Quantitative analysis method for inversion degree of microblog network public opinions

Abstract

The invention discloses a micro-blog network public opinion inversion degree quantitative analysis method, which comprises the following steps of: 1) constructing a micro-blog social network model, and simulating a micro-blog social network by using a BA directed scale-free network; 2) initializing parameters, and respectively calculating self-positive intensity, relative influence and tissue atmosphere; 3) improving a classic bounded trust HK model, recalculating viewpoint values of N nodes, and performing continuous iterative updating; 4) determining an opinion leader according to the network centrality index, identifying the opinion leader by using the network centrality, and quantifying a public opinion inversion degree parameter; and 5) the influence of the opinion leader number and the organization belonging sense on the public opinion inversion degree is explored. Based on a directed scale-free network, the opinion leader influence, individual heterogeneity and organization belonging factors are combined, a classical opinion dynamics HK model is improved, network centrality is used for identifying opinion leader, public opinion reversal degree parameters are quantified, and the influence of the number of opinion leader and organization belonging on the public opinion reversal degree is researched.

Description

Technical Field

[0001] The present invention relates to the field of network public opinion research, and in particular to a method for quantitatively analyzing the reversal degree of microblog network public opinion. Background Art

[0002] At present, research on online public opinion reversal is relatively limited. On the one hand, most of the research focuses on the group polarization phenomenon of public opinion reversal by considering the online group as a whole. However, in actual situations, online public opinion events often present complex and changeable characteristics. Only taking homogeneous individuals as research objects may lead to one-sided conclusions and fail to fully and accurately reflect the overall situation.

[0003] On the other hand, most of the research is concentrated in the fields of communication and social psychology, mainly through qualitative analysis to explore the causes, characteristics and propagation mechanisms of public opinion reversal, and put forward governance suggestions. However, the research on using quantitative models to reveal the laws of public opinion reversal is still in its early stages, and it is relatively difficult to quantify the degree of public opinion reversal in existing technologies. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the prior art and provide a method for quantitatively analyzing the reversal degree of Weibo network public opinion. Through the construction of a directed scale-free network and the research on the reversal quantification of Weibo network public opinion, the problem of the difficulty in quantitatively analyzing the reversal degree of public opinion is effectively solved.

[0005] The object of the present invention is achieved by: a method for quantitatively analyzing the reversal degree of microblog network public opinion, comprising the following steps:

[0006] 1) Construct a Weibo social network model and simulate the Weibo social network with a BA directed scale-free network;

[0007] 2) Initialize parameters, randomly generate N initial opinion values ​​that follow a normal distribution, initialize the number of nodes, trust threshold, and organizational sense of belonging; calculate the strength of self-affirmation, relative influence, and organizational atmosphere respectively;

[0008] 3) Considering the individual's self-affirmation strength, relative influence and external organizational atmosphere factors, the classic bounded trust HK model is improved; and the opinion values ​​of N nodes are recalculated and continuously updated;

[0009] 4) Determine opinion leaders based on network centrality indicators, use network centrality to identify opinion leaders, and quantify the degree of public opinion reversal parameters;

[0010] 5) Explore the impact of the number of opinion leaders and organizational belonging on the degree of public opinion reversal.

[0011] As a further limitation of the present invention, the step 1) specifically comprises:

[0012] 1-1) Construct a BA directed scale-free network: randomly construct an initial network containing N nodes, and then gradually increase the number of nodes. Each additional node will establish a connection with the nodes in the existing network.

[0013] The probability P of the new node connecting to the original node in the network and the degree k of the original node i is positively correlated, expressed as:

[0014]

[0015] 1-2) The specific algorithm process of building a microblog social network is as follows: create a directed graph object G and add N nodes to the graph;

[0016] Initialize the network. For each node i from 1 to m, through two nested loops, establish bidirectional edges between node i and other nodes j to form a complete graph.

[0017] For each new node i from m+1 to N, calculate the out-degree probability of each node, that is, the ratio of the out-degree of each node to the sum of the out-degrees of all nodes;

[0018] Then use the randsample function to randomly select m nodes as neighbor nodes of the new node i according to the out-degree probability, then establish outbound edges between the new node i and the selected nodes, and establish inbound edges between the selected nodes and the new node i; finally, return the constructed BA directed scale-free network object G.

