PCTSE smart park security risk assessment method and system

Abstract

The invention relates to a PCTSE smart park safety risk assessment method and system, and belongs to the field of industrial park risk level assessment. The method comprises the following steps: S1, establishing a basic principle of a security evaluation system based on a PCTSE smart park concept, and determining indexes of all levels in the system according to the basic principle; s2, obtaining a safety evaluation sample of the PCTSE smart park through actual investigation and collection of Internet data, and preprocessing the sample; s3, according to expert scoring and specifications, determining evaluation standards and grades of indexes at all levels in the system; s4, optimizing hyper-parameters of the support vector machine through a marine predator algorithm, and constructing a park safety evaluation model; according to the preprocessed sample, training the model by using an Adaboost integration algorithm; and S5, performing park safety risk quantitative evaluation by using the trained model. According to the invention, safety risk quantitative evaluation is carried out on the park through the trained model, managers are helped to comprehensively understand the safety condition of the park, and the overall safety management level of the smart park is improved.

Description

Technical Field

[0001] The present invention relates to the field of park safety risk assessment, and in particular to a PCTSE smart park safety risk assessment method and a PCTSE smart park safety risk assessment system. Background Art

[0002] With the rapid development of information technology, smart parks, as an emerging urban management and service model, have gradually become an important part of modern urban development. PCTSE smart parks refer to comprehensive solutions that improve the operational efficiency and service level of parks through advanced information technology and intelligent management methods. PCTSE represents the integrated application of technologies such as the Internet of Things, cloud computing, big data, artificial intelligence, and edge computing. The concept of PCTSE smart parks emphasizes the comprehensive management of physical security, network security, technical security, social security, and ecological security. Smart parks achieve intelligent management and efficient use of various resources in the park through technical means such as the Internet of Things, big data, cloud computing, and artificial intelligence. However, with the rapid development of smart parks, security risk issues have become increasingly prominent, and security risk assessment has become an important part of ensuring the safe operation of the park. Smart parks are usually composed of multiple subsystems, including intelligent security, intelligent transportation, intelligent buildings, and intelligent energy management. These subsystems are interconnected through the network to form a highly integrated management platform. Its main features include: collecting large amounts of data through sensors and monitoring equipment, using big data analysis technology for real-time monitoring and decision support, using artificial intelligence algorithms to achieve intelligent scheduling and management of various resources within the park, improving operational efficiency, and sharing information between subsystems through the network to form a collaborative working whole.

[0003] Although the risk level assessment technology of smart parks has made some progress at this stage, it still faces many challenges and shortcomings in practical applications; the existing security risk assessment models are mostly qualitative or quantitative analysis, lacking the ability to comprehensively consider multiple risk factors. Many models fail to fully consider the mutual influence between different risks during the assessment process, resulting in one-sided assessment results. For example, the correlation between physical security risks and network security risks is often ignored, resulting in the inability to fully respond to potential threats when formulating security strategies. The data of each subsystem in the smart park often exists in an island phenomenon, making data integration and sharing difficult. The data format, standards and interfaces between different systems are not unified, resulting in the inability to effectively circulate information. This data island phenomenon makes it impossible to conduct security risk assessment based on a comprehensive data foundation, affecting the accuracy and effectiveness of the assessment. The application of artificial intelligence technology in security risk assessment is still in its infancy. Although some parks have begun to try to use machine learning and deep learning algorithms for risk prediction, the actual effect is often unsatisfactory due to insufficient data volume and insufficient model training. Summary of the invention

[0004] Based on this, it is necessary to provide a PCTSE smart park safety risk assessment method and system to address the problems of insufficient integration and sharing of existing PCTSE smart park safety risk assessment technology data, which leads to one-sidedness of assessment results and unsatisfactory assessment effects of existing artificial intelligence models.

