Mutation Prediction and Control Method for an Online Group Opinion Evolution System

Abstract

The present invention discloses a method for predicting and controlling mutations in an online group opinion evolution system, belonging to the technical fields of text mining and mathematical modeling. By processing network text data, system variables of group opinion evolution are obtained; and a mutation model of group opinion evolution is established based on catastrophe theory; an elasticity index model of the system is constructed from the mutation model to analyze the possibility of opinion mutation and the danger of system state, thereby predicting mutations. A method for finding the threshold of the system elasticity index is proposed based on the elasticity index model, and opinion mutations are prevented according to the comparison result between the elasticity index and the threshold. Furthermore, a control method for the elasticity index is proposed, and a random search algorithm for reducing the system elasticity index is proposed to reduce the possibility of system mutations. Thus, the present invention can accurately describe and predict the mutation phenomenon of the online group opinion evolution system, and can reduce the possibility of group opinion mutations with less time consumption.

Description

Technical Field

[0001] The present invention belongs to the technical field of text mining and mathematical modeling, and more specifically, relates to a method for predicting and controlling mutations in an online group opinion evolution system. Background Art

[0002] With the development of Web 2.0, various media have become popular, and people tend to share knowledge and opinions online. The Internet of Things enables everything to be interconnected, creating an intelligent environment where people, things, and machines interact. In this intelligent environment, the interaction between people, things, and machines forms user-generated content (UGC). UGC reflects users' opinions and ideas to a certain extent and is an important research material. Obtaining the group attitude from UGC is an important task for managers. For managers, the polarization of attitudes and sudden changes in opinions are crucial issues. Converting online network text data into system variables requires the support of text mining technology. Text mining refers to obtaining valuable information and knowledge from text data. Text mining is a multidisciplinary field that encompasses various technologies such as machine learning, natural language processing, and mathematical statistics.

[0003] The online group opinion evolution system is a typical complex system with characteristics such as mutations, unpredictability, non-linearity, and instability. Mutability means that the system state suddenly changes at certain critical nodes and is unpredictable. Traditional linear models are difficult to accurately describe and predict such mutation phenomena. Mutation theory was proposed by the French mathematician Thom in the early 1970s and can better describe the mutation phenomena of complex systems. The cusp mutation model is the most commonly used, where the system is influenced by two control variables and the potential function is The equilibrium surface equation is -y 3 + uy + v = 0. Its equilibrium surface is as Figure 1 shown and consists of three parts: the upper surface, the lower surface, and the middle folded region. The upper and lower surfaces represent the two stable states of the system, and the fold is the mutation region of the system state. After obtaining the system variables, a mutation model can be established by fitting with cusp - package. Cusp - package is a mutation model fitting tool on the R platform developed by Raoul et al., and its principle is based on Cobb's maximum likelihood method.

[0004] However, existing mutation models are still difficult to accurately describe and predict the mutation phenomena of the online group opinion evolution system. Summary of the Invention

[0005] In view of the defects and improvement requirements of the prior art, the present invention provides a method for predicting and controlling mutations in an online group opinion evolution system. Based on the elasticity theory, an elasticity index is used to predict the possibility of system mutations; and when the system is predicted to have a possible mutation, a method for controlling the elasticity index of the system is proposed to reduce the possibility of system mutations.

[0006] To achieve the above object, in a first aspect, the present invention provides a method for predicting mutations in an online group opinion evolution system, including the following steps:

[0007] S1, obtain network text data related to the online group opinion evolution system, and determine the dependent variables Y1,...,Y p and independent variables X1,...,X q ;

[0008] S2, according to the dependent variables Y1,...,Y p and independent variables X1,...,X q fit and establish a cusp mutation model of the online group opinion evolution system; the equilibrium surface equation of the cusp mutation model is -f 3 +uf+v = 0, where f is the state variable, u and v are control variables, and satisfy f = w0+w1Y1+w2Y2+…+w p Y p , v = a0+a1X1+a2X2+…+a q X q , u = b0+b1X1+b2X2+…+b q X q , where w0,w1,...,w p , a0,a1,...,a q , b0,b1,...,b q are all fitting coefficients;

[0009] S3, according to the curve of the change of the state variable f with the control variable v when the control variable u is greater than 0 or less than 0, calculate the integral area between the current state and the stable state of the online group opinion evolution system, and use it as the elasticity index corresponding to the current state;

[0010] S4, if the elasticity index exceeds the threshold, it indicates that the online group opinion evolution system may have a mutation, otherwise, there is no possibility of mutation.

