Virus transmission path analysis system based on big data

Abstract

The present invention relates to the field of path analysis technology, and specifically refers to a virus propagation path analysis system based on big data, comprising a data screening module, a virus propagation path module, a path screening module, a spatiotemporal network construction module and a virus propagation influencing factor analysis module. The scheme introduces Poisson white noise into the virus propagation path module, adds random disturbances into the virus propagation process, and enables the model to adjust the diffusion speed over time through the dynamic change of the virus diffusion factor; a G‑DBSCAN algorithm is designed in the path screening module to select high-quality virus propagation paths from the predicted paths of virus propagation, and by introducing the WA algorithm into the G‑DBSCAN algorithm, the parallelization characteristics and the dynamically optimized decomposition mechanism have excellent performance in terms of computational efficiency, can process large-scale data sets, better process sparse or non-uniformly distributed virus propagation path data, and identify potential key propagation routes.

Description

Technical Field

[0001] The present invention relates to the field of path analysis technology, and in particular to a virus propagation path analysis system based on big data. Background Art

[0002] Virus transmission is a major challenge facing global public health. Existing transmission path analysis has problems of low prediction precision and accuracy. Due to the randomness caused by factors such as environment, population mobility, and policy intervention, the prediction results deviate from the actual situation. In addition, for virus transmission, it will be dynamically adjusted due to changes in external conditions in reality, making it difficult to cope with sudden transmission or relief stages. General path screening can use ordinary clustering algorithms, but traditional clustering algorithms have problems of inaccurate parameter selection and low adaptability. Parameter selection is usually based on manual experience or simple grid search, resulting in missed detection of important virus transmission paths or mistaking irrelevant paths for critical paths. In addition, for ordinary clustering algorithms, important paths in low-density areas are easily ignored, thus failing to fully reflect the laws of transmission. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a virus transmission path analysis system based on big data. In view of the problems of low prediction precision and low accuracy in the existing transmission path analysis, the present invention introduces Poisson white noise in the virus transmission path module, adds random disturbances in the virus transmission process, and more realistically simulates the randomness that may occur in the actual transmission process. Through the dynamic change of the diffusion factor, the model can adjust the diffusion speed over time, thereby accurately describing the nonlinear characteristics of virus transmission. At the same time, the state equation is introduced, combined with Poisson noise and diffusion terms, to fully reflect the complex process of virus transmission, including intra-regional diffusion and inter-regional transmission; in view of the problems of inaccurate parameter selection and In order to solve the problem of low adaptability, the present invention designs a G-DBSCAN algorithm in the path screening module to select high-quality virus transmission paths from the predicted paths of virus transmission. By introducing the WA algorithm (WASP-ANT) into the G-DBSCAN algorithm, the problem that parameter selection in the traditional DBSCAN algorithm relies on manual experience is overcome. The parallelization characteristics of the WA algorithm and the dynamically optimized decomposition mechanism make G-DBSCAN perform well in terms of computational efficiency and can process large-scale data sets. At the same time, it combines the density clustering characteristics of the traditional DBSCAN and the intelligent optimization ability of the WA algorithm to better process sparse or non-uniformly distributed virus transmission path data and identify potential key transmission routes.

[0004] The technical solution adopted by the present invention is as follows: The virus propagation path analysis system based on big data provided by the present invention includes a data screening module, a virus propagation path module, a path screening module, a spatiotemporal network construction module and a virus propagation law analysis module, specifically including the following contents:

[0005] The data screening module collects virus propagation data, grids the geographical area, maps the virus propagation data according to time and geographical location, and constructs a spatiotemporal data set;

[0006] The virus propagation path module defines a virus diffusion factor to represent the probability of the virus propagation path, and constructs a dynamic propagation model based on Poisson white noise to generate a predicted path for virus propagation;

[0007] The path screening module selects high-quality virus propagation paths from the predicted virus propagation paths by designing a G-DBSCAN algorithm;

[0008] The spatiotemporal network construction module builds a dynamic spatiotemporal network model based on high-quality virus transmission paths and combined with spatiotemporal data sets, where each node represents a geographical area and each edge represents a high-quality virus transmission path;

[0009] The virus propagation law analysis module analyzes the impact of edges and nodes on propagation based on a dynamic spatiotemporal network model and generates an evaluation report.