[0005] The present invention is implemented by the following technical scheme: A PCTSE smart park safety risk assessment method comprises the following steps: S1: Establish the basic principles of the safety assessment system based on the PCTSE smart park concept, and determine the indicators at all levels in the safety assessment system based on the basic principles; S2: Obtain safety evaluation samples of PCTSE smart parks through actual investigation and collection of Internet data, and pre-process the safety evaluation samples; S3: Determine the evaluation standards and evaluation levels of indicators at all levels in the safety assessment system based on expert scores and various safety standards for park safety assessment; S4: The penalty coefficient and kernel parameter of the support vector machine are optimized by the marine predator algorithm to construct a park safety evaluation model; based on the preprocessed safety evaluation samples, the Adaboost integrated algorithm is used to train the park safety evaluation model; S5: Use the trained park safety evaluation model to conduct quantitative evaluation of PCTSE smart park safety risks.

[0006] In one of the inventions, in step S1, the basic principles of the security assessment system are established based on the PCTSE smart park concept. The basic principles include: (1) Comprehensiveness principle: The security assessment system should cover all relevant security factors in the park, including physical security, network security, data security, personnel security, etc., to ensure the comprehensiveness of the assessment; (2) Operability principle: The design of evaluation indicators should be based on accurately reflecting the degree of security risk in the park, and try to select comprehensive indicators with commonality, and strive for data operability; (3) Systematic principle: The indicator system should take into account the overall structure of the park and the relationship between each subsystem to ensure coordination and consistency among various indicators; (4) Participation principle: In the process of establishing the indicator system, the opinions and suggestions of all stakeholders should be fully considered to ensure the rationality and applicability of the indicators.

[0007] According to the basic principles, the indicators in the safety assessment system include five first-level indicators, namely, park hazards, park safety protection equipment, park employees' safety awareness, park emergency response capabilities and domino effect, as well as several second-level indicators.

[0008] The basic principles of establishing a safety assessment system based on the PCTSE smart park concept emphasize comprehensiveness, operability, systematicness and participation; through the basic principles, a scientific, reasonable and effective safety assessment framework is jointly formed to provide support for building a park safety assessment model.

[0009] In one of the inventions, in step S2, the method for obtaining safety evaluation samples of PCTSE smart parks includes: (1) actually visiting different smart parks to observe their safety management measures, technical applications and personnel configuration; (2) consulting relevant academic papers and technical documents on the Internet to collect safety evaluation standards and cases of other PCTSE smart parks.

[0010] Preprocessing includes data cleaning and data standardization. Data cleaning includes identifying missing values ​​in the sample and filling the missing values ​​with the mean. Data standardization processes the data through Z-score standardization. The calculation formula is as follows: In the formula, Represents the original data value, represents the mean of the data set, Represents the standard deviation of the data set.

[0011] The feature vector and true label of the jth sample after preprocessing are expressed as follows: in, represents the feature vector of the jth sample, (i=1, 2, ... n) represents the standard value of the i-th safety index of the sample, Represents the safety level of the jth sample, j=1, 2,…m.

[0012] Several preprocessed sample sets are randomly selected as test sets, and the remaining sample sets are used as training sets; the K-fold cross-validation strategy is adopted, and a part of the training set is taken as the validation set for each training, and the rest is used as the training set for that training.

[0013] The present invention collects safety evaluation samples of PCTSE smart parks through actual investigation and Internet data, and performs data cleaning and standardization in the sample preprocessing process to ensure the quality and applicability of sample data, providing a reliable basis for subsequent safety evaluation analysis. The present invention adopts the K-fold cross-validation strategy, which can reduce the overfitting of model training and significantly improve the reliability and effectiveness of the model.

[0014] In one of the inventions, in step S3, specific evaluation criteria are formulated for each indicator based on the collected safety specifications and expert opinions; the evaluation criteria for indicators at each level in the safety assessment system are divided into level one, level two, level three, level four, and level five, with the safety level decreasing in order from high to low.

[0015] In one of the inventions, in step S4, the steps of constructing a park safety assessment model are as follows: S41: Determine a kernel function suitable for a park safety assessment model based on a support vector machine; S42: Optimize the hyperparameters of the support vector machine through the marine predator algorithm to determine the optimal parameter combination; S43: Establishing a park safety evaluation model through a support vector machine based on optimal hyperparameters; S44: Based on the preprocessed security evaluation samples, the campus security evaluation model is trained based on the AdaBoost ensemble algorithm, and the performance of the trained model is evaluated.

