector conversion matrix, and S is the semantic representation matrix of sensitive words. Through this formula, the semantic information of sensitive words can be converted into vector representation, and the W matrix describes the semantic relationship between different sensitive words. Specifically, each row of the W matrix represents the semantic feature of a sensitive word, and the S matrix maps each sensitive word to a semantic space, so that the similarity between words can be calculated, and then their relevance can be determined.

[0066] S14: The cleaning unit uses the Qwen big model to perform semantic analysis during the data cleaning process to further optimize data quality. In particular, when processing data, it can remove redundant information and noise to ensure that subsequent analysis is based on clean and high-quality data.

[0067] Step 2, propagation path analysis module: Build a propagation diffusion model, obtain the media data cleaned in step 1, obtain information such as propagation path values ​​through graph neural network and time series analysis, identify abnormal nodes and feed them back to the model; Figure 3Flowchart for propagation path analysis;

[0068] S21: The weighted sensitive words and subject word feature elements in step S1 are regarded as propagation nodes; by converting these propagation nodes into time series data, the timeliness and evolution trend of the propagation process can be reflected;

[0069] S22: Use the Qwen large model to process the text content of the communication nodes, convert these elements into time series data, and use graph neural network (GNN) and time series analysis to capture the communication relationship and communication intensity between different nodes, calculate the communication path effect value, and evaluate the overall effect of the communication activities, including the content’s dissemination breadth, depth, and interaction level;

[0070] The propagation path analysis unit tracks the propagation process of information based on these time series data and evaluates the diffusion of information at each node; the propagation path analysis unit will further evaluate the path effect of the entire propagation process, which can be quantified by the following formula:

[0071] O=Q·A·M·N·I

[0072] Among them, O represents the communication path analysis value, Q is the number of releases, A is the number of disseminations on different media platforms, M is the total number of reprints, N is the number of readers of the initial release content, and I is the communication path value; through this formula, the overall effect of the communication activity can be comprehensively evaluated, including the breadth, depth and interaction level of the content; the communication path value I is further determined by the communication path analysis unit, and the specific calculation method is:

[0073] I=L·U

[0074] Among them, I represents the propagation path value, L is the propagation level, and U is the propagation intensity value; the propagation level L reflects the number of levels of information diffusion, while the propagation intensity U is closely related to factors such as the propagation rate and participation of information at each node; through the comprehensive evaluation of these two parameters, the potential and effectiveness of information dissemination can be further revealed;

[0075] For example, assuming that a certain media content is published 10 times Q, the number of times it is disseminated on multiple platforms A is 50 times, the total number of reposts M is 100 times, the initial reading volume N is 500 times, and the propagation path value I is 2, the propagation effect value P can be calculated using the above formula; if the level of the propagation path L is 3 and the propagation intensity U is 0.8, then the propagation path value I is:

[0076] I = 3 * 0.8 = 2.4

[0077] Combining this information, we can obtain the propagation path effect value P1, thus providing a quantitative basis for subsequent propagation effect evaluation and optimization;

[0078] S23: Further improve the propagation path analysis by analyzing and finding abnormal nodes, conduct detailed analysis of the node status, identify abnormal nodes and take corresponding corrective measures; in the propagation path analysis unit, one of the core steps to determine the propagation path is to conduct detailed analysis of the node status in order to identify abnormal nodes and take corresponding corrective measures; the effectiveness of the propagation path and the accurate evaluation of the propagation path effect depend on the behavior and propagation of each node; specifically, the status of each node reflects the performance of information in the propagation process, and the abnormality of these states may affect the entire propagation process; therefore, the abnormality degree of the node needs to be calculated and analyzed in order to detect problems early and intervene;

[0079] The abnormal degree of the node ρ j It can be calculated by the following formula:

[0080] ρ j =k ij *(F j -R j )*S j

[0081] Among them, ρ j represents the abnormality of the jth node, k ij is the number of propagations from the i-th node to the j-th node, F j is the frequency of occurrence of the jth node during the propagation, R j is the frequency value of the jth node in the propagation process, S j is the influencing factor of the sensitive word type corresponding to the jth node; by calculating the abnormality degree of the node, we can determine whether the node is abnormal; the higher the abnormality degree, the more likely there is a problem with the node in the propagation process, such as the propagation volume is lower than expected, the sensitive word propagation is insufficient, or the propagation path deviates from the expected direction;

[0082] For example, suppose that in a certain propagation process, the number of propagations of the jth node is k ij is 30 times, the frequency of occurrence of this node is F j is 0.6, there is a frequency R j is 0.5, the impact factor of sensitive word type S j is 1.2, then the abnormality degree of the node is ρ j will be:

[0083] ρ j =30*(0.6-0.5)*1.2=30*0.1*1.2=3.6

[0084] If the calculated abnormality value ρ jIf it exceeds the preset threshold, it indicates that the node may have a propagation problem and further analysis and correction of the abnormal cause is required;

[0085] The analysis and search of abnormal nodes is an important step to further improve the propagation path analysis. First, the degree of abnormality of each node needs to be calculated and compared with other nodes. If the degree of abnormality of a node is significantly higher than that of other nodes, the node can be marked as an abnormal node. Then, the causes of abnormal nodes need to be deeply explored. Possible causes include but are not limited to insufficient propagation of sensitive words, poor information transmission channels, unreasonable node attribute settings, etc. After finding the causes of abnormal nodes, the system needs to correct these abnormal nodes. For example, it may be necessary to increase the frequency of information propagation of the node, adjust the propagation relationship between nodes, or optimize the propagation content and sensitive word settings involved in the node.