[0011] Further, in S1, the independent variables and dependent variables of the online group opinion evolution system are obtained by performing natural language processing and semantic network structure analysis on the network text data.

[0012] Further, in S3, the resilience index corresponding to the current state is specifically calculated in the following manner:

[0013] (1) When u0 < 0,

[0014] where resilience represents the resilience index, u0 is the value of the control variable u corresponding to the current state, v0 is the value of the control variable v corresponding to the current state, boundary is the value of the control variable v corresponding to the stable state, and f(v) is the unique real solution of the equilibrium surface equation -f 3 + u0f + v = 0;

[0015] (2) When u0 > 0,

[0016] where v i is the value of the control variable v corresponding to the boundary of the mutation domain; f1(v) and f3(v) are respectively the maximum and minimum real solutions of the equilibrium surface equation -f 3 + u0f + v = 0.

[0017] Further, in S4, the threshold is determined in the following manner:

[0018] First, calculate the resilience index corresponding to each state point of the online group opinion evolution system;

[0019] Without changing the value of the control variable u, change the control variable v with Δv as the fixed change amount, and calculate the change of the system state at each state point;

[0020] Finally, compare the resilience index with the change of the system state, and construct a change trajectory diagram of the system state with respect to the resilience index to determine the threshold of the resilience index;

[0021] where the system state change state change = max(|f(u, v + Δv) - f(u, v - Δv)|), and f(u, v + Δv) and f(u, v - Δv) are respectively the real solutions of the equilibrium surface equation -f 3 + uf + v = 0.

[0022] In a second aspect, the present invention provides a mutation control method for an online group opinion evolution system, characterized in that when it is predicted by the mutation prediction method of the online group opinion evolution system described in the first aspect that there is a possibility of mutation in the online group opinion evolution system, with the goal of minimizing adjust the independent variables X1,..., X q ;

[0023] where Δresilience iIt represents the change in the elasticity index of each system state point after adjusting the independent variable, and n represents the total number of state points; |ΔX i | represents the absolute value of the change in the independent variable X i , and q represents the total number of independent variables.

[0024] Furthermore, the particle swarm optimization algorithm is used to search for the optimal solution.

[0025] In a third aspect, the present invention provides a computer-readable storage medium, which includes a stored computer program. When the computer program is run by a processor, it controls the device where the storage medium is located to execute the mutation prediction method of the online group opinion evolution system as described in the first aspect and / or the mutation control method of the online group opinion evolution system as described in the second aspect.

[0026] Generally speaking, through the above technical solutions conceived by the present invention, the following beneficial effects can be achieved:

[0027] (1) Based on the elasticity theory and innovatively proposing a way to calculate the elasticity index on the basis of the cusp catastrophe model, the present invention can accurately predict the possibility of mutation in the online group opinion evolution system.

[0028] (2) On the basis of the mutation prediction method, the present invention proposes a method to control the elasticity index of the online group opinion evolution system to reduce the possibility of system mutation.

[0029] (3) The present invention integrates a random search algorithm with a time complexity of 0(mn 2 ), enabling the control algorithm to reduce the possibility of mutation of group opinions with less time consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is the equilibrium surface of the cusp catastrophe model;

[0031] Figure 2 is the flowchart of a mutation prediction method for an online group opinion evolution system provided by the present invention;

[0032] Figure 3 is the structural diagram of the text analysis and data processing method for online group opinions provided by the present invention;

[0033] Figure 4 is the change path of the system state in the equilibrium surface of the catastrophe model when the control variable u is less than 0;

[0034] Figure 5 is the change path of the system state in the equilibrium surface of the catastrophe model when the control variable u is greater than 0;

[0035] Figure 6When the control variables u < 0 and v < 0, it is the basis for the elastic index modeling of the present invention for online group opinions;

[0036] Figure 7 When the control variables u < 0 and v > 0, it is the basis for the elastic index modeling of the present invention for online group opinions;

[0037] Figure 8 When the control variable u > 0, v < 0 and outside the mutation domain, it is the basis for the elastic index modeling of the present invention for online group opinions;

[0038] Figure 9 When the control variable u > 0, v < 0 and within the mutation domain, it is the basis for the elastic index modeling of the present invention for online group opinions;