[0016] In one of the inventions, in step S41, the calculation formula of the kernel function is as follows: In the formula, represents any two different sample vectors, Represents the parameters of the kernel function.

[0017] In one of the inventions, in step S42, the steps of optimizing the hyperparameters of the support vector machine by the marine predator algorithm are as follows: A: Set the population size, number of iterations, search space boundary, FaDs function and objective function of the marine predator algorithm; B: Randomly generate the initial population, each individual in the population represents a set of penalty parameters C and kernel parameters g of the support vector machine; C: Calculate the fitness values ​​of all individuals to determine whether the preset end conditions are met; if not, use the dynamic lens imaging reverse learning method to select the best individual; update the position of the best individual according to the marine predator algorithm; D: Calculate the fitness value of the optimal individual; determine whether the preset end condition is met. If not, return to step C until the calculation ends; output the optimized penalty parameter C and kernel parameter g.

[0018] The present invention optimizes the penalty parameter C and kernel parameter g of the support vector machine through the marine predator algorithm, which not only improves the performance and accuracy of the model, but also achieves the advantages of automatic tuning, global search, and rapid convergence, ensuring that the support vector machine is more efficient and reliable when establishing a model to process high-dimensional complex data.

[0019] In one of the inventions, in step S43, the objective function of the support vector machine is as follows: The decision function of the support vector machine is as follows: The constraints of the support vector machine are as follows Where w represents the normal vector of the hyperplane, represents the slack variable, represents the fixed value of the constraint bias term, represents the feature vector of the sample, represents the true label of the sample, represents the Lagrange multiplier.

[0020] In one of the inventions, in step S44, the steps of training the park safety evaluation model based on the AdaBoost ensemble algorithm are as follows: E: Select N data set samples D= , determines the weight of the initial iteration of each sample and the maximum number of iterations T; F: Based on the weighted training set, train the park safety assessment model and calculate the error rate of the tth round of model training and learning weights ; G: Pass error rate and learning weights Update the sample weights for the next round of training, and normalize the updated sample weights to ensure that their sum is 1; H: Return to step F until the maximum number of iterations is reached, and output the prediction results of the final model.

[0021] The present invention trains the park safety assessment model based on the AdaBoost integrated algorithm, and utilizes the weighted multiple weak classifier function of AdaBoost to effectively improve the accuracy and robustness of the model. Through the iterative mechanism of AdaBoost, each round of iteration can automatically adjust the weights of unimportant features to enhance the learning ability of the model, achieve a more comprehensive safety risk assessment, and improve the applicability and credibility of the model.

[0022] The present invention also adopts a PCTSE smart park safety risk assessment system, including: Information collection module: It is used to obtain safety evaluation samples of PCTSE smart parks based on actual investigations and collection of Internet information, pre-process the safety evaluation samples, and send the pre-processed safety evaluation samples to the processing module.

[0023] Processing module: The processing module includes a hyperparameter optimization unit, a park safety evaluation unit and an output unit. The hyperparameter optimization unit is used to find the optimal combination value of the penalty coefficient and kernel parameter in the support vector machine through the marine predator algorithm; the park safety evaluation unit constructs a park safety evaluation model based on the support vector machine and the optimal parameter combination value output by the hyperparameter optimization unit; the output unit trains the park safety evaluation model through the AdaBoost integrated algorithm, and quantitatively evaluates the safety risks of the PCTSE smart park based on the trained model.