[0086] In the preprocessing unit, the system will first determine the frequency value of keyword occurrence, the type of sensitive words, the propagation channel and propagation path; these data are crucial for the subsequent propagation path analysis; by collecting and organizing this information, time series data can be constructed to accurately show the propagation changes of sensitive words at different time nodes; for example, if the frequency of occurrence of a sensitive word changes over time within a day, the system can display it in a time series, showing the propagation status and path of sensitive words in different propagation channels; the display form of these data not only helps to visualize the propagation process, but also provides detailed historical data support for subsequent analysis;

[0087] For example, suppose the propagation path of sensitive words is as follows: in the first hour, the number of propagation paths of sensitive words is 3 times; in the second hour, the number of propagation paths increases to 5 times; in the third hour, the number of propagation paths is 7 times; by arranging these data in chronological order, the system can accurately track the dynamic changes of sensitive word propagation and help users analyze the propagation effect and path;

[0088] Through node status analysis, calculation of abnormality, adjustment of propagation paths, and preprocessing of propagation data, the system can effectively identify problems in the propagation path and optimize it. These analysis processes can not only find abnormal nodes, but also provide data support for the optimization of propagation strategies. Through these precise evaluations and adjustments, the propagation path can be ensured to be more in line with expectations, thereby improving the effect and accuracy of information dissemination.

[0089] S24: Draw a directed graph of the network relationship between each sensitive word node and each piece of information through the propagation relationship, propagation strength, and propagation path value obtained in S22 and S23, and then build a propagation diffusion model by combining the anomaly analysis function;

[0090] By drawing a network diagram, we can show nodes (such as individuals, media, platforms, etc.) and the connections (edges) between them, which is very suitable for representing complex communication networks; the size of the node represents the influence or communication volume of the node, and the color is used to distinguish different types of nodes or communication states; the thickness of the edge can represent the communication intensity or the frequency of information flow;

[0091] A directed graph can clearly indicate the direction of information flow, that is, from one node to another, indicating that information is transmitted from the source node to the target node; this graph can vividly display the path and direction of information transmission; coupled with the time series data converted by the Qwen big model, it can show the transmission of information on different nodes over time, and the big model will analyze and output the dynamic changes and trends of information transmission; finally, combined with the anomaly analysis function, the abnormal node data is fed back to the big model in real time to build a propagation and diffusion model;

[0092] Step 3, communication behavior analysis module: obtain the communication behavior of the sensitive words in step 1 and classify them, and calculate the influence of each type of communication behavior;

[0093] S31: Based on the quantitative data in the dissemination process, the weighted sum of the number of different types of reposts is calculated each time a sensitive word appears, providing a quantitative indicator for the analysis of dissemination behavior and analyzing the impact of its behavioral characteristics on the dissemination effect; in the dissemination behavior analysis module, the core task is to accurately analyze various behaviors in the dissemination path to determine and classify the dissemination behavior of sensitive words; dissemination behavior refers to how sensitive words or information spread and spread in the network at a specific time point and node, and how users interact with them; therefore, the analysis of dissemination behavior not only focuses on the quantitative data in the dissemination process, but also on the impact of its behavioral characteristics on the dissemination effect; the analysis of dissemination behavior first needs to determine the performance of each dissemination behavior through the following formula:

[0094]

[0095] Among them, β i Indicates the reprinting and dissemination effect of sensitive words, T ik represents the number of reposts of the sensitive word of type k in the number of reposts of the i-th occurrence, P ik represents the total number of reposts of the i-th occurrence of a sensitive word, and w represents the number of types in the dissemination behavior. Through this formula, the weighted sum of the number of reposts of different types each time a sensitive word appears can be calculated, thereby providing a quantitative indicator for the analysis of dissemination behavior.

