A social network sentiment evolution model fusing user influence and activity

Abstract

This invention relates to the field of social network analysis technology and proposes a social network sentiment evolution model that integrates user influence and activity. To address the shortcomings of existing technologies, such as inaccurate simulations and inflexible predictions in current network sentiment analysis and sentiment evolution processes, this invention provides a social network sentiment evolution model that integrates user influence and activity. The model analysis method involves: collecting text sentiment data from target users; collecting user influence data from each user; collecting the activity level of the target social network at time t; collecting the forgetting probability of each user at time t as a measure of sentiment evolution capability; establishing a dynamic model for the sentiment evolution of the social network; optimizing the evolution probability of the information propagation model based on text sentiment; and analyzing the evolution process of the dynamic model for the sentiment evolution of the social network based on the optimized information propagation model. This model is suitable for application in social network analysis.

Description

Technical Field

[0001] This relates to the field of social network analysis technology, specifically to a social network sentiment evolution model that integrates user influence and activity. Background Technology

[0002] In the field of social network analytics, research on social network sentiment analysis and sentiment evolution has become a hot topic in recent years. Social platforms such as Facebook, Twitter, and in China, Sina Weibo and WeChat, have become an indispensable part of the daily lives of hundreds of millions of users. Users post various emotional content on social platforms, and the evolution of this content has an increasingly significant impact on public opinion and sentiment. Existing research mainly focuses on sentiment analysis based on sentiment lexicons, the application of information dissemination models, and the assessment of the impact on user behavior.

[0003] Current sentiment analysis research commonly uses text-based sentiment lexicons to classify the sentiment of user-posted text. These methods help determine the sentiment polarity of text, such as those using sentiment lexicons like BosonNLP and SentiWordNet. On the other hand, sentiment propagation and evolution models are mostly based on the SIR (Susceptible-Infectious-Recovered) model to simulate the propagation of sentiment in social networks. However, existing technologies often simply equate the propagation of user sentiment with the spread of epidemics, ignoring the differences between individual users in social networks.

[0004] For example, some studies based on the SIR model only categorize users' states as susceptible, infected, and immune, without considering the differences in influence between individuals. Influence is a key factor in social networks; the magnitude of influence of different users can significantly affect the breadth and speed of emotion transmission. Meanwhile, changes in social network activity also have a significant impact on the emotion evolution process, but many existing studies have failed to fully quantify and utilize activity. Furthermore, the phenomenon of individual emotion forgetting and gradual decay during emotion transmission—that is, the process by which nodes gradually change from a positive or negative state to an immune state—is not fully reflected in existing models.

[0005] Therefore, the existing technology mainly has the following problems:

[0006] 1. The emotion propagation model failed to effectively integrate the differences in user influence, resulting in an inaccurate simulation of the evolution of real emotions.

[0007] 2. The model does not fully consider the impact of social network activity on emotion transmission, making it less flexible in predicting the evolution of emotions in dynamic social environments.

[0008] 3. The lack of modeling of the emotional forgetting process fails to reflect the natural decay mechanism in the evolution of emotions, reducing the realism of the simulation results. Summary of the Invention

[0009] To address the shortcomings of existing technologies, such as insufficient accuracy in simulating real-world emotional evolution, lack of flexibility in predicting emotional evolution in dynamic social environments, and inability to guarantee the authenticity of simulation results in current online sentiment analysis and emotion evolution analysis processes, the technical solution provided by this invention is as follows:

[0010] A social network sentiment evolution model that integrates user influence and activity, the model analysis method is as follows:

[0011] The steps to collect textual sentiment analysis from target users;

[0012] The steps to collect user influence data for each user within the target social network;

[0013] The steps to collect the activity level of the target social network at time t;

[0014] Collect the forgetting probability of each user at time t in the target social network as a step in the process of emotional evolution ability;

[0015] Steps in the dynamics model of emotion evolution in social networks;

[0016] The steps for collecting information from a pre-defined information propagation model;

[0017] The steps to optimize the evolution probability of the information dissemination model based on the sentiment of the text;

[0018] The steps involve analyzing the evolution process of the emotional evolution dynamics model of the social network based on the optimized information propagation model.