[0024] Compared with the prior art, the present invention has the following beneficial effects: The present invention establishes the basic principles of the safety assessment system based on the PCTSE smart park concept, and through the basic principles, a scientific, reasonable and effective safety assessment framework is jointly formed to provide support for the construction of the park safety assessment model. The present invention collects safety assessment samples of the PCTSE smart park through actual investigation and Internet data, and performs data cleaning and standardization in the preprocessing process of the samples to ensure the quality and applicability of the sample data, and adopts the K-fold cross-validation strategy to reduce the overfitting of the model training and significantly improve the reliability and effectiveness of the model. The present invention optimizes the penalty parameter C and the kernel parameter g of the support vector machine through the marine predator algorithm, which not only improves the performance and accuracy of the model, but also realizes the advantages of automatic tuning, global search, and rapid convergence; it ensures that the support vector machine is more efficient and reliable when establishing a model to process high-dimensional complex data. The present invention trains the park safety assessment model based on the AdaBoost integrated algorithm, and uses the weighted multiple weak classifier function of AdaBoost to effectively improve the accuracy and robustness of the model. Through the iterative mechanism of AdaBoost, each round of iteration can automatically adjust the weights of unimportant features to enhance the learning ability of the model and achieve a more comprehensive safety risk assessment. The trained park safety index evaluation model can be used to accurately and quantitatively evaluate park safety, helping managers to fully understand the park’s safety status and identify potential safety hazards and risk points, thereby improving the overall safety management level of the smart park. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a step diagram of the PCTSE smart park safety risk assessment method of this embodiment 1; Figure 2 This is a step diagram for constructing a park safety assessment model in this embodiment 1; Figure 3This is a flow chart of the PCTSE smart park security risk assessment system in this embodiment 2. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] It should be noted that when a component is referred to as being "mounted on" another component, it may be directly on the other component or there may be a central component. When a component is considered to be "set on" another component, it may be directly set on the other component or there may be a central component at the same time. When a component is considered to be "fixed to" another component, it may be directly fixed on the other component or there may be a central component at the same time.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "or / and" used herein includes any and all combinations of one or more of the related listed items.

[0029] Example 1 See also Figure 1-2 , Figure 1 A PCTSE smart park security risk assessment method according to embodiment 1 of the present invention comprises the following steps: S1: Establish the basic principles of the safety assessment system based on the PCTSE smart park concept, and determine the indicators at all levels in the safety assessment system based on the basic principles.

[0030] The basic principles of the security assessment system include: (1) Comprehensiveness principle: The security assessment system should cover all relevant security factors in the park, including physical security, network security, data security, personnel security, etc., to ensure the comprehensiveness of the assessment; (2) Operability principle: The design of evaluation indicators should be based on accurately reflecting the degree of security risk in the park, and try to select comprehensive indicators with commonalities, and strive to make the data operable; (3) Systematic principle: The indicator system should take into account the overall structure of the park and the relationship between each subsystem to ensure the coordination and consistency between each indicator; (4) Participation principle: In the process of establishing the indicator system, the opinions and suggestions of all stakeholders should be fully considered to ensure the rationality and applicability of the indicators.

[0031] According to the basic principles, the indicators in the safety assessment system include five first-level indicators, namely, park hazard sources, park safety protection equipment, park employees' safety awareness, park emergency handling capabilities and domino effect, as well as several second-level indicators. The second-level indicators selected in this embodiment include: hazard source identification rate, hazard source assessment frequency, hazard source control measures implementation rate, accident hidden danger rectification rate, safety protection equipment integrity rate, safety protection equipment regular maintenance rate, safety protection equipment use training coverage rate, safety protection equipment failure response time, safety protection equipment configuration rate, safety training participation rate, emergency plan formulation rate, emergency drill frequency, emergency response time, emergency material reserve rate, emergency handling success rate, safety knowledge test pass rate, safety hidden danger reporting rate, safety behavior observation pass rate, safety culture activity participation rate, chain reaction identification rate, domino effect assessment frequency, and domino effect control measures implementation rate.

[0032] This embodiment establishes a safety assessment system based on the PCTSE smart park concept. The basic principles emphasize comprehensiveness, operability, systematicness and participation. Through the basic principles, a scientific, reasonable and effective safety assessment framework is formed to provide support for building a park safety assessment model.

[0033] S2: Obtain safety evaluation samples of PCTSE smart parks through actual investigation and collection of Internet information, and pre-process the safety evaluation samples.

[0034] The methods for obtaining safety evaluation samples for PCTSE smart parks include: (1) actually visiting different smart parks to observe their safety management measures, technology applications, and personnel configurations; and (2) consulting relevant academic papers and technical documents on the Internet to collect safety evaluation standards and cases for other PCTSE smart parks.