[0039] Figure 10 When the control variable u > 0, v > 0 and within the mutation domain, it is the basis for the elastic index modeling of the present invention for online group opinions;

[0040] Figure 11 When the control variable u > 0, v > 0 and inside and outside the mutation domain, it is the basis for the elastic index modeling of the present invention for online group opinions;

[0041] Figure 12 It is the change of the system state caused by the change of the control variable v;

[0042] Figure 13 For reducing the possibility of system mutation, it is the algorithm flow of the elastic index control method of the present invention for online group opinions;

[0043] Figure 14 It is an example of the post text data of the Dada Forum selected by the present invention;

[0044] Figure 15 It is the meaning of the system variables of the group opinions of the Dada Forum example selected by the present invention;

[0045] Figure 16 It is the goodness of fit of the mutation model of the Dada Forum example selected by the present invention;

[0046] Figure 17 It is the system state surface diagram of the Dada Forum example selected by the present invention;

[0047] Figure 18 It is the elastic index surface diagram of the group opinions of the Dada Forum example selected by the present invention;

[0048] Figure 19 It is the diagram of the change of the system state with the elastic index in the Dada Forum example selected by the present invention;

[0049] Figure 20 It is a graph showing the changes of the system state and the elasticity index over time in the Dada Forum example selected by the present invention within a certain period of time;

[0050] Figure 21 It is the operation process of the search algorithm of the group opinion elasticity index control method of the Dada Forum example selected by the present invention. Detailed implementation manners

[0051] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0052] In the present invention, the terms "first", "second", etc. (if any) in the present invention and the accompanying drawings are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence.

[0053] The existing system elasticity theory defines elasticity as the ability of a system to maintain a stable state or restore its original state. Jean's research believes that elasticity is the corresponding energy of external stimuli. When the external stimuli exceed the maximum energy that the system can withstand, the system state will mutate. Based on these elasticity theories, the present invention provides a mutation prediction for an online group opinion evolution system, using an elasticity index to predict the possibility of system mutation; and when the system is predicted to have a possible mutation, a method for controlling the system elasticity index is proposed to reduce the possibility of system mutation.

[0054] Refer to Figure 2 , the present invention provides a mutation prediction method for an online group opinion evolution system, and the method includes operations S1 to S4.

[0055] Operation S1: Obtain network text data related to the online group opinion evolution system, and determine the dependent variables Y1,..., Y p and independent variables X1,..., X q .

[0056] In this embodiment, the network text data is obtained from an online network forum. In the network forum, the thread starter makes the first post, which causes others to reply and discuss. The post of the thread starter is used as an external stimulus, and the floor discussion is the research object. The research object of the present invention is the group opinion. The sentiment analysis method in the text mining method is used to obtain the sentiment value of the text. The average value of the sentiment values of the text of each person's post in the floor discussion is used as the dependent variable of the system, that is, the state variable of the system.

[0057] The independent variables X1,..., X of the system q are the factors that affect the dependent variable of the system. First, the speech of the host as an external stimulus, its emotional value should be included in the independent variables of the system. The present invention also obtains the independent variables of the system by performing natural language processing and semantic network structure analysis on the speech text of the floor discussion. The specific methods include the emotional analysis and word frequency analysis of the keywords in the speech text, and the complex network analysis of the language network of the speech text. The specific data analysis method structure of the system variables is as Figure 3 shown.

[0058] Operation S2, according to the dependent variables Y1,..., Y p and independent variables X1,..., X q fit to establish the cusp catastrophe model of the online group opinion evolution system; the equilibrium surface equation of the cusp catastrophe model is -f 3 +uf+v = 0, where f is the state variable, u and v are the control variables, and satisfy f = w0+w1Y1+w2Y2+…+w p Y p , v = a0+a1X1+a2X2+…+a q X q , u = b0+b1X1+b2X2+…+b q X q , where w0, w1,..., w p , a0, a1,..., a q , b0, b1,..., b q are all fitting coefficients.

[0059] In this embodiment, the system variables (dependent variables Y1,..., Y p and independent variables X1,..., X q ) are input into cusp - package to fit and establish the cusp catastrophe model. The fitting of the state variable and control variable of the cusp catastrophe model is based on f = w0+w1Y1+w2Y2+…+w p Y p , v = a0+a1X1+a2X2+…+a q X q , u = b0+b1X1+b2X2+…+b q X q q to obtain. After fitting these parameters, the formulas of the state variable and control variable can be obtained, and thus the catastrophe model can be determined.