[0019] As a further limitation of the present invention, the calculation of the self-affirmation strength, relative influence and organizational atmosphere in step 2) specifically includes:

[0020] Self-affirmation strength ρ i , that is, the degree to which an individual adheres to his or her own views. The corresponding self-affirmation strength is calculated according to the degree of different nodes. That is, the larger the degree of the node, the more confident in one's own views, and the less likely one is to be influenced by the views of neighbors. The specific formula is as follows:

[0021]

[0022] Among them, k i is the out-degree of node i, k max is the maximum node degree in the network;

[0023] The relative influence function is F(i,j). In the process of opinion interaction, due to the heterogeneity of nodes, they will have different influences on neighboring nodes. Select node i in the network and define its influence relative to node j. The specific formula is as follows:

[0024]

[0025] Among them, ω(i) is the set of all neighbor nodes of node i, ∑ r∈ω(i) k r is the sum of the out-degrees of all neighboring nodes of node i; that is, the greater the out-degree of node i, the more fans it has, and the greater its influence; similarly, the relative influence of its neighbor node j is:

[0026]

[0027] The relative influence function F(i,j) of node i on node j is defined as:

[0028]

[0029] In general, F(i,j)≠F(j,i);

[0030] The organizational atmosphere is O i (t), the evolution of individual opinions is related to the organizational atmosphere to a certain extent, especially in social networks, where individuals are closely connected with organizations. Individual opinions will not only directly affect other individuals, but also form group pressure and indirectly affect the opinions of other individuals, ultimately making different opinions in the group tend to be consistent; the impact of organizational atmosphere on different individuals i at time t is:

[0031]

[0032] Among them, λ represents the sense of belonging to the organization, that is, the greater the degree to which individual i is influenced by the organizational atmosphere, the stronger the sense of belonging to the organization and the more easily the individual is influenced by the organizational atmosphere, so that the individual's views are consistent with those of the organization as a whole; Indicates the tendency of the mainstream opinion in the organization; when When , it means that most people in the organization hold positive views, that is, the organizational atmosphere has a positive impact on personal views; When , it means that most people hold negative views and the organizational atmosphere has a negative impact on individual views.

[0033] As a further limitation of the present invention, the improvement of the classic bounded trust HK model by considering the individual's self-affirmation strength, relative influence and external organizational atmosphere factors in step 3) specifically includes:

[0034] Taking into account the impact of changes in surrounding individual opinions and the combined influence of organizational atmosphere on individual opinions, the self-affirmation strength is introduced, so that each individual has different levels of confidence in his or her own opinions, thereby making the improved model more in line with the actual situation. The improved HK model is specifically expressed as follows:

[0035]

[0036] x i (t+1) represents the opinion value of individual i at time t+1; n represents the total number of individuals in the group; a trust threshold ε is set, and the opinion of individual j will only affect individual i if the difference between the opinion value of individual i and the opinion value of individual j is less than ε;

[0037] Among them, ρ i The strength of self-affirmation is the degree to which an individual adheres to his or her own views; i (t) is the organizational atmosphere, F(j,x) is the relative influence function of node j on node i, is the weight of the influence of individual i on individual j, is the number of neighbors of individual i at time t. Trust threshold ε∈(0,1). x j (t) is the opinion value of individual j at time t.

[0038] As a further limitation of the present invention, the step 4) specifically includes: according to the centrality index, finding the top M nodes in the centrality ranking in the constructed directed scale-free network and marking them as opinion leaders; when the algorithm executes to time t1, it indicates that the public opinion information has erupted for the first time, and the opinion of the opinion leader at this time is set to Opinion1; when the algorithm runs to time t2, it indicates the second spread caused by the disclosure of new information, and the opinion of the opinion leader is set to Opinion2; by looping step 3), the opinion values ​​of all nodes are continuously updated, and the evolution stops when the algorithm running time reaches the maximum value T, otherwise it returns to execute step 3).