[0096] For example, suppose the number of reposts of a sensitive word when it first appears is 1,000, and there are two types of reposts, type A with 600 reposts and type B with 400 reposts; for this dissemination behavior, the formula can be calculated as follows:

[0097] β i =(600*1)+(400*1)=1000

[0098] In this way, the reprinting and dissemination effect of sensitive words can be quantified, helping to analyze which types of dissemination behaviors have a greater impact on the dissemination effect;

[0099] S32: Combined with the deep learning capabilities of the Qwen large model, it is able to identify highly influential user nodes and their communication preferences through sentiment analysis and user behavior prediction, thereby analyzing the behavior patterns of different user groups to determine the communication behavior of sensitive words and classify them;

[0100] S33: Further, according to the importance of the communication behavior in S32, the behaviors are divided into core behaviors and sub-core behaviors; considering the influence of different communication behaviors on the communication effect of sensitive words, the influence of various communication behaviors is calculated; core behaviors refer to behaviors that have a significant impact on the communication effect, while sub-core behaviors have a relatively small impact on the communication effect; the purpose of behavior classification is to help identify the key driving factors in the communication process and optimize the communication strategy;

[0101] In the impact analysis unit, the analysis of communication behavior is not limited to the number and type of its occurrence, but also needs to further consider the degree of influence of different communication behaviors on the communication effect of sensitive words; the influence of communication behavior can be calculated by the following formula:

[0102]

[0103] Among them, U j represents the influence factor of the propagation behavior of the jth node, F j represents the propagation behavior characteristics of the jth node, R j represents the propagation behavior characteristics of the node, C j Indicates the number of times sensitive words appear, E j It represents the average increase in the number of reposts of sensitive words when they first appear. This formula can be used to quantify the impact of dissemination behaviors and analyze which behaviors have a greater impact on dissemination effects.

[0104] For example, if the propagation behavior characteristic F of the jth node j is 50, the propagation behavior characteristic R j is 30, the number of occurrences of sensitive words is C j The average growth rate of the first reprint is E j is 200 times, then the influence factor of the propagation behavior is U j It can be calculated as:

[0105]

[0106] This value shows that the propagation behavior of the jth node has a greater impact on the propagation effect of sensitive words, which is specifically reflected in the relationship between the change in propagation behavior characteristics and the growth of sensitive words;

[0107] Next, during the process of analyzing and calculating the communication behavior, the system also needs to analyze the behavior characteristics of sensitive words to further quantify the influence of the communication behavior; the calculation formula is as follows:

[0108]

[0109] Among them, α represents the quantitative impact factor of the effect of sensitive word dissemination behavior, F j represents the propagation behavior characteristics of the jth node, G j It represents the average increase in the number of reposts caused by highly sensitive communication behaviors. Through this formula, the system can comprehensively consider the communication behaviors of each node in the communication process and their contribution to the communication effect, and calculate the overall communication effect of sensitive words.

[0110] For example, suppose that sensitive words behave differently on multiple nodes during the propagation process. The propagation behavior characteristic of a certain node is 100, and the sensitivity of the node is high, and the increase in the number of reprints is 50; the propagation behavior characteristic of another node is 50, the sensitivity is low, and the increase in the number of reprints is 10; then, the propagation behavior effect of sensitive words is calculated as follows:

[0111] α=(100*5)+(50*10)=5000+500=5500

[0112] This shows that the propagation effect of sensitive words on high-sensitivity nodes is much greater than that on low-sensitivity nodes, thus providing a quantitative basis for evaluating the propagation effect;

[0113] Through these steps of communication behavior analysis, the system can comprehensively analyze the key behaviors in the communication path, the influence relationship between nodes, the classification of communication behaviors and their impact on the communication effect, so as to optimize the communication strategy and improve the accurate communication effect of information;

[0114] Step 4, communication effect evaluation module: analyze the data obtained in step 2 and step 3 in parallel, obtain various influencing factors to achieve data fitting, and determine the communication effect value;

[0115] S41: Determine the dissemination effect of sensitive words and make corrections to them; the average growth of the reprint volume of words with high sensitivity and low sensitivity is different, and the reprint volume is corrected according to the sensitivity so that the dissemination effects of different sensitivities have the same dimension;

[0116] Correction of sensitive words:

[0117] When evaluating the communication effect of sensitive words, the average growth of reprint volume is different for words with high sensitivity and words with low sensitivity. In the communication process, words with high sensitivity are usually more likely to gain a larger growth in reprint volume, while words with low sensitivity may face a smaller communication expansion. Therefore, it is necessary to correct the reprint volume according to the sensitivity so that the communication effects of different sensitivities have the same dimension.

[0118] The correction formula is:

[0119]

[0120] Among them, S j represents the behavior correction value of the jth node, R j represents the number of reposts of sensitive words corresponding to the jth node, T j represents the number of each type in the number of reprints corresponding to the jth node, Q j Represents the sensitivity coefficient of the node; through this formula, the system can adjust the propagation behavior correction value of different sensitivity levels, thereby ensuring that the propagation growth of high and low sensitivity has the same dimension;

[0121] For example, suppose a node has 100 sensitive words reprinted, of which type A accounts for 70 and type B accounts for 30. The sensitivity coefficient of the node is Q j is 1.5; then, the corrected behavior value can be calculated as:

[0122]

[0123] Thus, through this correction method, the propagation growth of type A with high sensitivity and type B with low sensitivity are unified and adjusted;

[0124] S42: fitting the reprint volume of sensitive words through a quadratic function, and using an exponential or logarithmic function to fit the reading volume and reprint volume;