[0060] At the same time, after cusp - package fits to obtain the formula of the cusp catastrophe model, it can also obtain the goodness of fit of the model. The fitting result can be based on the goodness of fit index R 2, AIC and BIC are used to compare the mutation model, the logistic model and the linear model. The optimal fitting result is that R 2 is the largest and AIC and BIC are the smallest.

[0061] In operation S3, according to the curve of the change relationship of the state variable f with respect to the control variable v when the control variable u is greater than 0 or less than 0, calculate the integral area between the current state and the stable state of the online group opinion evolution system, and use it as the resilience index corresponding to the current state.

[0062] In this embodiment, a resilience index model is established based on the cusp mutation model of the system to predict the possibility of the system state mutation. Observe the equilibrium surface of the mutation model, as Figure 1 shown, it can be found that when the control variable u takes different values, the change paths of the system state are not the same. When u is less than 0, the system state will experience a generally relatively gentle continuous change, and at the same time, when v approaches 0, the system state will experience a faster change, as Figure 4 shown. When u is greater than 0, the system state will experience a large discontinuous mutation, as Figure 5 shown. Therefore, the control variable u determines whether the mutation will occur, and the control variable v determines when the mutation will occur.

[0063] Based on the existing elasticity theory, the present invention takes the internal pressure currently borne by the system as the resilience index. The source of this pressure is the internal energy gap between the stable state and the unstable state of the system. When the control variable u takes different values, the present invention calculates the resilience index according to different change paths of the system state. When u is less than 0, the resilience index is obtained according to the integral area between the current state and the stable state of the system, and the rules are as Figure 6 and Figure 7 shown. When u is greater than 0, the resilience index is also obtained from the integral area between the current state and the stable state of the system, and the rules are as Figures 8 - 11 shown, and it is divided into four cases according to the magnitude of v and whether it is within the mutation domain. Figures 6 - 11 Among them, u0 is the value of the control variable u corresponding to the current state, v0 is the value of the control variable v corresponding to the current state, v i is the value of the control variable v corresponding to the boundary of the mutation domain, boundary is the value of the control variable v corresponding to the stable state, and it is an exogenous parameter.

[0064] From the above calculation rules of the resilience index, the obtained resilience index model is as follows:

[0065] (1) When u0 < 0,

[0066] where resilience represents the resilience index, f(v) is the equilibrium surface equation -f3 The unique real solution of +u0f + v = 0;

[0067] (2) When u0 > 0,

[0068] where resilience represents the resilience index, and f1(v) and f3(v) are respectively the maximum and minimum real solutions of the equilibrium surface equation -f 3 +u0f + v = 0.

[0069] In operation S4, if the resilience index exceeds the threshold, it indicates that there may be a mutation in the online group opinion evolution system; otherwise, there is no possibility of mutation.

[0070] Refer to Figure 1 , there is a mutation domain on the equilibrium surface of the mutation model. Within this mutation domain, a small change in the control variable will cause a mutation in the system state. As Figure 12 shown, assuming that the control variable u remains unchanged and the control variable v undergoes a fixed change amount Δv, the state change of the system: state change = max(|f(u, v + Δv) - f(u, v - Δv)|). Where f(u, v + Δv) and f(u, v - Δv) are the real solutions of the equilibrium surface equation -f 3 +uf + v = 0. This system state change is the maximum value of the difference between two adjacent states of the system.

[0071] The present invention finds the threshold of the system resilience index through the system state change caused by the change of the control variable v. After establishing the resilience index model of the system, the resilience index of each state point of the system is calculated. Then, by comparing the resilience index with the system state change, a change trajectory graph of the system state change with respect to the resilience index is constructed, and the threshold point of the resilience index is found by finding the key points of the system state change.

[0072] After obtaining the threshold of the resilience index, the resilience index value of the current state of the system is compared with the threshold, so as to judge the stability of the current state of the system.

[0073] In addition, since the resilience index model can predict the possibility of system mutation, the greater the resilience index, the higher the possibility of system mutation. Therefore, by controlling the control variables u and v to control the system state and reducing the resilience index, the possibility of mutation can be reduced.