[0039] As a further limitation of the present invention, the centrality index specifically includes degree centrality, betweenness centrality and closeness centrality;

[0040] The calculation formula of degree centrality is:

[0041]

[0042] Among them, m represents the number of other nodes connected to node i, and n represents the total number of nodes in the entire social network;

[0043] The calculation formula for betweenness centrality is:

[0044]

[0045] Among them, count i (j,k) represents the number of paths passing through node i in the shortest path from node j to node k, and count(j,k) represents the total number of all shortest paths from node j to node k;

[0046] Closeness centrality is defined as the inverse of the average shortest path length from a node to other nodes, specifically expressed as:

[0047]

[0048] Among them, d ij represents the average distance from node i to node j.

[0049] As a further limitation of the present invention, the degree of public opinion reversal in step 5) is defined as:

[0050]

[0051] Among them, x max represents the maximum value of the average opinion value during the process of public opinion dissemination, x min represents the minimum value of the average opinion value in the process of public opinion dissemination; since the research object is the public opinion reversal event, x max Generally positive, and x min To avoid the situation where the denominator is 0, the numerator and denominator in the formula are smoothed. p The larger the value is, the greater the tendency of public opinion reversal is, that is, the greater the possibility of reversal.

[0052] The present invention adopts the above technical solution, and compared with the prior art, the beneficial effects are as follows: the present invention constructs a BA directed scale-free network to simulate the Weibo social network environment, and determines opinion leaders based on the network centrality index, while considering factors such as individual self-affirmation strength, relative influence, and external organizational atmosphere to improve the classic bounded trust HK model, effectively solving the problem of difficulty in quantitative analysis of the degree of public opinion reversal. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 Flowchart of the present invention.

[0054] Figure 2 This is the node viewpoint evolution diagram of the present invention.

[0055] Figure 3 This is the group average opinion evolution diagram of the present invention.

[0056] Figure 4 This is a diagram showing the impact of the number of different opinion leaders on the evolution of the average opinion of the group according to the present invention.

[0057] Figure 5 This is a diagram showing the impact of organizational belonging on the evolution of individual opinions in the public opinion reversal of the present invention. DETAILED DESCRIPTION

[0058] like Figure 1 The method for quantitatively analyzing the reversal degree of microblog network public opinion shown in the figure comprises the following steps:

[0059] 1) Construct a Weibo social network model and simulate the Weibo social network with a BA directed scale-free network;

[0060] 1-1) Construct a BA directed scale-free network: randomly construct an initial network containing N nodes, and then gradually increase the number of nodes. Each additional node will establish a connection with the nodes in the existing network.

[0061] The probability P of the new node connecting to the original node in the network and the degree k of the original node i is positively correlated, expressed as:

[0062]

[0063] 1-2) The specific algorithm process of building a microblog social network is as follows: create a directed graph object G and add N nodes to the graph;

[0064] Initialize the network. For each node i from 1 to m, through two nested loops, establish bidirectional edges between node i and other nodes j to form a complete graph.

[0065] For each new node i from m+1 to N, calculate the out-degree probability of each node, that is, the ratio of the out-degree of each node to the sum of the out-degrees of all nodes;

[0066] Then use the randsample function to randomly select m nodes as neighbor nodes of the new node i according to the out-degree probability, then establish outbound edges between the new node i and the selected nodes, and establish inbound edges between the selected nodes and the new node i; finally, return the constructed BA directed scale-free network object G.

[0067] 2) Initialize parameters, randomly generate N initial opinion values ​​that follow a normal distribution, initialize the number of nodes, trust threshold, and organizational sense of belonging; calculate the strength of self-affirmation, relative influence, and organizational atmosphere respectively;

[0068] Calculating the strength of self-affirmation, relative influence and organizational atmosphere include:

[0069] Self-affirmation strength ρ i , that is, the degree to which an individual adheres to his or her own views. The corresponding self-affirmation strength is calculated according to the degree of different nodes. That is, the larger the degree of the node, the more confident in one's own views, and the less likely one is to be influenced by the views of neighbors. The specific formula is as follows:

[0070]

[0071] Among them, k i is the out-degree of node i, k max is the maximum node degree in the network;

[0072] The relative influence function is F(i,j). In the process of opinion interaction, due to the heterogeneity of nodes, they will have different influences on neighboring nodes. Select node i in the network and define its influence relative to node j. The specific formula is as follows:

[0073]

[0074] Among them, ω(i) is the set of all neighbor nodes of node i, ∑ r∈ω(i) k r is the sum of the out-degrees of all neighboring nodes of node i. That is, the greater the out-degree of node i, the more fans it has, and the greater its influence. Similarly, the relative influence of its neighbor node j can be obtained as:

[0075]

[0076] The relative influence function F(i,j) of node i on node j is defined as:

[0077]

[0078] In general, F(i,j)≠F(j,i).