[0035] Preprocessing includes data cleaning and data standardization. Data cleaning includes identifying missing values ​​in the sample and filling the missing values ​​with the mean. Data standardization processes the data through Z-score standardization. The calculation formula is as follows: In the formula, Represents the original data value, represents the mean of the data set, Represents the standard deviation of the data set.

[0036] The feature vector and true label of the jth sample after preprocessing are expressed as follows: in, represents the feature vector of the jth sample, (i=1, 2, ... n) represents the standard value of the i-th safety index of the sample, Represents the safety level of the jth sample, j=1, 2,…m.

[0037] Several preprocessed sample sets are randomly selected as test sets, and the remaining sample sets are used as training sets; the K-fold cross-validation strategy is adopted, and a part of the training set is taken as the validation set for each training, and the rest is used as the training set for that training.

[0038] This embodiment collects safety evaluation samples of PCTSE smart parks through actual investigation and Internet data, and performs data cleaning and standardization in the sample preprocessing process to ensure the quality and applicability of sample data, providing a reliable basis for subsequent safety evaluation analysis. The present invention adopts the K-fold cross-validation strategy, which can reduce model training overfitting and significantly improve the reliability and effectiveness of the model.

[0039] S3: Determine the evaluation standards and evaluation levels of indicators at all levels in the safety assessment system based on expert scores and various safety standards for park safety assessment. According to the collected safety standards and expert opinions, formulate specific evaluation standards for each indicator; divide the evaluation standards of indicators at all levels in the safety assessment system into level 1, level 2, level 3, level 4, and level 5, with the safety level decreasing from high to low.

[0040] S4: The penalty coefficient and kernel parameter of the support vector machine are optimized by the marine predator algorithm to construct a park safety evaluation model; based on the preprocessed safety evaluation samples, the park safety evaluation model is trained using the Adaboost ensemble algorithm. AdaBoost is a powerful ensemble learning algorithm designed to improve the accuracy of the classifier. It builds a strong classifier by combining multiple weak classifiers.

[0041] The core idea of ​​AdaBoost is to gradually train weak classifiers, each of which focuses on samples that were misclassified by the previous classifier. In each iteration, AdaBoost assigns a weight to each sample, increasing the weight of misclassified samples and decreasing the weight of correctly classified samples. The prediction results of all weak classifiers are weighted and combined according to their accuracy to form the final prediction result.

[0042] The steps to construct a park safety assessment model are as follows: S41: Determine a kernel function suitable for the park safety assessment model based on a support vector machine. In this embodiment, a radial basis kernel function is selected, and the calculation formula of the kernel function is as follows: In the formula, represents any two different sample vectors, Represents the parameters of the kernel function.

[0043] S42: The marine predator algorithm optimizes the hyperparameters of the support vector machine to determine the optimal parameter combination. The steps of optimizing the hyperparameters of the support vector machine using the marine predator algorithm are as follows: A: Set the population size, number of iterations, search space boundary, FaDs function and objective function of the marine predator algorithm. In this embodiment, the population size is set to 30, the number of iterations is set to 50, the penalty parameter C and the kernel parameter g are in the range of [0.01,50], the F1 score is selected as the objective function, and the FaDs function expression is determined as follows: Where a and b are the solution ranges of the dynamic lens imaging reverse learning method on the x-axis, and x represents the current optimal projection position of the individual x-axis. represents the optimal x-axis projection position of the image generated by the individual projection, , Represent the current number of iterations and the maximum number of iterations respectively.

[0044] B: Randomly generate the initial population, each individual in the population represents a set of penalty parameters C and kernel parameters g of the support vector machine.

[0045] C: Calculate the fitness values ​​of all individuals to determine whether the preset end condition is met; if not, use the dynamic lens imaging reverse learning method to select the best individual; update the position of the best individual according to the marine predator algorithm; In this embodiment, the fitness value of the individual is calculated by using the F1 score as the objective function. The calculation formula of the F1 score is as follows: In the formula, TP represents the number of samples correctly predicted as positive by the model, FP represents the number of samples incorrectly predicted as positive by the model, and FN represents the number of samples correctly predicted as negative by the model.