[0010] Furthermore, the virus propagation module defines a virus diffusion factor to represent the probability of the virus propagation path, and constructs a dynamic propagation model based on Poisson white noise to generate a predicted path for virus propagation. The specific implementation steps are as follows:

[0011] Step S1: Introduce Poisson white noise, perform random perturbations based on the Poisson process, and add zero-mean white noise, which is defined as follows:

[0012] ;

[0013] in, Indicates the moment in the current virus spreading process, is Poisson white noise, indicating that at time The random perturbation of represents the Poisson intensity, is zero-mean white noise, satisfying ;

[0014] Step S2: Define the virus diffusion factor and initialize the virus diffusion factor based on Poisson white noise, which is expressed as follows:

[0015] ;

[0016] in, The virus spreading factor is the basic diffusion rate, which indicates the diffusion speed of the virus without random disturbance;

[0017] Step S3: Define the state equation to describe the virus propagation process. The state equation is expressed as follows:

[0018] ;

[0019] in, is the virus transmission density, indicating the and time The virus density, is the diffusion term, is the Poisson noise term;

[0020] Step S4: construct a model, construct a dynamic virus propagation model, set the initial infection area and the initial virus propagation density, and the initial virus propagation density is expressed as follows:

[0021] ;

[0022] in, for Initial virus transmission density at time ;

[0023] The boundary conditions of the initial infection area are expressed as follows:

[0024] ;

[0025] Step S5: Discretize the equation and use the finite difference method to perform spatial discretization to obtain the numerical solution of virus propagation. The formula used is as follows:

[0026] ;

[0027] in, represents the number of time steps, is the time step, is the discrete index of the location, is the spatial step length, For the spatial position and time step The virus density distribution at

[0028] Step S6: Predicting a path, simulating the virus propagation process according to the numerical solution of virus propagation, and generating a predicted path of virus propagation.

[0029] Furthermore, the G-DBSCAN algorithm in the path screening module specifically includes the following steps:

[0030] Step Q1: Parameter generation, using the WA algorithm to generate the neighborhood radius and the minimum number of samples;

[0031] Step Q2: Dataset preparation, putting the predicted path of virus transmission into the predicted path dataset;

[0032] Step Q3: In the clustering stage, the prediction path dataset is clustered using the DBSCAN algorithm to obtain the clustering results;

[0033] Step Q4: Iterative optimization, set the maximum number of iterations, and repeat steps Q2 to Q4 until the maximum number of iterations is reached to obtain the best clustering result;

[0034] Step Q5: Output high-quality virus propagation paths and extract high-quality virus propagation paths from the best clustering results.

[0035] Further, step Q2 uses the WA algorithm to generate the neighborhood radius and minimum number of samples required by DBSCAN, which specifically includes the following steps:

[0036] Step Q11: Initialization phase, define a neighborhood radius and a minimum number of samples as a parameter combination, define the search space, and randomly generate multiple initial solutions in the search space;

[0037] Step Q12: Update the initial solution, calculate the fitness value of the initial solution, sort according to the fitness value, and the solution with the smallest fitness value is the current optimal solution;

[0038] Step Q13: Iterate the loop and set the number of loop iterations;

[0039] Step Q14: In the chasing phase, a solution is randomly selected in the search space as the position of the wasp, and the remaining solutions are used as the positions of the ants. The positions of the ants are updated to move closer to the position of the wasp. The update formula used is as follows:

[0040] ;

[0041] in, Represents the new position of the ant, which is The position of the ant at the iteration, represents the current position of the ant, is a random number coefficient in the range [0,3], is the current optimal solution;

[0042] Step Q15: Attack phase, define perception signals , generate a random number between [0,1] as the perception signal, specifically including the following steps:

[0043] if , the wasp moves to the current optimal solution position:

[0044] ;

[0045] if , the wasp randomly jumps to a new location:

[0046] ;

[0047] in, Indicates the new position of the wasp, which is the current coordinate of the wasp, are the updated coordinates of the wasp, To control the wasp's stride length, is the lower bound of the search space, is the upper bound of the search space, is a random number;

[0048] Step Q16: Integration of solutions, integration and To integrate the group;

[0049] Step Q17: Fitness evaluation, perform fitness evaluation on all solutions in the integrated group and select the final solution.