[0079] The organizational atmosphere is O i (t), the evolution of individual opinions is related to the organizational atmosphere to a certain extent, especially in social networks, where individuals are closely connected with organizations. Individual opinions will not only directly affect other individuals, but also form group pressure and indirectly affect the opinions of other individuals, ultimately making different opinions in the group tend to be consistent; the impact of organizational atmosphere on different individuals i at time t is:

[0080]

[0081] Among them, λ represents the sense of belonging to the organization, that is, the greater the degree to which individual i is influenced by the organizational atmosphere, the stronger the sense of belonging to the organization and the more easily the individual is influenced by the organizational atmosphere, so that the individual's views are consistent with those of the organization as a whole; Indicates the tendency of the mainstream opinion in the organization; when When , it means that most people in the organization hold positive views, that is, the organizational atmosphere has a positive impact on personal views; When , it means that most people hold negative views and the organizational atmosphere has a negative impact on individual views.

[0082] 3) Considering the individual's self-affirmation strength, relative influence and external organizational atmosphere factors, the classic bounded trust HK model is improved; and the opinion values ​​of N nodes are recalculated and continuously updated;

[0083] Taking into account the influence of changes in surrounding individual opinions and the comprehensive influence of organizational atmosphere on individual opinions, the self-affirmation strength is introduced, so that each individual has different levels of confidence in his or her own opinions, which makes the improved model more in line with the actual situation and can better describe the complex relationships and opinion dissemination between individuals in the social network. The improved HK model is specifically expressed as follows:

[0084]

[0085] Among them, ρ i The strength of self-affirmation is the degree to which an individual adheres to his or her own views; i (t) is the organizational atmosphere, x i (t+1) represents the opinion value of individual i at time t+1. n represents the total number of individuals in the group. The trust threshold ε is set. The opinion of individual j will affect individual i only when the difference between the opinion value of individual i and the opinion value of individual j is less than ε.

[0086] F(j,x) is the relative influence function of node j on node i, is the weight of the influence of individual i on individual j, is the number of neighbors of individual i at time t. Trust threshold ε∈(0,1); x j (t) is the opinion value of individual j at time t.

[0087] 4) Determine opinion leaders based on network centrality indicators, use network centrality to identify opinion leaders, and quantify the degree of public opinion reversal parameters;

[0088] According to the centrality index, find the top M nodes in the centrality ranking of the constructed directed scale-free network and mark them as opinion leaders. When the algorithm is executed to time t1, it means that the public opinion information has exploded for the first time, and the opinion of the opinion leader at this time is set to Opinion1; when the algorithm runs to time t2, it means the second spread caused by the disclosure of new information, and the opinion of the opinion leader is set to Opinion2; through the loop step 3), the opinion values ​​of all nodes are continuously updated. When the algorithm running time reaches the maximum value T, the evolution stops, otherwise it returns to step 3);

[0089] Centrality measurement of Weibo social network,In social network, the indicators for evaluating key nodes mainly include three,centrality measures, namely degree centrality, betweenness centrality, and closeness centrality;,in directed networks, degree centrality can be divided into in-degree centrality and out-degree centrality,,which respectively represent the number of edges pointing to the node and,from the node to other nodes.

[0090] The calculation formula of degree centrality is:

[0091]

[0092] Among them, m represents the number of other nodes connected to node i, and n represents the total number of nodes in the entire social network;

[0093] The calculation formula for betweenness centrality is:

[0094]

[0095] Among them, count i (j,k) represents the number of paths passing through node i in the shortest path from node j to node k, and count(j,k) represents the total number of all shortest paths from node j to node k;

[0096] Closeness centrality is defined as the inverse of the average shortest path length from a node to other nodes, specifically expressed as:

[0097]

[0098] Among them, d ij represents the average distance from node i to node j.