[0046] D: Calculate the fitness value of the optimal individual; determine whether the preset end condition is met. If not, return to step C until the calculation ends; output the optimized penalty parameter C and kernel parameter g.

[0047] This embodiment optimizes the penalty parameter C and kernel parameter g of the support vector machine through the marine predator algorithm, which not only improves the performance and accuracy of the model, but also achieves the advantages of automated tuning, global search, and rapid convergence; ensuring that the support vector machine is more efficient and reliable when establishing a model to process high-dimensional complex data.

[0048] S43: Establish a park safety evaluation model through a support vector machine based on optimal hyperparameters. The objective function of the support vector machine is as follows: The decision function of the support vector machine is as follows: The constraints of the support vector machine are as follows Where w represents the normal vector of the hyperplane, represents the slack variable, represents the fixed value of the constraint bias term, represents the feature vector of the sample, represents the true label of the sample, represents the Lagrange multiplier.

[0049] S44: Based on the preprocessed security evaluation samples, the park security evaluation model is trained based on the AdaBoost ensemble algorithm, and the performance of the trained model is evaluated. The steps for training the park security evaluation model based on the AdaBoost ensemble algorithm are as follows: E: Select N data set samples D= , determines the weight of the initial iteration of each sample and the maximum number of iterations T; this embodiment sets the weight of the initial iteration =1 / N, the maximum number of iterations T=30.

[0050] F: Based on the weighted training set, train the park safety assessment model and calculate the error rate of the tth round of model training and learning weights ; Error rate of model training The calculation formula is as follows: In the formula, represents the weight of the tth iteration, I(x) represents the discriminant about x, if x is correct, I takes the value 1, otherwise I takes the value 0.

[0051] The calculation formula for the learning weights of model training is as follows: G: Pass error rate and learning weights Update the sample weights for the next round of training, and normalize the updated sample weights to ensure that their sum is 1; the calculation formula for updating the training sample weights in this embodiment is as follows: In the formula, It represents the predicted value of the model after training the park safety evaluation model t times, and C represents the normalization factor.

[0052] H: Return to step F until the maximum number of iterations is reached, and output the prediction result of the final model. In this embodiment, the final model prediction result S is as follows: This example tests the prediction performance of the trained model using the F1 score, with the threshold set to 0.5.

[0053] This embodiment uses the AdaBoost integrated algorithm to train the park security assessment model, and uses the weighted multiple weak classifier function of AdaBoost to effectively improve the accuracy and robustness of the model. Through the iterative mechanism of AdaBoost, each round of iteration can automatically adjust the weights of unimportant features to enhance the learning ability of the model, achieve a more comprehensive security risk assessment, and improve the applicability and credibility of the model.

[0054] S5: Use the trained park security assessment model to perform a quantitative assessment of PCTSE smart park security risks. The standards for the quantitative assessment of PCTSE smart park security risks in this embodiment are as follows: This embodiment establishes the basic principles of the safety assessment system based on the PCTSE smart park concept, and through the basic principles, a scientific, reasonable and effective safety assessment framework is formed to provide support for the construction of the park safety assessment model. This embodiment collects safety assessment samples of the PCTSE smart park through actual investigation and Internet data, and performs data cleaning and standardization in the preprocessing process of the samples to ensure the quality and applicability of the sample data, and adopts the K-fold cross-validation strategy to reduce the overfitting of the model training and significantly improve the reliability and effectiveness of the model. This embodiment optimizes the penalty parameter C and kernel parameter g of the support vector machine through the marine predator algorithm, which not only improves the performance and accuracy of the model, but also realizes the advantages of automatic tuning, global search, and rapid convergence; ensuring that the support vector machine is more efficient and reliable when establishing a model to process high-dimensional complex data. This embodiment trains the park safety assessment model based on the AdaBoost integrated algorithm, and uses the weighted multiple weak classifier function of AdaBoost to effectively improve the accuracy and robustness of the model. Through the iterative mechanism of AdaBoost, each round of iteration can automatically adjust the weights of unimportant features to enhance the learning ability of the model and achieve a more comprehensive safety risk assessment. The trained park safety index evaluation model can be used to accurately and quantitatively evaluate park safety, helping managers to fully understand the park’s safety status and identify potential safety hazards and risk points, thereby improving the overall safety management level of the smart park.