[0019] Furthermore, a preferred implementation method is provided, wherein the text sentiment of the target user specifically refers to the text sentiment of the target Weibo user, and the method of acquisition is as follows:

[0020] The steps to perform sentence and word segmentation on Weibo text to obtain a set of segmented words;

[0021] The steps involve iterating through the sentences in the word segmentation set and detecting sentiment words, degree adverbs, and negation words.

[0022] The steps to obtain the weight of each sentiment word, degree adverb, and negation word by comparing with the dictionary;

[0023] The steps for determining sentiment tendency based on the weights are as follows.

[0024] Furthermore, a preferred implementation method is provided, in which the user influence of each user is obtained based on the number of followers and fans of each user.

[0025] Furthermore, a preferred implementation is provided, in which the activity level of the target social network at time t is obtained based on the number of active nodes, negative nodes, and immune nodes in the target social network at time t.

[0026] Furthermore, a preferred implementation method is provided, which obtains the forgetting probability of each user at time t based on the forgetting mechanism and the user influence mechanism.

[0027] Furthermore, a preferred implementation method is provided, in which the information dissemination model is obtained based on the user influence of each user, the activity level of the target social network at time t, and the emotional evolution capability.

[0028] Based on the same inventive concept, this invention also provides an analysis device for a social network sentiment evolution model that integrates user influence and activity, comprising:

[0029] A module for collecting textual sentiment from target users;

[0030] A module for collecting user influence data from the target social network;

[0031] A module for collecting the activity level of the target social network at time t;

[0032] Collect the forgetting probability of each user at time t in the target social network as a module for emotional evolution ability;

[0033] Modules of the dynamics model of emotion evolution in social networks;

[0034] A module for collecting pre-defined information propagation models;

[0035] A module that optimizes the evolution probability of the information dissemination model based on the sentiment of the text;

[0036] This module analyzes the evolution process of the emotional evolution dynamics model of the social network based on the optimized information propagation model.

[0037] Based on the same inventive concept, the present invention also provides a computer storage medium for storing a computing program, wherein when the computer reads the computer program, the computer executes the method described thereon.

[0038] Based on the same inventive concept, the present invention also provides a computer, including a processor and a storage medium, wherein when the processor reads a computer program stored in the storage medium, the computer executes the method described thereon.

[0039] Based on the same inventive concept, the present invention also provides a computer program product, which, when executed, implements the method described.

[0040] Compared with the prior art, the advantages of the technical solution provided by the present invention are as follows:

[0041] By introducing a method for calculating user influence, the model can more accurately simulate the effect of individuals on the spread of emotions within social networks. Compared to existing models that only consider user status, this approach quantifies a user's influence based on the number of their followers and followers, thereby improving the accuracy of the emotional dissemination path and scope, and making the dissemination effect more consistent with the influence distribution in real-world social networks.

[0042] By incorporating the calculation of social network activity, this approach can more flexibly simulate the dynamic process of emotion propagation. In existing research, activity is often insufficiently considered; however, this approach enhances the dynamic responsiveness to the rate and scope of emotion propagation by introducing changes in the activity of different nodes within the social network. This approach allows the model to exhibit higher adaptability and predictive accuracy when dealing with complex social network environments.

[0043] By introducing the node forgetting probability, this scheme successfully simulates the natural decay mechanism in emotion evolution, making the transition of node states more closely resemble the emotion propagation process in real social networks. Compared to the single state transition in the traditional SIR model, this scheme uses the forgetting probability to simulate the gradual weakening of emotions in nodes, making the transformation process of positive and negative emotions smoother and more realistic, thereby effectively improving the rationality of the emotion evolution results.