[0099] 5) Explore the impact of the number of opinion leaders and organizational belonging on the degree of public opinion reversal;

[0100] The degree of public opinion reversal is defined as:

[0101]

[0102] Among them, x max represents the maximum value of the average opinion value during the process of public opinion dissemination, x min represents the minimum value of the average opinion value in the process of public opinion dissemination; since the research object is the public opinion reversal event, x max Generally positive, and x min To avoid the situation where the denominator is 0, the numerator and denominator in the formula are smoothed. p The larger the value is, the greater the tendency of public opinion reversal is, that is, the greater the possibility of reversal.

[0103] In order to more intuitively illustrate the technical solution used in the present invention, a specific embodiment is used here. The above improved model is simulated using MATLAB R2019a, and the specific parameter settings are shown in the following table. The model will continue to run until the simulation time limit T, and all viewpoint value results will be taken as the average value after 100 runs.

[0104]

[0105] Based on the parameter settings in the table above, the effect diagram obtained after running the simulation program is as follows Figure 2 and 3shown.

[0106] Figure 2 The evolution of individual opinions in the improved HK model is shown. When the public opinion outbreak event t1=10, the node opinion quickly approaches the opinion of the opinion leader (0.5), and the change process is smooth and the convergence speed is fast. When the public opinion reversal event t2=40, the node opinion quickly approaches the new opinion leader's opinion (-0.8), with a small fluctuation range and a fast convergence speed. Over time, the node's opinion quickly stabilizes, showing the efficient convergence and stability of the system. Figure 3 The figure shows the change of the average value of all individual opinions in the improved HK model over time. The improved model makes the change of the average opinion smoother, without large jumps, showing a more stable convergence characteristic. This is because the improved HK model can guide other individuals in the group due to the participation of opinion leaders, making it easier to converge opinions. The sense of belonging of individuals to the organization increases their enthusiasm for interacting with other members of the group, thus promoting consistency. The relative influence of individuals adjusts their ability to influence others in the group, making the model more realistic and flexible.

[0107] In different public opinion events, the number of opinion leaders will vary due to factors such as the sensitivity, impact, and harmfulness of the event. In this embodiment, other parameters are set unchanged, and the number of opinion leaders is set to 10, 30, 50, and 80 respectively. The model is run to obtain the simulation results of the average opinion evolution of the group. Figure 4 shown.

[0108] Figure 4 The simulation results of the model show that the number of opinion leaders has a significant impact on the individual opinion value in the process of public opinion reversal. The more opinion leaders there are, the greater the absolute value of the average opinion value of the group. When M = 10, the maximum average opinion value before the reversal is about 0.14, and the minimum is about -0.04 after the reversal, so the degree of public opinion reversal is 0.16; when M = 30, the maximum average opinion value before the reversal is about 0.18, and the minimum is about -0.07 after the reversal, so the degree of public opinion reversal is 0.23; when M = 50, the maximum average opinion value before the reversal is about 0.21, and the minimum is about -0.12 after the reversal, and the degree of public opinion reversal is 0.30; when M = 80, the maximum average opinion value before the reversal is 0.22, and the minimum is -0.13 after the reversal, and the degree of public opinion is about 0.32. In summary, the more opinion leaders participate in the reversal of public opinion, the greater their appeal in the dissemination of public opinion events, the more obvious the direction of guiding opinions, the stronger the change of opinion values ​​before and after, the greater the degree of public opinion reversal, and the greater the possibility of public opinion reversal.

[0109] Due to the existence of group pressure in the organization, the "spiral of silence" effect often occurs in the process of propagation and evolution of public opinion events, that is, when people express their own opinions, if they are consistent with the group's opinions, they will boldly express and spread them. If they are inconsistent with the group's opinions, they will remain silent or even change their opinions to cater to the group, resulting in one party's opinions becoming the mainstream and the other party's opinions gradually disappearing. This example explores the impact of this factor on the process of public opinion reversal. The organizational sense of belonging λ is set to 0.02, 0.04, 0.06, and 0.08 respectively. The results obtained by running the model are as follows: Figure 5 shown.