[0125] The work focus of the communication effect evaluation module is to calculate the communication effect of sensitive words through fitting and correction. First, fit the reprint volume and behavior analysis results of sensitive words, and then correct and associate them with the reading volume, reprint frequency, etc.; by calculating the correlation between the reprint volume and the reading volume, an accurate communication effect value can be obtained;

[0126] Fitting of reprint volume and reading volume:

[0127] In the evaluation of communication effects, the reprint volume of sensitive words is fitted by quadratic functions to help identify their dissemination trends; for the fitting of reading volume and reprint volume, exponential or logarithmic functions are used. Such fitting methods can reflect the expansion speed and form of information in the dissemination process; Figure 4Flowchart for communication effectiveness evaluation;

[0128] To further quantify the communication effect, the correlation between the reprint frequency and the reprint volume and the reading volume is calculated; the calculation formula for the reprint frequency is:

[0129]

[0130] Among them, F represents the reprint frequency of sensitive words, A represents the reprint volume of sensitive words, and B represents the reading volume of sensitive words. Through this formula, the reprint frequency of each sensitive word in the process of dissemination can be calculated, and the dissemination effect can be further evaluated based on this frequency.

[0131] Calculation of the correlation of communication effect:

[0132] In order to further quantify the communication effect, we also need to calculate the correlation between the reprint frequency and the reprint volume and the reading volume. The formula is as follows:

[0133]

[0134] Among them, τ represents the correlation between the number of reprints and the number of readings, A represents the number of reprints, B represents the number of readings, and F represents the frequency of reprints. The function of this formula is to calculate the correlation between the number of reprints and the number of readings in order to determine the effectiveness of the communication effect.

[0135] For example, assuming that the number of reposts of a sensitive word is 5,000, the number of reads is 10,000, and the repost frequency F is 0.5; the relevance is calculated according to the formula:

[0136]

[0137] S43: Finally, the system combines all the corrected propagation paths, propagation behaviors, reprinting frequencies and relevances, and obtains the final propagation effect value through fitting results; when analyzing the propagation results, if the propagation results meet expectations, the next step of the output module is entered; if the propagation results do not meet expectations, an evaluation function needs to be constructed to provide feedback and optimize the propagation effect; the evaluation function can be comprehensively evaluated through factors such as the propagation effect and impact factor of sensitive words, and adjusted according to the feedback results;

[0138] In the communication effect analysis unit, the analysis and evaluation of the communication process is mainly achieved by combining the analysis of the communication path and communication behavior in S2 and S3, and then obtaining various impact factors; this process covers multiple steps, from determining the impact factor, the communication distance to the calculation of the communication intensity and the communication level, all of which directly affect the final communication effect evaluation; the impact factors include the subject word importance coefficient of sensitive words, the number of sensitive words cited, the frequency of occurrence of sensitive words, etc. These factors reflect the various influences on the information during the communication process;

[0139] Specifically, the impact factor can be determined by the following formula:

[0140]

[0141] Among them, σ represents the impact factor, D is the set of all impact factors, and w d is the topic importance coefficient of each sensitive word, c d is the number of citations of the subject word, U is the total number of occurrences of the sensitive word, K is the number of attribute types of the sensitive word, F is the number of times the sensitive word is published, and T is the number of channels in the dissemination channel; through this formula, the influence of multiple factors can be comprehensively considered to evaluate the role of each sensitive word in the dissemination process;

[0142] Next, the determination of the propagation distance is the key to evaluating the propagation scope and expansion effect; the propagation distance reflects the propagation coverage of information from the source to the final audience, which can be calculated by the following formula:

[0143] shi(μ=E·C

[0144] Among them, μ represents the dissemination distance, E is the dissemination effect value of different social media platforms, and C is the number of channels or paths involved in the dissemination process; through this formula, the depth and breadth of information dissemination between different platforms can be quantified;

[0145] Propagation intensity is another key evaluation indicator that determines the speed at which information spreads during the propagation process; the value of propagation intensity U can be determined by the following formula:

[0146]

[0147] in, is the communication intensity value, F is the number of times the sensitive word is released, and T is the number of channels in the communication channel; the calculation of communication intensity takes into account the frequency of information release and the diversity of communication channels, and can reflect the explosiveness and breadth of information dissemination;

[0148] The propagation level is an important indicator for measuring the level of information dissemination, which usually reflects the complexity of the path from the source to the target audience. The calculation method of the propagation level is based on the reprinting and publishing volume of sensitive words at each propagation node. According to the reprinting and publishing volume at each propagation level, the value L of the propagation level can be quantified by the following formula:

[0149]

[0150] Among them, L is the propagation level, R i is the number of reposts of the i-th level of communication, P iis the amount of information released at the i-th level of communication; this formula can further reveal the diffusion effect of information at different communication levels;

[0151] Finally, based on the comprehensive evaluation of propagation distance, propagation intensity and propagation level, the propagation effect value P can be calculated by the following formula:

[0152]

[0153] Among them, P is the propagation effect value, μ is the propagation distance, is the transmission intensity value, and L is the transmission level; this formula comprehensively considers the coverage, transmission rate and layer depth of information transmission, and provides a comprehensive quantitative indicator for the transmission effect;