[0074] Since the control variables u and v are affected by the independent variables X1,..., X q the resilience index of the system is ultimately controlled by adjusting the independent variables. The control function of the system is control(ΔX1,..., ΔX q) The values of each independent variable are adjusted through this interface to control the system state. The control objective is to minimize the adjustment of the control variables while minimizing the system resilience index, and the objective function is to find the minimum value. Where |ΔX i | is the absolute value of the change in the independent variable X i , and the smaller the sum of its absolute values, the better; while Δresilience i is the change in the resilience index of each system state point after adjusting the independent variable, and the smaller the sum of its absolute values, the better; therefore, the objective function is to find the minimum value.

[0075] After obtaining the control function and the objective function, the present invention searches for the optimal solution by establishing a particle swarm optimization algorithm to obtain the optimal objective function value and the adjusted value of the independent variable. The particle swarm algorithm is a stochastic search algorithm based on group collaboration, and its algorithm idea is to simulate the foraging process of a flock of birds and is applicable to optimization problems in continuous space. The algorithm flow established by the present invention is as Figure 13 shown. The time complexity of the algorithm is 0(mn 2 ), where m is the number of iterations and n is the number of particles.

[0076] Hereinafter, the present invention takes the Dada Delivery Bar on Baidu Tieba as an example to further elaborate in detail on the mutation prediction and control method of the online group opinion evolution system. The Dada Bar is a forum for Dada crowdsourcing logistics employees to communicate. Dada crowdsourcing logistics employees obtain and complete orders through smartphones and also communicate on the forum through smartphones. Therefore, the online group opinion system of the Dada Bar is a typical complex system, and its system state is the group opinion.

[0077] The present invention has collected 21,494 post text data. After data screening and preliminary processing, there are 17,956 valid data. The post text data is as Figure 14 shown. One piece of data is a system state point. After text analysis and data processing, 1 dependent variable Y and 5 independent variables X1,..., X5 are obtained. The specific meanings of the variables are as Figure 15 shown.

[0078] After obtaining the system variables, next, a cusp mutation model of the online group opinion evolution system is established, and the cusp-package is used for the fitting of the mutation model. The fitting results obtained after inputting the system independent variables and the dependent variable are as Figure 16 shown. It can be found that the goodness of fit R 2 of the cusp mutation model is the highest, reaching 0.92. Therefore, the present invention believes that the group opinion system of the Dada forum example is most suitable for modeling with the cusp mutation model.

[0079] Further fitting is performed to obtain the formula and various parameters of the mutation model. State variable: y = 1.120638 + 2.430648Y, control variables: v = 0.159945 + 0.132075X1 + 0.143132X2 + 0.340591X3 + 0.044981X4 - 0.190836X5, u = 1.686694 + 0.047870X3 - 0.093710X4 + 0.102391X5. Therefore, the mutation model -y 3 + uy + v = 0 can be determined. In this way, the values of the control variables u, v and the state variable y of each text data can be calculated. Taking the control variables u, v as the x, y axes and the state variable as the z axis, a three-dimensional surface diagram of the system state of the group view is plotted, as Figure 17 shown. It can be observed that there is a mutation region in the system state surface diagram, and the fluctuations between system state points are very large.

[0080] After establishing the mutation model of the group view of the Dada Forum, an elasticity index model of the group view is established, and the elasticity index of each text data in the Dada Forum is calculated. Taking the control variables u, v as the x, y axes and the elasticity index as the z axis, a three-dimensional surface of the elasticity index of the group view is plotted, as Figure 18 shown. It can be found that there is also a mutation region in the elasticity index surface, and the system elasticity index in this region is significantly higher than that in other regions.

[0081] After establishing the elasticity index model of the Dada Forum, according to step 4, its threshold is found. A trajectory diagram of the state change of the Dada Forum with the change of the elasticity index is plotted, as Figure 19 shown. It can be found that there are two thresholds 4.8 and 5.7 for the elasticity index of the system, and the state change of the system generally increases with the increase of the elasticity index. When the elasticity index is less than 4.8, the state change increases slowly; when the elasticity index exceeds 4.8, the state change increases rapidly; when the elasticity index exceeds 5.7, the state change starts to increase slowly again. That is to say, when the elasticity index reaches 4.8, the system continuously enters an unstable state; when the elasticity index exceeds 5.7, the system is still in a dangerous state, but the degree of danger does not increase significantly.

[0082] At this time, observe the changes of the system state (group view) and the elasticity index of the Dada Forum over a period of time, as Figure 20 shown. It can be found that when the system enters the dangerous state and undergoes a state mutation, the value of the elasticity index is also very large.