[0050] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0051] (1) In view of the problems of low prediction precision and accuracy in the existing propagation path analysis, the present invention introduces Poisson white noise into the virus propagation path module and adds random disturbances during the virus propagation process to more realistically simulate the randomness that may occur in the actual propagation process. Through the dynamic change of the diffusion factor, the model can adjust the diffusion speed over time, thereby accurately describing the nonlinear characteristics of virus propagation. At the same time, the state equation is introduced, combined with Poisson noise and diffusion terms, to fully reflect the complex process of virus propagation, including intra-regional and inter-regional propagation.

[0052] (2) In view of the problems of inaccurate parameter selection and low adaptability of traditional clustering algorithms, the present invention designs a G-DBSCAN algorithm in the path screening module to select high-quality virus transmission paths from the predicted virus transmission paths. By introducing the WA algorithm (WASP-ANT) into the G-DBSCAN algorithm, the problem of parameter selection relying on manual experience in the traditional DBSCAN algorithm is overcome. The parallelization characteristics of the WA algorithm and the dynamically optimized decomposition mechanism make G-DBSCAN perform well in terms of computational efficiency and can process large-scale data sets. At the same time, it combines the density clustering characteristics of the traditional DBSCAN and the intelligent optimization capabilities of the WA algorithm to better process sparse or non-uniformly distributed virus transmission path data and identify potential key transmission routes. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a schematic diagram of the virus propagation path analysis system based on big data proposed in the present invention.

[0054] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0056] Example 1, see Figure 1 The virus propagation path analysis system based on big data provided by the present invention includes a data screening module, a virus propagation path module, a path screening module, a spatiotemporal network construction module and a virus propagation law analysis module, which specifically includes the following contents:

[0057] The data screening module collects virus propagation data, grids the geographical area, maps the virus propagation data according to time and geographical location, and constructs a spatiotemporal data set;

[0058] The virus propagation path module defines a virus diffusion factor to represent the probability of the virus propagation path, and constructs a dynamic propagation model based on Poisson white noise to generate a predicted path for virus propagation;

[0059] The path screening module selects high-quality virus propagation paths from the predicted virus propagation paths by designing a G-DBSCAN algorithm;

[0060] The spatiotemporal network construction module builds a dynamic spatiotemporal network model based on high-quality virus transmission paths and combined with spatiotemporal data sets, where each node represents a geographical area and each edge represents a high-quality virus transmission path;

[0061] The virus propagation law analysis module analyzes the impact of edges and nodes on propagation based on a dynamic spatiotemporal network model and generates an evaluation report.

[0062] Embodiment 2: This embodiment is based on the above embodiment. The virus propagation module defines a virus diffusion factor to represent the probability of the virus propagation path. A dynamic propagation model based on Poisson white noise is constructed to generate a predicted path for virus propagation. The specific implementation steps are as follows:

[0063] Step S1: Introduce Poisson white noise, perform random perturbations based on the Poisson process, and add zero-mean white noise, which is defined as follows:

[0064] ;

[0065] in, Indicates the moment in the current virus spreading process, is Poisson white noise, indicating that at time The random perturbation of represents the Poisson intensity, is zero-mean white noise, satisfying ;

[0066] Step S2: Define the virus diffusion factor and initialize the virus diffusion factor based on Poisson white noise, which is expressed as follows:

[0067] ;

[0068] in, The virus spreading factor is the basic diffusion rate, which indicates the diffusion speed of the virus without random disturbance;

[0069] Step S3: Define the state equation to describe the virus propagation process. The state equation is expressed as follows:

[0070] ;

[0071] in, is the virus transmission density, indicating the and time The virus density, is the diffusion term, is the Poisson noise term;

[0072] Step S4: construct a model, construct a dynamic virus propagation model, set the initial infection area and the initial virus propagation density, and the initial virus propagation density is expressed as follows:

[0073] ;

[0074] in, for Initial virus transmission density at time ;

[0075] The boundary conditions of the initial infection area are expressed as follows:

[0076] ;

[0077] Step S5: Discretize the equation and use the finite difference method to perform spatial discretization to obtain the numerical solution of virus propagation. The formula used is as follows:

[0078] ;

[0079] in, represents the number of time steps, is the time step, is the discrete index of the location, is the spatial step length, For the spatial position and time step The virus density distribution at

[0080] Step S6: Predicting a path, simulating the virus propagation process according to the numerical solution of virus propagation, and generating a predicted path of virus propagation.