[0154] When analyzing the communication results, first calculate the values ​​of various communication indicators through the above formula; if the communication results meet expectations, enter the next step of output; if the communication results do not meet expectations, it is necessary to construct an evaluation function to provide feedback and optimize the communication effect; the evaluation function can be comprehensively evaluated through factors such as the communication effect and impact factor of sensitive words, and adjusted according to the feedback results;

[0155] For example, suppose a sensitive word is released 100 times during its dissemination process, and the number of dissemination channels T is 5. The dissemination intensity value U is calculated as:

[0156] μ=100*5=500

[0157] The communication effect value E on different social media platforms is 1.2, the number of communication channels C is 5, and the communication distance D is calculated as:

[0158]

[0159] According to the amount of reprints and publications P, assuming that the amount of reprints under the first level of dissemination is 50 and the amount of publications is 80, the level of dissemination L is:

[0160] L=50*80=4000

[0161] Finally, the propagation effect value P is:

[0162] P = 6 * 500 * 4000 = 12,000,000

[0163] Through the above calculations, we can comprehensively evaluate the effect of the sensitive word in the entire communication process; if the communication result is lower than expected, we can adjust the communication strategy according to the feedback mechanism, optimize the communication path and strategy, and improve the communication effect;

[0164] The communication effect evaluation module can effectively quantify the communication effect by correcting and fitting the communication path, reprint volume, reading volume and behavioral characteristics of sensitive words. Through the above formulas and algorithms, the system can accurately evaluate the impact of communication behavior, provide feedback mechanisms, and adjust communication strategies to achieve the best communication effect in practical applications.

[0165] Step 5, output module: Input the results of step 4 into the Qwen big model to generate an automated report that explains the key factors, influence nodes and communication trends in the communication process. Combined with the Tableau data visualization tool, it generates interactive charts to help users understand the communication effect and support multi-dimensional analysis.

[0166] Through accurate communication effect evaluation, media organizations, brands and enterprises can fully understand the effectiveness of communication activities and identify the strengths and weaknesses in communication. This evaluation system can provide real-time feedback on the communication intensity, communication effect and content influence during the communication process, helping brands adjust content and optimize communication channels, ensuring that brand image and information can be effectively delivered to the target audience and improving brand communication efficiency.

[0167] The system provides data support for the optimization of communication strategies through multi-level communication behavior analysis, communication path tracking, and communication effect evaluation. Based on big data communication effect analysis and model prediction suggestions, the system can provide users with accurate communication optimization suggestions in real time, helping users to achieve accurate communication in a changing communication environment. By continuously adjusting and optimizing communication strategies, the success rate of information dissemination can be improved, and the overall communication effect can be enhanced.

[0168] Through a highly integrated communication effect evaluation system, a variety of advanced technologies such as large model technology, data processing, behavior analysis and communication evaluation are combined to analyze media data from all aspects and angles, accurately identify key factors and communication paths in communication, and improve the accuracy and operability of communication effects. This not only helps users effectively evaluate communication effects, but also provides them with real-time optimization strategies to improve the efficiency and influence of information dissemination on a global scale.

[0169] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A method for accurately evaluating the effect of international communication based on a large language model, characterized in that: include: S1. Read media data, and perform sensitive word recognition and data cleaning on the media data; S2. Construct a propagation diffusion model, obtain the media data cleaned in step S1, obtain propagation path analysis value information through graph neural network and time series analysis, identify abnormal nodes and feed them back to the propagation diffusion model; S3, obtaining and classifying the dissemination behaviors of the sensitive words in step S1, and calculating the influence of each type of dissemination behavior; S4, analyzing the data obtained in step S2 and step S3 in parallel, obtaining various influencing factors to achieve data fitting, and determining the communication effect value; S5. Input the results obtained in step S4 into the Qwen large model, generate an evaluation report, and perform visual output.

2. According to claim 1, a method for accurately evaluating the international communication effect based on a large language model is characterized in that: Step S1, the specific process is: S11: Build a data collection module to obtain text data such as posts, news, and comments from foreign mainstream social media, news websites, and video websites to generate original data sets; S12: Use the Qwen large model to identify sensitive words in the original data set obtained in step S11, and filter out words and sentences that are irrelevant to the analysis target; S13: constructing a semantic model based on the vector relationship between sensitive words, and using vector decomposition technology to analyze the association relationship between the sensitive words in step S12; S14: Use the Qwen big model to perform semantic analysis during data cleaning, remove noise information, merge similar sentences, and ensure the accuracy and quality of the data.