[0110] Figure 5 The simulation results show that when the value of organizational belonging λ increases, that is, the stronger the organizational atmosphere and the greater the group pressure, the individual's opinion is likely to change under the influence of group pressure or choose to remain silent in order to pursue consistency with the group. This phenomenon further increases the possibility of public opinion reversal. From the simulation results, it can be seen that when λ=0.02, the degree of public opinion reversal is about 0.10; when λ=0.04, the degree of public opinion reversal is about 0.15; when λ=0.06, the degree of public opinion reversal is about 0.16; and when λ=0.08, the degree of public opinion reversal is about 0.21. Therefore, it is necessary to pay attention to the impact of organizational atmosphere on groups and individuals, as well as the role of group pressure in public opinion reversal, so as to better understand and respond to the phenomenon of public opinion reversal and promote a good social interaction and public opinion environment.

[0111] The present invention is not limited to the above-mentioned embodiments. On the basis of the technical solution disclosed in the present invention, technicians in this field can make some substitutions and deformations to some technical features therein according to the disclosed technical content without creative labor, and these substitutions and deformations are all within the protection scope of the present invention.

Claims

1. A quantitative analysis method for the reversal degree of microblog network public opinion, characterized in that: The following steps are involved: 1) Construct a Weibo social network model and simulate the Weibo social network with a BA directed scale-free network; 2) Initialize parameters, randomly generate N initial opinion values ​​that follow a normal distribution, initialize the number of nodes, trust threshold, and organizational sense of belonging; Calculate the strength of self-affirmation, relative influence, and organizational climate separately; 3) Considering the individual's self-affirmation strength, relative influence and external organizational atmosphere factors, the classic bounded trust HK model is improved; and the opinion values ​​of N nodes are recalculated and continuously updated; 4) Determine opinion leaders based on network centrality indicators, use network centrality to identify opinion leaders, and quantify the degree of public opinion reversal parameters; 5) Explore the impact of the number of opinion leaders and organizational belonging on the degree of public opinion reversal.

2. According to claim 1, a method for quantitatively analyzing the reversal degree of microblog network public opinion is characterized in that: The step 1) specifically includes: 1-1) Construct a BA directed scale-free network: randomly construct an initial network containing N nodes, and then gradually increase the number of nodes. Each additional node will establish a connection with the nodes in the existing network. The probability P of the new node connecting to the original node in the network and the degree k of the original node i is positively correlated, expressed as: 1-2) The specific algorithm process of building a microblog social network is as follows: create a directed graph object G and add N nodes to the graph; Initialize the network. For each node i from 1 to m, through two nested loops, establish bidirectional edges between node i and other nodes j to form a complete graph. For each new node i from m+1 to N, calculate the out-degree probability of each node, that is, the ratio of the out-degree of each node to the sum of the out-degrees of all nodes; Then use the randsample function to randomly select m nodes as neighbor nodes of the new node i according to the out-degree probability, then establish outbound edges between the new node i and the selected nodes, and establish inbound edges between the selected nodes and the new node i; finally, return the constructed BA directed scale-free network object G.