[0081] In this embodiment, the set parameters are:

[0082] ;

[0083] ;

[0084] ;

[0085] : The initial infection area is [0,1], the virus density within this range is 1, and the rest of the area is 0;

[0086] ;

[0087] The maximum number of iterations was set to 1000 iterations;

[0088] At every moment , by generating a normally distributed random value Simulate white noise;

[0089] At the moment hour:

[0090] Random Generation =0.02 (sampled from N(0,0.01));

[0091] = =0.5+0.02=0.52.

[0092] In this embodiment, at each moment , using the generated in step S1 To calculate the diffusion factor

[0093] At the moment hour, =0.52,

[0094] The diffusion factor =0.1+(1+0.52)=1.62.

[0095] In this embodiment, at time ,Location :

[0096] Assume that Position time =0.5, =0.6, =0.4.

[0097] Diffusion Factor 1.62, Noise =0.52.

[0098] calculate ;

[0099] ;

[0100] The virus density at the next moment:

[0101] .

[0102] Embodiment 3, based on the above embodiment, step S6 specifically includes the following steps:

[0103] Step S61: Initialize the propagation path, define the propagation threshold, and initialize the propagation path matrix , record the virus transmission status, It is expressed as follows:

[0104] ;

[0105] in, is the propagation threshold;

[0106] Step S62: Dynamically update the propagation path. At each moment, update the propagation path matrix according to the numerical solution of virus propagation. The update rules are as follows:

[0107] ;

[0108] Step S63: Store the propagation path and record the propagation path matrix at each moment, expressed as ,

[0109] Step S64: Visualize the propagation path and draw the propagation path matrix at each moment;

[0110] Step S65: Calculate the virus transmission coverage rate, that is, calculate the proportion of the virus transmission area to the global area at the current moment;

[0111] Step S66: Analyze the propagation direction, analyze the propagation center according to the propagation path matrix at each moment, and obtain the main direction of virus propagation;

[0112] Step S67: fitting a mathematical model of virus propagation according to the propagation path matrix at each moment and the main direction of virus propagation;

[0113] Step S68: Generate a predicted path, based on a mathematical model that fits the virus propagation, generate a predicted path for the virus propagation.

[0114] Embodiment 4, based on the above embodiment, the G-DBSCAN algorithm in the path screening module specifically includes the following steps:

[0115] Step Q1: Parameter generation, using the WA algorithm to generate the neighborhood radius and minimum number of samples; (WASP-ANT)

[0116] Step Q2: Dataset preparation, putting the predicted path of virus transmission into the predicted path dataset;

[0117] Step Q3: In the clustering stage, the prediction path dataset is clustered using the DBSCAN algorithm to obtain the clustering results;

[0118] Step Q4: Iterative optimization, set the maximum number of iterations, and repeat steps Q2 to Q4 until the maximum number of iterations is reached to obtain the best clustering result;

[0119] Step Q5: Output high-quality virus propagation paths and extract high-quality virus propagation paths from the best clustering results.

[0120] Embodiment 5, based on the above embodiment, step Q2, uses the WA algorithm to generate the neighborhood radius and minimum number of samples required by DBSCAN, specifically including the following steps:

[0121] Step Q11: Initialization phase, define a neighborhood radius and a minimum number of samples as a parameter combination, define the search space, and randomly generate multiple initial solutions in the search space;

[0122] Step Q12: Update the initial solution, calculate the fitness value of the initial solution, sort according to the fitness value, and the solution with the smallest fitness value is the current optimal solution;

[0123] Step Q13: Iterate the loop and set the number of loop iterations;

[0124] Step Q14: In the chasing phase, a solution is randomly selected in the search space as the position of the wasp, and the remaining solutions are used as the positions of the ants. The positions of the ants are updated to move closer to the position of the wasp. The update formula used is as follows:

[0125] ;

[0126] in, Represents the new position of the ant, which is The position of the ant at the iteration, represents the current position of the ant, is a random number coefficient in the range [0,3], is the current optimal solution;

[0127] Step Q15: Attack phase, define perception signals , generate a random number between [0,1] as the perception signal, specifically including the following steps:

[0128] if , the wasp moves to the current optimal solution position:

[0129] ;

[0130] if , the wasp randomly jumps to a new location:

[0131] ;

[0132] in, Indicates the new position of the wasp, which is the current coordinate of the wasp, are the updated coordinates of the wasp, To control the wasp's stride length, is the lower bound of the search space, is the upper bound of the search space, is a random number;

[0133] Step Q16: Integration of solutions, integration and To integrate the group;

[0134] Step Q17: Fitness evaluation, perform fitness evaluation on all solutions in the integrated group and select the final solution.