3. According to claim 1, a method for accurately evaluating the international communication effect based on a large language model is characterized in that: Step S2, the specific process is: S21: The sensitive words and topic word feature elements weighted in step S1 are regarded as propagation nodes; S22: Use the Qwen large model to process the text content of the communication nodes, convert these elements into time series data, and use graph neural network (GNN) and time series analysis to capture the communication relationship and communication intensity between different nodes, calculate the communication path value, and evaluate the overall effect of the communication activities, including the content’s communication breadth, depth, and interaction level; S23: Further improve the propagation path analysis by analyzing and finding abnormal nodes, conduct detailed analysis of the node status, identify abnormal nodes and take corresponding corrective measures; S24: Based on the propagation relationship, propagation strength, and propagation path value obtained in step S22 and step S23, a directed graph of the network relationship between each sensitive word node and each information is drawn, and then combined with the anomaly analysis function, a propagation diffusion model is constructed.

4. According to claim 3, a method for accurately evaluating the international communication effect based on a large language model is characterized in that: Step S22, the specific process is: The Qwen large model is used to process the text content of the communication nodes, and these elements are converted into time series data. Through graph neural network (GNN) and time series analysis, the communication relationship and communication intensity of information between different nodes are captured, the communication path analysis value is calculated, and the overall effect of the communication activities is evaluated, including the breadth, depth and interaction level of the content. The propagation path analysis unit tracks the propagation process of information based on these time series data and evaluates the diffusion of information at each node; the propagation path analysis unit will further evaluate the path effect of the entire propagation process, which can be quantified by the following formula: O=Q·A·M·N·I Among them, O represents the communication path analysis value, Q is the number of releases, A is the number of disseminations on different media platforms, M is the total number of reprints, N is the number of readers of the initial release content, and I is the communication path value; through this formula, the overall effect of the communication activity can be comprehensively evaluated, including the breadth, depth and interaction level of the content; the communication path value I is further determined by the communication path analysis unit, and the specific calculation method is: I=L·U Among them, I represents the propagation path value, L is the propagation level, and U is the propagation intensity value; the propagation level L reflects the number of levels of information diffusion, while the propagation intensity U is closely related to factors such as the propagation rate and participation of information at each node; through the comprehensive evaluation of these two parameters, the potential and effectiveness of information dissemination can be further revealed.

5. According to claim 3, a method for accurately evaluating the international communication effect based on a large language model is characterized in that: Step S23, the specific process is: The propagation path analysis is further improved by analyzing and finding abnormal nodes, analyzing the node status in detail, identifying abnormal nodes and taking corresponding corrective measures; in the propagation path analysis unit, one of the core steps to determine the propagation path is to analyze the node status in detail in order to identify abnormal nodes and take corresponding corrective measures; the effectiveness of the propagation path and the accurate evaluation of the propagation path effect depend on the behavior and propagation of each node; The state of each node reflects the performance of information during the propagation process. Abnormalities in these states may affect the entire propagation process. Therefore, the degree of abnormality of the node needs to be calculated and analyzed in order to detect problems early and intervene. The abnormal degree of the node ρ j It can be calculated by the following formula: r j =k ij *(F j -R j )*S j Among them, ρ j represents the abnormality of the jth node, k ij is the number of propagations from the i-th node to the j-th node, F j is the frequency of occurrence of the jth node during the propagation, R j is the frequency value of the jth node in the propagation process, S j is the influencing factor of the sensitive word type corresponding to the jth node; by calculating the abnormality degree of the node, we can determine whether the node is abnormal; the higher the abnormality degree, the more likely there is a problem with the node during the propagation process, such as the propagation volume is lower than expected, the sensitive word propagation is insufficient, or the propagation path deviates from the expected direction.

6. According to claim 1, a method for accurately evaluating the international communication effect based on a large language model is characterized in that: Step S3, the specific process is: S31: Based on the quantitative data during the dissemination process, the weighted sum of the number of different types of reposts is calculated each time a sensitive word appears, providing a quantitative indicator for the analysis of dissemination behavior and analyzing the impact of its behavioral characteristics on the dissemination effect; S32: Combined with the deep learning capabilities of the Qwen large model, it is able to identify highly influential user nodes and their communication preferences through sentiment analysis and user behavior prediction, thereby analyzing the behavior patterns of different user groups to determine the communication behavior of sensitive words and classify them; S33: According to the importance of the communication behavior in step S32, the behavior is divided into core behavior and sub-core behavior, and the influence of different communication behaviors on the communication effect of sensitive words is considered, and the influence of each type of communication behavior is calculated.