3. The method for quantitatively analyzing the reversal degree of microblog network public opinion according to claim 1 is characterized in that: The calculation of self-affirmation strength, relative influence and organizational atmosphere described in step 2) specifically includes: Self-affirmation strength ρ i , that is, the degree to which an individual adheres to his or her own views. The corresponding self-affirmation strength is calculated according to the degree of different nodes. That is, the larger the degree of the node, the more confident in one's own views, and the less likely one is to be influenced by the views of neighbors. The specific formula is as follows: Among them, k i is the out-degree of node i, k max is the maximum node degree in the network; The relative influence function is F(i,j). In the process of opinion interaction, due to the heterogeneity of nodes, they will have different influences on neighboring nodes. Select node i in the network and define its influence relative to node j. The specific formula is as follows: Among them, ω(i) is the set of all neighbor nodes of node i, ∑ r∈ω(i) k r is the sum of the out-degrees of all neighboring nodes of node i; that is, the greater the out-degree of node i, the more fans it has, and the greater its influence; similarly, the relative influence of its neighbor node j is: The relative influence function F(i,j) of node i on node j is defined as: In general, F(i,j)≠F(j,i); The organizational atmosphere is O i (t), the evolution of individual opinions is related to the organizational atmosphere to a certain extent, especially in social networks, where individuals are closely connected with organizations. Individual opinions will not only directly affect other individuals, but also form group pressure and indirectly affect the opinions of other individuals, ultimately making different opinions in the group tend to be consistent; the impact of organizational atmosphere on different individuals i at time t is: Among them, λ represents the sense of belonging to the organization, that is, the greater the degree to which individual i is influenced by the organizational atmosphere, the stronger the sense of belonging to the organization and the more easily the individual is influenced by the organizational atmosphere, so that the individual's views are consistent with those of the organization as a whole; Indicates the tendency of the mainstream opinion in the organization; when When , it means that most people in the organization hold positive views, that is, the organizational atmosphere has a positive impact on personal views; When , it means that most people hold negative views and the organizational atmosphere has a negative impact on individual views.

4. According to claim 3, a method for quantitatively analyzing the reversal degree of microblog network public opinion is characterized in that: The improvement of the classic bounded trust HK model by considering the individual's self-affirmation strength, relative influence and external organizational atmosphere factors mentioned in step 3) specifically includes: Taking into account the impact of changes in surrounding individual opinions and the combined influence of organizational atmosphere on individual opinions, the self-affirmation strength is introduced, so that each individual has different levels of confidence in his or her own opinions, thereby making the improved model more in line with the actual situation. The improved HK model is specifically expressed as follows: x i (t+1) represents the opinion value of individual i at time t+1; n represents the total number of individuals in the group; a trust threshold ε is set, and the opinion of individual j will only affect individual i if the difference between the opinion value of individual i and the opinion value of individual j is less than ε; Among them, ρ i The strength of self-affirmation is the degree to which an individual adheres to his or her own views; i (t) is the organizational atmosphere, F(j,x) is the relative influence function of node j on node i, is the weight of the influence of individual i on individual j, is the number of neighbors of individual i at time t, the trust threshold ε∈(0,1), x j (t) is the opinion value of individual j at time t.

5. The method for quantitatively analyzing the reversal degree of microblog network public opinion according to claim 1 is characterized in that: The step 4) specifically includes: according to the centrality index, finding the top M nodes in the centrality ranking in the constructed directed scale-free network and marking them as opinion leaders; when the algorithm is executed to time t1, it indicates that the public opinion information has erupted for the first time, and the opinion of the opinion leader at this time is set to Opinion1; when the algorithm runs to time t2, it indicates the second spread caused by the disclosure of new information, and the opinion of the opinion leader is set to Opinion2; by looping step 3), the opinion values ​​of all nodes are continuously updated, and when the algorithm running time reaches the maximum value T, the evolution stops, otherwise it returns to execute step 3).

6. A method for quantitatively analyzing the degree of reversal of microblog network public opinion according to claim 5, characterized in that: The centrality index specifically includes degree centrality, betweenness centrality and closeness centrality; the calculation formula of degree centrality is: Among them, m represents the number of other nodes connected to node i, and n represents the total number of nodes in the entire social network; The calculation formula for betweenness centrality is: Among them, count i (j,k) represents the number of paths passing through node i in the shortest path from node j to node k, and count(j,k) represents the total number of all shortest paths from node j to node k; Closeness centrality is defined as the inverse of the average shortest path length from a node to other nodes, specifically expressed as: Among them, d ij represents the average distance from node i to node j.

7. The method for quantitatively analyzing the reversal degree of microblog network public opinion according to claim 1 is characterized in that: The degree of public opinion reversal in step 5) is defined as: Among them, x max represents the maximum value of the average opinion value during the process of public opinion dissemination, x min represents the minimum value of the average opinion value in the process of public opinion dissemination; since the research object is the public opinion reversal event, x max Generally positive, and x min To avoid the situation where the denominator is 0, the numerator and denominator in the formula are smoothed. p The larger the value is, the greater the tendency of public opinion reversal is, that is, the greater the possibility of reversal.

Images