[0044] Compared to the traditional SIR model, this SIR-IASF sentiment evolution model incorporates a comprehensive consideration of user influence, activity, and forgetting mechanisms, making the sentiment propagation process more complex and diverse. This improvement enables the model to more accurately reflect the patterns of sentiment propagation in social networks, especially in complex scenarios where multiple emotions coexist and node states are constantly changing, allowing for a better description of the evolutionary path and final state of sentiment.

[0045] It is suitable for application in social network analysis. Attached Figure Description

[0046] Figure 1 A flowchart for a social network sentiment evolution model that integrates user influence and activity;

[0047] Figure 2 This is a graph showing the probability of an infected node becoming an immune node. Detailed Implementation

[0048] To make the advantages and benefits of the technical solution provided by the present invention clearer, the technical solution provided by the present invention will now be described in further detail with reference to the accompanying drawings, specifically:

[0049] Implementation Method 1: This implementation method provides a social network sentiment evolution model that integrates user influence and activity. The model analysis method is as follows:

[0050] The steps to collect textual sentiment analysis from target users;

[0051] The steps to collect user influence data for each user within the target social network;

[0052] The steps to collect the activity level of the target social network at time t;

[0053] Collect the forgetting probability of each user at time t in the target social network as a step in the process of emotional evolution ability;

[0054] Steps in the dynamics model of emotion evolution in social networks;

[0055] The steps for collecting information from a pre-defined information propagation model;

[0056] The steps to optimize the evolution probability of the information dissemination model based on the sentiment of the text;

[0057] The steps involve analyzing the evolution process of the emotional evolution dynamics model of the social network based on the optimized information propagation model.

[0058] Specifically, including:

[0059] Step 1: Weibo Text Segmentation and Sentiment Extraction

[0060] First, the jieba Chinese word segmentation component is used to segment the microblog texts of social network users. By segmenting sentences into words, the text is divided into a vocabulary set. Then, a text-based sentiment evolution analysis method is used to extract the sentiment of the microblog users' texts, specifically detecting sentiment words, degree adverbs, and negation words, and calculating the sentiment polarity score and sentiment tendency for each microblog text based on a sentiment lexicon. The output of this step is the sentiment polarity score and sentiment tendency for each user's microblog text.

[0061] Step Two: User Influence Calculation

[0062] Next, the influence of each user in the social network is calculated based on an influence model. When creating the social network, each user is first assigned an initial influence value. Then, the user's actual influence is calculated based on the number of followers and followers. To prevent excessively large user influence from affecting the experimental results, a standardization method is used to adjust the influence. The output of this step is the standardized influence value for each user.

[0063] Step 3: Calculation of Social Network Activity

[0064] In this step, the activity level of the social network at a given moment is calculated based on the number of active, negative, and immune nodes. Social network activity reflects the activity of users within the network and is an important dynamic parameter for sentiment propagation. The output of this step is the social network activity value at that specific moment.

[0065] Step 4: Calculate the probability of node forgetting

[0066] Based on the forgetting mechanism and the user influence mechanism, this method calculates the forgetting probability of social network users at a specific moment. The forgetting probability reflects the natural decay process of user emotional evolution and is calculated based on the Ebbinghaus forgetting curve. The greater the user's influence, the slower the rate of change in their forgetting probability, meaning the longer the user influences surrounding nodes. The output of this step is the forgetting probability of each node at a specific moment.

[0067] Step 5: Calculation of Emotion Evolution Probability and Construction of Dynamic Model

[0068] By leveraging user influence, social network activity, and node forgetting probability, the emotional evolution probability of social network users is calculated, and an emotional evolution dynamics model is constructed. User node states are categorized as susceptible, positive, negative, and immune. The SIR-IASF emotional evolution analysis model is used to simulate the emotional propagation process until all user nodes become immune. The output of this step is the change in the emotional state of user nodes.

[0069] Step Six: Emotional Evolution Analysis

[0070] Finally, by traversing all nodes in the social network and calculating the sentiment evolution process of each node based on text sentiment analysis and the aforementioned dynamic model, the sentiment state of each node is updated, and it is determined whether its sentiment has evolved. This calculation process is repeated until all nodes become immune nodes, yielding the final result of the social network user sentiment evolution.