[0083] After establishing the elasticity index model of the online group view evolution, a particle swarm optimization algorithm is used to search for the direction and optimal value with the fastest decline of the elasticity index. The running process of the search algorithm in the Dada Forum example is as Figure 21As shown, the horizontal axis is the number of iterations and the vertical axis is the objective function value. The direction of the fastest decline in the elasticity index of the Dada Forum and the optimal value obtained by the final search algorithm are [0.06719399628179326, 0.05181991089446525, 1.5448918771280307, -14.853762843365354, -1.0468903041477826] and -58.63260341354989 respectively.

[0084] It is easy for those skilled in the art to understand that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A mutation prediction method for an online group opinion evolution system, characterized in that Including the following steps: S1. Obtain network text data related to the online group opinion evolution system and determine the dependent variables Y1,..., Y p and independent variables X1,..., X q ; S2. According to the dependent variables Y1,..., Y p and independent variables X1,..., X q a cusp catastrophe model of the online group opinion evolution system is established by fitting; the equilibrium surface equation of the cusp catastrophe model is -f 3 + uf + v = 0, where f is the state variable, u and v are control variables, and f satisfies f = w0 + w1Y1 + w2Y2 + … + w p Y p , v = a0 + a1X1 + a2X2 + … + a q X q , u = b0 + b1X1 + b2X2 + … + b q X q , where w0, w1,..., w p , a0, a1,..., a q , b0, b1,..., b q are all fitting coefficients; S3. According to the curve of the change of the state variable f with the control variable v when the control variable u is greater than 0 or less than 0, calculate the integral area between the current state and the stable state of the online group opinion evolution system, and use it as the elasticity index corresponding to the current state; S4. If the elasticity index exceeds the threshold, it indicates that there may be a mutation in the online group opinion evolution system; otherwise, there is no possibility of mutation.

2. The mutation prediction method for the online group opinion evolution system according to claim 1, characterized in that In S1, the independent variable and the dependent variable of the online group opinion evolution system are obtained by performing natural language processing and semantic network structure analysis on the network text data.

3. The mutation prediction method for the online group opinion evolution system according to claim 1, characterized in that In S3, the elasticity index corresponding to the current state is specifically calculated in the following manner: (1) When u0 < 0, Among them, resilience represents the resilience index, u0 is the value of the control variable u corresponding to the current state, v0 is the value of the control variable v corresponding to the current state, boundary is the value of the control variable v corresponding to the stable state, and f(v) is the unique real solution of the equilibrium surface equation -f 3 + u0f + v = 0; (2) When u0 > 0, wherein, v i is the value of the control variable v corresponding to the boundary of the mutation domain; f1(v) and f3(v) are respectively the maximum and minimum real number solutions of the equilibrium surface equation -f 3 + u0f + v = 0.

4. The mutation prediction method of the online group opinion evolution system according to claim 3, characterized in that In S4, the threshold is determined in the following manner: First, calculate the elasticity index corresponding to each state point of the online group opinion evolution system; Without changing the value of the control variable u, change the control variable v with a fixed change amount of Δv, and calculate the system state change at each state point; Finally, compare the elasticity index with the system state change, and construct a change trajectory diagram of the system state change with the elasticity index to determine the threshold of the elasticity index; Among them, the system state change state change = max(|f(u, v + Δv) - f(u, v - Δv)|), where f(u, v + Δv) and f(u, v - Δv) are respectively the real number solutions of the equilibrium surface equation -f 3 + uf + v = 0.

5. A mutation control method for an online group opinion evolution system, characterized in that, When using the mutation prediction method of the online group opinion evolution system described in any one of claims 1 to 4 to predict the possibility of mutation in the online group opinion evolution system, with the goal of minimizing as the objective, adjust the independent variables X1,..., X q ; Among them, Δresilience i represents the change in the resilience index of each system state point after adjusting the independent variable, and n represents the total number of state points; |ΔX i | represents the absolute value of the change in the independent variable X i , and q represents the total number of independent variables.

6. The mutation control method of the online group opinion evolution system according to claim 5, characterized in that The particle swarm optimization algorithm is used to search for the optimal solution.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein when the computer program is run by a processor, it controls the device where the storage medium is located to execute the mutation prediction method of the online group opinion evolution system as described in any one of claims 1 to 4 and / or the mutation control method of the online group opinion evolution system as described in claim 5 or 6.

Images