[0135] In this embodiment, the codes used are as follows:

[0136] import numpy as np

[0137] from sklearn.metrics import silhouette_score

[0138] from sklearn.cluster import DBSCAN

[0139] from sklearn.datasets import make_blobs

[0140] # Sample data generation

[0141] data, _ = make_blobs(n_samples=300, centers=4, cluster_std=0.6,random_state=42)

[0142] #Fitness function definition

[0143] def fitness_function(data, eps, min_samples):

[0144] try:

[0145] db = DBSCAN(eps=eps, min_samples=min_samples).fit(data)

[0146] labels = db.labels_

[0147] # Ignore the case where all noise is present

[0148] if len(set(labels)) <= 1:

[0149] return float('inf')

[0150] score = silhouette_score(data, labels)

[0151] return -score # Negative silhouette coefficient as optimization target

[0152] except Exception:

[0153] return float('inf')

[0154] # Parameters

[0155] define_lb = [0.1, 2] # The lower bound of the neighborhood radius eps and the lower bound of the minimum number of samples

[0156] define_ub = [2.0, 10] # Upper bound of neighborhood radius eps and upper bound of minimum number of samples

[0157] num_solutions = 10 # The number of initial solutions

[0158] max_iter = 50 # Maximum number of iterations

[0159] # Initialization

[0160] def initialize_solutions(lb, ub, num_solutions):

[0161] solutions = []

[0162] for _ in range(num_solutions):

[0163] eps = np.random.uniform(lb[0], ub[0])

[0164] min_samples = np.random.randint(lb[1], ub[1] + 1)

[0165] solutions.append([eps, min_samples])

[0166] return np.array(solutions)

[0167] # Update ant position

[0168] def update_ant_position(ant_position, best_position, beta):

[0169] return ant_position + beta * (best_position - ant_position)

[0170] # Update the wasp position

[0171] def update_wasp_position(wasp_position, best_position, r):

[0172] return wasp_position + r * (best_position - wasp_position)

[0173] # Random jump

[0174] def random_jump(lb, ub):

[0175] return [np.random.uniform(lb[i], ub[i]) for i in range(len(lb))]

[0176] # WA-DBSCAN parameter optimization algorithm implementation

[0177] def wa_dbscan(data, lb, ub, num_solutions, max_iter):

[0178] solutions = initialize_solutions(lb, ub, num_solutions)

[0179] fitness = [fitness_function(data, sol[0], sol[1]) for sol insolutions]

[0180] best_solution = solutions[np.argmin(fitness)]

[0181] for t in range(max_iter):

[0182] # Chase Phase

[0183] beta = np.random.uniform(0, 3)

[0184] for i in range(len(solutions)):

[0185] if np.random.rand() < 0.5:

[0186] solutions[i] = update_ant_position(solutions[i],best_solution, beta)

[0187] # Attack Phase

[0188] for i in range(len(solutions)):

[0189] AS = np.random.rand()

[0190] if AS >= 0.5:

[0191] r = np.random.uniform(0, 1)

[0192] solutions[i] = update_wasp_position(solutions[i],best_solution, r)

[0193] else:

[0194] solutions[i] = random_jump(lb, ub)

[0195] # Fitness evaluation

[0196] fitness = [fitness_function(data, sol[0], sol[1]) for sol in solutions]

[0197] best_solution = solutions[np.argmin(fitness)]

[0198] return best_solution

[0199] # Run the algorithm

[0200] optimized_params = wa_dbscan(data, define_lb, define_ub, num_solutions, max_iter)

[0201] print("Optimized DBSCAN parameters:", optimized_params)。

[0202] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0203] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