[0071] Implementation Method Two: This implementation method further defines the social network sentiment evolution model that integrates user influence and activity provided in Implementation Method One. Specifically, the text sentiment of the target user refers to the text sentiment of the target Weibo user, and the method for obtaining this sentiment is as follows:

[0072] The steps to perform sentence and word segmentation on Weibo text to obtain a set of segmented words;

[0073] The steps involve iterating through the sentences in the word segmentation set and detecting sentiment words, degree adverbs, and negation words.

[0074] The steps to obtain the weight of each sentiment word, degree adverb, and negation word by comparing with the dictionary;

[0075] The steps for determining sentiment tendency based on the weights are as follows.

[0076] Implementation Method 3: This implementation method is a further refinement of the social network sentiment evolution model that integrates user influence and activity provided in Implementation Method 1. The user influence of each user is obtained based on the number of followers and fans of each user.

[0077] Implementation Method Four: This implementation method further defines the social network sentiment evolution model that integrates user influence and activity provided in Implementation Method One. Based on the number of positive nodes, negative nodes, and immune nodes in the target social network at time t, the activity level of the target social network at time t is obtained.

[0078] Implementation Method 5: This implementation method is a further refinement of the social network sentiment evolution model that integrates user influence and activity provided in Implementation Method 1. Based on the forgetting mechanism and the user influence mechanism, the forgetting probability of each user at time t is obtained.

[0079] Implementation Method Six: This implementation method further defines the social network sentiment evolution model that integrates user influence and activity provided in Implementation Method One. The information dissemination model is obtained based on the user influence of each user, the activity of the target social network at time t, and the sentiment evolution capability.

[0080] Implementation Method Seven: This implementation method provides an analysis device for a social network sentiment evolution model that integrates user influence and activity, comprising:

[0081] A module for collecting textual sentiment from target users;

[0082] A module for collecting user influence data from the target social network;

[0083] A module for collecting the activity level of the target social network at time t;

[0084] Collect the forgetting probability of each user at time t in the target social network as a module for emotional evolution ability;

[0085] Modules of the dynamics model of emotion evolution in social networks;

[0086] A module for collecting pre-defined information propagation models;

[0087] A module that optimizes the evolution probability of the information dissemination model based on the sentiment of the text;

[0088] This module analyzes the evolution process of the emotional evolution dynamics model of the social network based on the optimized information propagation model.

[0089] Implementation Method 8: This implementation method provides a computer storage medium for storing a computing program. When the computer reads the computer program, the computer executes the method provided in Implementation Method 1.

[0090] Implementation Method Nine: This implementation method provides a computer, including a processor and a storage medium. When the processor reads a computer program stored in the storage medium, the computer executes the method provided in Implementation Method One.

[0091] Implementation Method 10: This implementation method provides a computer program product. As a computer program, when the computer program is executed, it implements the method provided in Implementation Method 1.

[0092] Implementation Method Eleven: Combination Figure 1-2 This embodiment describes the technical solution provided above in further detail through specific examples. Specifically:

[0093] Reference Figure 1 As shown, this embodiment proposes an example of a social network sentiment evolution model that integrates user influence and activity. The construction method includes the following steps:

[0094] Step 1: Use the jieba Chinese word segmentation component to segment the microblog text of social network users, and then use the text-based sentiment evolution analysis method to extract the sentiment of microblog users' text.

[0095] Step 2: Based on the influence model, each user has a different level of influence on the surrounding nodes. The user influence of each user on the same social network is calculated based on the number of followers and fans of a user.

[0096] Step 3: Calculate the social network activity popular(t) at time t by using the number of active nodes, negative nodes, and immune nodes in the social network at a certain time t.

[0097] Step 4: Based on the forgetting mechanism and the user influence mechanism, calculate the forgetting probability of social network users at time t, that is, the extent of their ability to adapt to emotional evolution;

[0098] Step 5: Utilize user influence, social network activity, and node forgetting probability to obtain the emotional evolution probability of social network users, and obtain the emotional evolution dynamic model of the social network.