[0055] Example 2: Please refer to Figure 3 , this embodiment provides a PCTSE smart park security risk assessment system, including: Information collection module: It is used to obtain safety evaluation samples of PCTSE smart parks based on actual investigations and collection of Internet information, pre-process the safety evaluation samples, and send the pre-processed safety evaluation samples to the processing module.

[0056] Processing module: The processing module includes a hyperparameter optimization unit, a park safety evaluation unit and an output unit. The hyperparameter optimization unit is used to find the optimal combination value of the penalty coefficient and kernel parameter in the support vector machine through the marine predator algorithm; the park safety evaluation unit constructs a park safety evaluation model based on the support vector machine and the optimal parameter combination value output by the hyperparameter optimization unit; the output unit trains the park safety evaluation model through the AdaBoost integrated algorithm, and quantitatively evaluates the safety risks of the PCTSE smart park based on the trained model.

[0057] The above embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for those of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A PCTSE smart park safety risk assessment method, characterized in that: The following steps are involved: S1: Establish the basic principles of the safety assessment system based on the PCTSE smart park concept, and determine the indicators at all levels in the safety assessment system based on the basic principles; S2: Obtain safety evaluation samples of PCTSE smart parks through actual investigation and collection of Internet data, and pre-process the safety evaluation samples; S3: Determine the evaluation standards and evaluation levels of indicators at all levels in the safety assessment system based on expert scores and various safety standards for park safety assessment; S4: Optimize the penalty coefficient and kernel parameter of the support vector machine through the marine predator algorithm to build a park safety evaluation model; According to the pre-processed safety evaluation samples, the Adaboost integrated algorithm is used to train the park safety evaluation model; S5: Use the trained park safety evaluation model to conduct quantitative evaluation of PCTSE smart park safety risks.

2. The PCTSE smart park safety risk assessment method according to claim 1 is characterized in that: In step S1, the basic principles of the security assessment system are established based on the PCTSE smart park concept. The basic principles include: (1) Comprehensiveness principle: The security assessment system covers all relevant security factors in the park, including physical security, network security, data security, and personnel security, to ensure the comprehensiveness of the assessment; (2) Operability principle: The design of evaluation indicators should be based on accurately reflecting the degree of security risk in the park, selecting common comprehensive indicators, and ensuring the operability of the data; (3) Systematic principle: The indicator system considers the overall structure of the park and the relationship between each subsystem to ensure the coordination and consistency between each indicator; (4) Participation principle: In the process of establishing the indicator system, the opinions and suggestions of all stakeholders are fully considered to ensure the rationality and applicability of the indicators; According to the basic principles, the indicators in the safety assessment system include five first-level indicators, namely, park hazards, park safety protection equipment, park employees' safety awareness, park emergency response capabilities and domino effect, as well as several second-level indicators.

3. The PCTSE smart park safety risk assessment method according to claim 1 is characterized in that: In step S2, the methods for obtaining safety evaluation samples of PCTSE smart parks include: (1) actually visiting different smart parks to observe their safety management measures, technology applications, and personnel configuration; (2) consulting relevant academic papers and technical documents on the Internet to collect safety evaluation standards and cases of other PCTSE smart parks; The preprocessing includes data cleaning and data standardization. The data cleaning includes identifying missing values ​​in the sample and filling the missing values ​​with the mean. The data standardization processes the data through Z-score standardization, and the calculation formula is as follows: In the formula, Represents the original data value, represents the mean of the data set, Represents the standard deviation of the data set; The feature vector and true label of the jth sample after preprocessing are expressed as follows: in, represents the feature vector of the jth sample, (i=1, 2, ... n) represents the standard value of the i-th safety index of the sample, represents the safety level of the jth sample, j=1, 2, ...m; Several preprocessed sample sets are randomly selected as test sets, and the remaining sample sets are used as training sets; a K-fold cross-validation strategy is adopted, and a part of the training set is taken as the validation set for each training, and the rest is taken as the training set for this training.