[0204] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. A virus propagation path analysis system based on big data, characterized by: It includes data screening module, virus transmission path module, path screening module, spatiotemporal network construction module and virus transmission law analysis module, specifically including the following contents: The data screening module collects virus propagation data, grids the geographical area, maps the virus propagation data according to time and geographical location, and constructs a spatiotemporal data set; The virus propagation path module defines a virus diffusion factor to represent the probability of the virus propagation path, and constructs a dynamic propagation model based on Poisson white noise to generate a predicted path for virus propagation; The path screening module selects high-quality virus propagation paths from the predicted virus propagation paths by designing a G-DBSCAN algorithm; The spatiotemporal network construction module builds a dynamic spatiotemporal network model based on high-quality virus transmission paths and combined with spatiotemporal data sets, where each node represents a geographical area and each edge represents a high-quality virus transmission path; The virus propagation law analysis module analyzes the impact of edges and nodes on propagation based on a dynamic spatiotemporal network model and generates an evaluation report; The virus propagation path module defines the virus diffusion factor to represent the probability of the virus propagation path. It uses a dynamic propagation model based on Poisson white noise to generate a predicted path for virus propagation. The specific implementation steps are as follows: Step S1: Introduce Poisson white noise, perform random perturbations based on the Poisson process, and add zero-mean white noise, which is defined as follows: ； in, Indicates the moment in the current virus spreading process, is Poisson white noise, indicating that at time The random perturbation of represents the Poisson intensity, is zero-mean white noise, satisfying ; Step S2: Define the virus diffusion factor and initialize the virus diffusion factor based on Poisson white noise, which is expressed as follows: ； in, The virus spreading factor is the basic diffusion rate, which indicates the diffusion speed of the virus without random disturbance; Step S3: Define the state equation to describe the virus propagation process. The state equation is expressed as follows: ； in, is the virus transmission density, indicating the and time The virus density, is the diffusion term, is the Poisson noise term; Step S4: construct a model, construct a dynamic virus propagation model, set the initial infection area and the initial virus propagation density, and the initial virus propagation density is expressed as follows: ； in, for Initial virus transmission density at time ; The boundary conditions of the initial infection area are expressed as follows: ； Step S5: Discretize the equation and use the finite difference method to perform spatial discretization to obtain the numerical solution of virus propagation. The formula used is as follows: ； in, represents the number of time steps, is the time step, is the discrete index of the location, is the spatial step length, For the spatial position and time step The virus density distribution at Step S6: predicting a path, simulating the virus propagation process according to the numerical solution of virus propagation, and generating a predicted path of virus propagation; The G-DBSCAN algorithm in the path screening module specifically includes the following steps: Step Q1: Parameter generation, using the WA algorithm to generate the neighborhood radius and the minimum number of samples; Step Q2: Dataset preparation, putting the predicted path of virus transmission into the predicted path dataset; Step Q3: In the clustering stage, the prediction path dataset is clustered using the DBSCAN algorithm to obtain the clustering results; Step Q4: Iterative optimization, set the maximum number of iterations, and repeat steps Q2 to Q4 until the maximum number of iterations is reached to obtain the best clustering result; Step Q5: Output high-quality virus propagation paths, and extract high-quality virus propagation paths from the best clustering results; Step Q2, using the WA algorithm to generate the neighborhood radius and minimum number of samples required by DBSCAN, specifically includes the following steps: Step Q11: Initialization phase, define a neighborhood radius and a minimum number of samples as a parameter combination, define the search space, and randomly generate multiple initial solutions in the search space; Step Q12: Update the initial solution, calculate the fitness value of the initial solution, sort according to the fitness value, and the solution with the smallest fitness value is the current optimal solution; Step Q13: Iterate the loop and set the number of loop iterations; Step Q14: In the chasing phase, a solution is randomly selected in the search space as the position of the wasp, and the remaining solutions are used as the positions of the ants. The positions of the ants are updated to move closer to the position of the wasp. The update formula used is as follows: ； in, Represents the new position of the ant, which is The position of the ant at the iteration, represents the current position of the ant, is a random number coefficient in the range [0,3], is the current optimal solution; Step Q15: Attack phase, define perception signals , generate a random number between [0,1] as the perception signal, specifically including the following steps: if , the wasp moves to the current optimal solution position: ； if , the wasp randomly jumps to a new location: ； in, Indicates the new position of the wasp, which is the current coordinate of the wasp, are the updated coordinates of the wasp, To control the wasp's stride length, is the lower bound of the search space, is the upper bound of the search space, is a random number; Step Q16: Integration of solutions, integration and To integrate the group; Step Q17: Fitness evaluation, perform fitness evaluation on all solutions in the integrated group and select the final solution.

Images