7. According to claim 6, a method for accurately evaluating the international communication effect based on a large language model is characterized in that: Step S3 is specifically: S31: Based on the quantitative data in the dissemination process, the weighted sum of the number of different types of reposts is calculated each time a sensitive word appears, providing a quantitative indicator for the analysis of dissemination behavior and analyzing the impact of its behavioral characteristics on the dissemination effect; in the dissemination behavior analysis module, the core task is to accurately analyze various behaviors in the dissemination path to determine and classify the dissemination behavior of sensitive words; dissemination behavior refers to how sensitive words or information spread and spread in the network at a specific time point and node, and how users interact with them; therefore, the analysis of dissemination behavior not only focuses on the quantitative data in the dissemination process, but also on the impact of its behavioral characteristics on the dissemination effect; the analysis of dissemination behavior first needs to determine the performance of each dissemination behavior through the following formula: Among them, β i Indicates the reprinting and dissemination effect of sensitive words, T ik represents the number of reposts of the sensitive word of type k in the number of reposts of the i-th occurrence, P ik represents the total number of reposts of the i-th occurrence of a sensitive word, and w represents the number of types in the dissemination behavior. Through this formula, the weighted sum of the number of reposts of different types each time a sensitive word appears can be calculated, thereby providing a quantitative indicator for the analysis of dissemination behavior. S32: Combined with the deep learning capabilities of the Qwen large model, it is able to identify highly influential user nodes and their communication preferences through sentiment analysis and user behavior prediction, thereby analyzing the behavior patterns of different user groups to determine the communication behavior of sensitive words and classify them; S33: Based on the importance of the communication behaviors in S32, the behaviors are divided into core behaviors and sub-core behaviors; considering the influence of different communication behaviors on the communication effect of sensitive words, the influence of various communication behaviors is calculated; core behaviors refer to behaviors that have a significant impact on the communication effect, while sub-core behaviors have a relatively small impact on the communication effect; the purpose of behavior classification is to help identify the key driving factors in the communication process and optimize the communication strategy; In the impact analysis unit, the analysis of communication behavior is not limited to the number and type of its occurrence, but also needs to further consider the degree of influence of different communication behaviors on the communication effect of sensitive words; the influence of communication behavior can be calculated by the following formula: Among them, U j represents the influence factor of the propagation behavior of the jth node, F j represents the propagation behavior characteristics of the jth node, R j represents the propagation behavior characteristics of the node, C j Indicates the number of times sensitive words appear, E j It represents the average increase in the number of reposts of sensitive words when they first appear. This formula can be used to quantify the impact of dissemination behaviors and analyze which behaviors have a greater impact on dissemination effects. Next, in the calculation process of communication behavior analysis, the system also needs to analyze based on the behavioral characteristics of sensitive words to further quantify the influence of communication behavior; the calculation formula is as follows: Among them, α represents the quantitative impact factor of the effect of sensitive word dissemination behavior, F j represents the propagation behavior characteristics of the jth node, G j It represents the average increase in the number of reposts caused by highly sensitive communication behaviors. Through this formula, the system can comprehensively consider the communication behaviors of each node in the communication process and their contribution to the communication effect, and calculate the overall communication effect of sensitive words. Through these steps of communication behavior analysis, the system can comprehensively analyze the key behaviors in the communication path, the influence relationship between nodes, the classification of communication behaviors and their impact on communication effects, thereby optimizing communication strategies and improving the accuracy of information dissemination.

8. According to claim 1, a method for accurately evaluating the international communication effect based on a large language model is characterized in that: Step S4, the specific process is: S41: Determine the dissemination effect of sensitive words and make corrections to them. The average growth of the reprint volume of words with high sensitivity and low sensitivity is different. The reprint volume is corrected according to the sensitivity so that the dissemination effects of different sensitivities have the same dimension. S42: fitting the reprint volume of sensitive words through a quadratic function, and using an exponential or logarithmic function to fit the reading volume and reprint volume; S43: Further quantify the dissemination effect and calculate the correlation between the frequency of reprinting and the amount of reprinting and reading; S44: The system combines all the corrected communication behaviors, communication paths, reprint frequencies, and relevances, and obtains the communication effect value through the fitting results; When analyzing the communication results, if the communication results meet expectations, the next step of the output module will be entered; if the communication results do not meet expectations, it is necessary to construct an evaluation function to provide feedback and optimize the communication effect; the evaluation function can be comprehensively evaluated through factors such as the communication effect and influencing factor of sensitive words, and adjusted according to the feedback results.