4. The PCTSE smart park safety risk assessment method according to claim 1 is characterized in that: In step S3, specific evaluation criteria are formulated for each indicator based on the collected safety specifications and expert opinions; the evaluation criteria for indicators at all levels in the safety assessment system are divided into level one, level two, level three, level four, and level five, with the safety level decreasing in descending order.

5. The PCTSE smart park safety risk assessment method according to claim 1 is characterized in that: In step S4, the steps of constructing the park safety assessment model are as follows: S41: Determine a kernel function suitable for a park safety assessment model based on a support vector machine; S42: Optimize the hyperparameters of the support vector machine through the marine predator algorithm to determine the optimal parameter combination; S43: Establishing a park safety evaluation model through a support vector machine based on optimal hyperparameters; S44: Based on the preprocessed security evaluation samples, the campus security evaluation model is trained based on the AdaBoost ensemble algorithm, and the performance of the trained model is evaluated.

6. The PCTSE smart park safety risk assessment method according to claim 5 is characterized in that: In step S41, the calculation formula of the kernel function is as follows: In the formula, represents any two different sample vectors, Represents the parameters of the kernel function.

7. The PCTSE smart park safety risk assessment method according to claim 5 is characterized in that: In step S42, the steps of optimizing the hyperparameters of the support vector machine by using the marine predator algorithm are as follows: A: Set the population size, number of iterations, search space boundary, FaDs function and objective function of the marine predator algorithm; B: Randomly generate the initial population, each individual in the population represents a set of penalty parameters C and kernel parameters g of the support vector machine; C: Calculate the fitness values ​​of all individuals to determine whether the preset end conditions are met; If it is not satisfied, the dynamic lens imaging reverse learning method is used to select the best individual; the position of the best individual is updated according to the marine predator algorithm; D: Calculate the fitness value of the optimal individual; determine whether the preset end condition is met. If not, return to step C until the calculation ends; output the optimized penalty parameter C and kernel parameter g.

8. The PCTSE smart park safety risk assessment method according to claim 5 is characterized in that: In step S43, the objective function of the support vector machine is as follows: The decision function of the support vector machine is as follows: The constraints of the support vector machine are as follows Where w represents the normal vector of the hyperplane, represents the slack variable, represents the fixed value of the constraint bias term, represents the feature vector of the sample, represents the true label of the sample, represents the Lagrange multiplier.

9. The PCTSE smart park safety risk assessment method according to claim 5 is characterized in that: In step S44, the steps of training the park safety evaluation model based on the AdaBoost integrated algorithm are as follows: E: Select N data set samples D= , determines the weight of the initial iteration of each sample and the maximum number of iterations T; F: Based on the weighted training set, train the park safety assessment model and calculate the error rate of the tth round of model training and learning weights ; G: Pass error rate and learning weights Update the sample weights for the next round of training, and normalize the updated sample weights to ensure that their sum is 1; H: Return to step F until the maximum number of iterations is reached, and output the prediction results of the final model.

10. A PCTSE smart park safety risk assessment system, which adopts the PCTSE smart park safety risk assessment method as described in any one of claims 1-9, characterized in that: include: Information collection module and processing module; The information collection module is used to obtain safety evaluation samples of the PCTSE smart park based on actual investigations and collection of Internet data, pre-process the safety evaluation samples, and send the pre-processed safety evaluation samples to the processing module; The processing module includes a hyperparameter optimization unit, a park safety evaluation unit and an output unit. The hyperparameter optimization unit is used to find the optimal combination value of the penalty coefficient and the kernel parameter in the support vector machine through the marine predator algorithm; the park safety evaluation unit constructs a park safety evaluation model based on the support vector machine and the optimal parameter combination value output by the hyperparameter optimization unit; the output unit trains the park safety evaluation model through the AdaBoost integrated algorithm, and quantitatively evaluates the safety risks of the PCTSE smart park based on the trained model.

Images