[0099] Step 6: Use the SIR-IASF sentiment evolution analysis model to analyze the sentiment evolution process of users in the social network until all user nodes become immune nodes, and return the evolved social network user sentiment.

[0100] In step one, text-based sentiment evolution analysis is used to extract the sentiment of Weibo users' texts, specifically:

[0101] Step 11: Perform sentence and word segmentation on the Weibo text to obtain a collection of Items. i =[[Word 11 Word 12 Word 1n1 ],…,[Word k1 Word k2 Word knk ]], where i is the i-th Weibo user, n k k represents the number of words in each sentence, and k represents the number of sentences in this Weibo post.

[0102] Steps 1 and 2: Traverse the segmented sentences and detect sentiment words, degree adverbs, and negation words, etc.

[0103] Step 13: Compare the sentiment dictionary, degree adverb dictionary, and negation word dictionary to obtain the weight set {w1, w2, ..., w} for each sentiment word. n If the negative word precedes the sentiment word, its sentiment value should be multiplied by the sentiment coefficient -1 to obtain a new sentiment value;

[0104] Step 14: Compare with the sentiment dictionary to obtain the score for each sentiment word {e1, e2, ..., e...} n The sentiment score for each Weibo post is calculated using the following method:

[0105]

[0106] Step 15: If the sentiment score Ei is greater than 0, then the sentiment tendency is positive; if the sentiment score Ei is less than 0, then the sentiment score is negative. Return the sentiment tendency and sentiment score.

[0107] In step two, the influence of each user in the social network is calculated. When creating the social network, each user is initially assigned an influence value of 1 (Influence(i)). Then, the actual influence of each user is calculated. The specific calculation method is as follows:

[0108]

[0109] Where d is a damping coefficient, and experiments have shown that d=0.8 yields the best results. Influence(i) represents the influence of user i, Follower(i) is the number of followers of user i, and Followee(i) is the number of followers of user i. To avoid the calculated user influence being too large and affecting the experimental results, it needs to be standardized. The standardization calculation method is as follows:

[0110]

[0111] In social networks, the number of followers of a node is its in-degree, while the number of its followers is its out-degree.

[0112] In step three, the specific method for calculating the social network activity at time t is as follows:

[0113] people(t) = P(t) + N(t) + R(t)

[0114]

[0115] Where popular(t) represents the activity level of the social network at time t, and people(t) represents the number of people who have been infected by a certain emotion at a certain time.

[0116] In step four, after a user in a social network is infected by a certain emotion, they acquire the ability to evolve and spread that emotion, becoming either a positive node P or a negative node N. However, over time, their ability to evolve their emotion gradually weakens, meaning their ability to become an immune node gradually increases. To make this model more closely resemble the emotional evolution in the real world, the forgetting probability of nodes is introduced. Referring to the Ebbinghaus forgetting curve, its calculation method is as follows:

[0117]

[0118] Where, γ i (t) represents the probability of forgetting a positive node P or a negative node N, denoted by node i, at time t. This probability is the probability of becoming an immune node R and losing the ability to influence surrounding nodes. δ i This represents the rate of change of the forgetting probability. In social networks, the more followers a user has, the greater their influence, and the longer it should take for them to affect the emotions of those around them. Therefore, the rate of change of the forgetting probability should be relatively slower. Hence, δ... i The calculation method is as follows:

[0119]

[0120] Taking a node in a social network as an example, after it evolves from a susceptible node S into an active node P or a negative node N, its probability of forgetting changes to the probability of becoming an immune node R, as shown below. Figure 2 As shown.

[0121] Depend on Figure 2 It can be seen that when a susceptible node S first becomes an active node P or a negative node N, its ability to influence the emotions of surrounding susceptible nodes S is 0. As time goes by, the probability of it losing the ability to influence surrounding nodes, i.e., becoming an immune node R, gradually increases. And after a certain period of time, it will definitely lose its ability to influence, which is more consistent with the emotional evolution process in real life.