9. According to claim 8, a method for accurately evaluating the international communication effect based on a large language model is characterized in that: Step S4 is specifically: S41: Determine the dissemination effect of sensitive words and make corrections to them; the average growth of the reprint volume of words with high sensitivity and low sensitivity is different, and the reprint volume is corrected according to the sensitivity so that the dissemination effects of different sensitivities have the same dimension; When evaluating the communication effect of sensitive words, the average growth of reprint volume is different for words with high sensitivity and words with low sensitivity. In the communication process, words with high sensitivity are usually more likely to gain a larger growth in reprint volume, while words with low sensitivity may face a smaller communication expansion. Therefore, it is necessary to correct the reprint volume according to the sensitivity so that the communication effects of different sensitivities have the same dimension. The correction formula is: Among them, S j represents the behavior correction value of the jth node, R j represents the number of reposts of sensitive words corresponding to the jth node, T j represents the number of each type in the number of reprints corresponding to the jth node, Q j Represents the sensitivity coefficient of the node; through this formula, the system can adjust the propagation behavior correction value of different sensitivity levels, thereby ensuring that the propagation growth of high and low sensitivity has the same dimension; S42: fitting the reprint volume of sensitive words through a quadratic function, and using an exponential or logarithmic function to fit the reading volume and reprint volume; The work focus of the communication effect evaluation module is to calculate the communication effect of sensitive words through fitting and correction. First, fit the reprint volume and behavior analysis results of sensitive words, and then correct and associate them with the reading volume, reprint frequency, etc.; by calculating the correlation between the reprint volume and the reading volume, an accurate communication effect value can be obtained; In the evaluation of communication effects, the reprint volume of sensitive words is fitted by quadratic functions to help identify their dissemination trends; for the fitting of reading volume and reprint volume, exponential or logarithmic functions are used. Such fitting methods can reflect the expansion speed and form of information in the dissemination process; To further quantify the communication effect, the correlation between the reprint frequency and the reprint volume and the reading volume is calculated; the calculation formula for the reprint frequency is: Among them, F represents the reprint frequency of sensitive words, A represents the reprint volume of sensitive words, and B represents the reading volume of sensitive words. Through this formula, the reprint frequency of each sensitive word in the process of dissemination can be calculated, and the dissemination effect can be further evaluated based on this frequency. In order to further quantify the communication effect, we also need to calculate the correlation between the reprint frequency and the reprint volume and the reading volume. The formula is as follows: Among them, τ represents the correlation between the number of reprints and the number of readings, A represents the number of reprints, B represents the number of readings, and F represents the frequency of reprints. The function of this formula is to calculate the correlation between the number of reprints and the number of readings in order to determine the effectiveness of the communication effect. S43: Finally, the system combines all the corrected propagation paths, propagation behaviors, reprinting frequencies and relevances, and obtains the final propagation effect value through fitting results; when analyzing the propagation results, if the propagation results meet expectations, the next step of the output module is entered; if the propagation results do not meet expectations, an evaluation function needs to be constructed to provide feedback and optimize the propagation effect; the evaluation function can be comprehensively evaluated through factors such as the propagation effect and impact factor of sensitive words, and adjusted according to the feedback results; In the communication effect analysis unit, the analysis and evaluation of the communication process is mainly achieved by combining the analysis of the communication path and communication behavior in S2 and S3, and then obtaining various impact factors; this process covers multiple steps, from determining the impact factor, the communication distance to the calculation of the communication intensity and the communication level, all of which directly affect the final communication effect evaluation; the impact factors include the subject word importance coefficient of sensitive words, the number of sensitive words cited, the frequency of occurrence of sensitive words, etc. These factors reflect the various influences on the information during the communication process; The impact factor can be determined using the following formula: Where I represents the value of the impact factor, D is the set of all impact factors, and w d is the topic importance coefficient of each sensitive word, c d is the number of citations of the subject word, U is the total number of occurrences of the sensitive word, K is the number of attribute types of the sensitive word, F is the number of publications of the sensitive word, and T is the number of channels in the dissemination channel; through this formula, the influence of multiple factors can be comprehensively considered to evaluate the role of each sensitive word in the dissemination process; Next, the determination of the propagation distance is the key to evaluating the propagation scope and expansion effect; the propagation distance reflects the propagation coverage of information from the source to the final audience, which can be calculated by the following formula: D=E·C Among them, D represents the dissemination distance, E is the dissemination effect value of different social media platforms, and C is the number of channels or paths involved in the dissemination process; through this formula, the depth and breadth of information dissemination between different platforms can be quantified; Propagation intensity is another key evaluation indicator that determines the speed at which information spreads during the propagation process; the value of propagation intensity U can be determined by the following formula: U=F·T Among them, U is the communication intensity value, F is the number of times the sensitive word is released, and T is the number of channels in the communication channel; the calculation of communication intensity takes into account the frequency of information release and the diversity of communication channels, which can reflect the explosiveness and breadth of information dissemination; The propagation level is an important indicator for measuring the level of information dissemination, which usually reflects the complexity of the path from the source to the target audience. The calculation method of the propagation level is based on the reprinting and publishing volume of sensitive words at each propagation node. According to the reprinting and publishing volume at each propagation level, the value L of the propagation level can be quantified by the following formula: Among them, L is the propagation level, R i is the number of reposts of the i-th level of communication, P i is the amount of information released at the i-th level of communication; this formula can further reveal the diffusion effect of information at different communication levels; Finally, based on the comprehensive evaluation of propagation distance, propagation intensity and propagation level, the propagation path effect value P can be calculated by the following formula: P=D·U·L Among them, P is the propagation effect value, D is the propagation distance, U is the propagation intensity value, and L is the propagation level; this formula comprehensively considers the coverage, propagation rate and hierarchical depth of information dissemination, and provides a comprehensive quantitative indicator for the propagation effect; When analyzing the communication results, first calculate the various communication index values ​​​​through the above formula; if the communication results meet expectations, proceed to the next output link; if the communication results do not meet expectations, it is necessary to construct an evaluation function to provide feedback and optimize the communication effect; the evaluation function can be comprehensively evaluated through factors such as the communication effect and influencing factors of sensitive words, and adjusted according to the feedback results.

10. A precise evaluation system for international communication effects based on a large language model, characterized in that: It includes data processing module, communication path analysis module, communication behavior analysis module, communication effect evaluation module and output module.

Images