[0122] In step five, the SIR-IASF model for analyzing the evolution of social network sentiment is constructed using methods such as the SIR information propagation model, user influence, social network activity, and node forgetting probability. Specifically:

[0123] Step 51: Divide user node states into susceptible nodes S, active nodes P, negative nodes N, and immune nodes R. Susceptible nodes S are nodes not yet infected by a certain emotion, but can evolve into active or negative nodes at any time through infection by surrounding infected nodes. They also possess a certain emotional tendency obtained through sentiment analysis of their Weibo texts. Active nodes P are usually already infected and have the ability to infect surrounding susceptible nodes S, thus becoming immune nodes R through forgetting probability. Negative nodes N are already infected and have the ability to infect surrounding susceptible nodes S, becoming immune nodes S through forgetting probability. Immune nodes S are new nodes that appear after active nodes P or negative nodes N lose their immunity and no longer have the ability to infect surrounding susceptible nodes. Their evolutionary dynamics formula is:

[0124]

[0125] Where α1 is the probability that a susceptible node S is infected by an infected node P or N and evolves into an active node P, α2 is the probability that a susceptible node S is infected by an active node P or a negative node N and evolves into a negative node N, β1 is the probability that an active node P loses the ability to evolve emotionally and evolves into an immune node R, and β2 is the probability that a negative node N loses the ability to evolve emotionally and evolves into an immune node R. The number of nodes in the social network is constant, i.e., S(t) + P(t) + N(t) + R(t) = K.

[0126] Step 52: Calculate the probability of user emotional evolution based on user influence, social network activity, and emotional polarity score.

[0127]

[0128] Where Spreader(i) represents the followers of Weibo user i who are either positive nodes P or negative nodes N, Emotion(i) represents the sentiment polarity of the Weibo posts published by user i after sentiment analysis, where 1 represents positive polarity and -1 represents negative polarity. Popular(t) represents the activity level of the social network at time t, and α represents the parameters used to optimize the results.

[0129] Step 5.3: Next, calculate ω1 and ω2 based on the sentiment polarity score. The greater the user's sentiment polarity, the greater their influence on sentiment evolution, i.e., the larger ω1 is. Conversely, the smaller the influence of social network activity on sentiment evolution, the smaller ω2 is. The method for processing ω1 and ω2 using the sentiment polarity score is as follows:

[0130]

[0131] score(i) is the sentiment polarity score of social network user i, |score| max The value representing the highest sentiment polarity score among social network users, |score| min It is the value with the smallest emotional polarity score in social networks.

[0132] Step 54: Calculate the probabilities β1 and β2 of nodes evolving from positive nodes P and negative nodes N into immune nodes R and losing their ability to evolve emotionally based on the forgetting probability. The calculation method is as follows:

[0133]

[0134] Here, γ is used to adjust the rate of change of the forgetting probability, and it is related to user influence. The greater the user influence, the slower the increase in the forgetting probability should be; therefore, it is inversely proportional to user influence.

[0135] According to its dynamic model, the change of its nodes is as follows: susceptible nodes S are gradually transformed into active nodes P and passive nodes N due to infection by surrounding active nodes P and passive nodes N, thus gradually decreasing until they reach 0. Since susceptible nodes S continuously transform into active nodes P, and then gradually transform into immune nodes R as the forgetting probability increases, the change of its nodes should be a gradual increase followed by a gradual decrease, eventually all becoming immune nodes S.

[0136] In step six, the SIR-based information propagation model is used to analyze the emotional evolution of social networks, specifically as follows:

[0137] Step 61: Traverse all nodes in the social network and perform text-based sentiment analysis on them to obtain their sentiment labels and sentiment polarity scores;

[0138] Step 62: Calculate the node influence and current social network activity of each node, then calculate its sentiment evolution probability and determine whether its sentiment has evolved.

[0139] Step 63: Calculate the forgetting probability for each node, and determine whether the node state has changed based on the forgetting probability.

[0140] Step 64: Repeat steps 62 to 63 until all node states become immune nodes, thus obtaining the final social network user sentiment.

[0141] The above description of several specific embodiments further details the technical solution provided by the present invention in order to highlight the advantages and benefits of the technical solution provided by the present invention. However, the above-described specific embodiments are not intended to limit the present invention. Any reasonable modifications and improvements to the present invention, combinations of embodiments, and equivalent substitutions based on the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A social network sentiment evolution model fusing user influence and activity, characterized in that, The model analysis method lies in: The steps to collect textual sentiment analysis from target users; The steps to collect user influence data for each user within the target social network; The steps to collect the activity level of the target social network at time t; Collect the forgetting probability of each user at time t in the target social network as a step in the process of emotional evolution ability; Steps in the dynamics model of emotion evolution in social networks; The steps for collecting information from a pre-defined information propagation model; The steps to optimize the evolution probability of the information dissemination model based on the sentiment of the text; The steps for analyzing the evolution process of the emotional evolution dynamics model of the social network based on the optimized information propagation model; Specifically: The text-based sentiment evolution analysis method is used to extract the sentiment of Weibo users’ texts. Specifically, sentiment words, degree adverbs and negation words are detected, and the sentiment polarity score and sentiment tendency of each Weibo text are calculated according to the sentiment dictionary. Each user's user influence is determined by the number of their followers and fans. The activity level of the target social network at time t is obtained based on the number of active nodes, negative nodes, and immune nodes in the target social network at time t. Based on the forgetting mechanism and the user influence mechanism, the forgetting probability of each user at time t is obtained; Based on text sentiment, optimize the evolution probability of the information dissemination model. 2.The social network sentiment evolution model fusing user influence and activity according to claim 1, wherein, The text sentiment of the target user specifically refers to the text sentiment of the target Weibo user, and the method of acquisition is as follows: The steps to perform sentence and word segmentation on Weibo text to obtain a set of segmented words; The steps involve iterating through the sentences in the word segmentation set and detecting sentiment words, degree adverbs, and negation words. The steps to obtain the weight of each sentiment word, degree adverb, and negation word by comparing with the dictionary; The steps for determining sentiment tendency based on the weights are as follows. 3.The social network sentiment evolution model fusing user influence and activity according to claim 1, wherein, The information dissemination model is derived based on the user influence of each user, the activity level of the target social network at time t, and the emotional evolution capability.

4. An analysis device of a social network sentiment evolution model fusing user influence and activity, characterized in that, include: A module for collecting textual sentiment from target users; A module for collecting user influence data from the target social network; A module for collecting the activity level of the target social network at time t; Collect the forgetting probability of each user at time t in the target social network as a module for emotional evolution ability; Modules of the dynamics model of emotion evolution in social networks; A module for collecting pre-defined information propagation models; A module that optimizes the evolution probability of the information dissemination model based on the sentiment of the text; This module analyzes the evolution process of the emotional evolution dynamics model of the social network based on the optimized information propagation model. Specifically: The text-based sentiment evolution analysis method is used to extract the sentiment of Weibo users’ texts. Specifically, sentiment words, degree adverbs and negation words are detected, and the sentiment polarity score and sentiment tendency of each Weibo text are calculated according to the sentiment dictionary. Each user's user influence is determined by the number of their followers and fans. The activity level of the target social network at time t is obtained based on the number of active nodes, negative nodes, and immune nodes in the target social network at time t. Based on the forgetting mechanism and the user influence mechanism, the forgetting probability of each user at time t is obtained; Based on text sentiment, optimize the evolution probability of the information dissemination model.

5. A computer storage medium for storing computing programs, characterized in that, When the computer reads the computer program, the computer executes the method of claim 1.

6. A computer, comprising a processor and a storage medium, characterized in that, When the processor reads the computer program stored in the storage medium, the computer executes the method of claim 1.

7. A computer program product, as a computer program, is characterized by: When the computer program is executed, it implements the method of claim 1.

Images