Entertainment equipment safety risk assessment and monitoring system based on artificial intelligence

Abstract

The invention relates to the technical field of entertainment equipment safety risk assessment, in particular to an entertainment equipment safety risk assessment and monitoring system based on artificial intelligence. The system comprises a first risk assessment module, a second risk assessment module and a monitoring module. The first risk assessment module is used for acquiring a data set D of the entertainment equipment, performing tree splitting on the data set D by using a gain of a structural component, forming a plurality of trees in a residual error fitting manner, and performing logistic regression on each leaf node of each tree to obtain N logistic regression models, N being the number of information of the entertainment equipment, and N being the number of information of the entertainment equipment; each piece of entertainment equipment corresponds to one piece of entertainment equipment information; the second risk assessment module is used for substituting the N logistic regression models into an Adaboost algorithm to obtain a risk assessment result; and the monitoring module is used for analyzing the risk assessment result to obtain an analysis result, and performing early warning according to the analysis result. According to the invention, the accuracy of security risk assessment of the entertainment equipment and the efficiency of management of the entertainment equipment can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of amusement equipment safety risk assessment, and in particular to an amusement equipment safety risk assessment and monitoring system based on artificial intelligence. Background Art

[0002] With the rapid development of science and technology, artificial intelligence technology has been widely used in various fields, bringing unprecedented changes to modern society. In the amusement equipment industry, the application of AI technology is also gradually increasing, especially in the aspects of safety risk assessment and monitoring, AI technology plays an increasingly important role.

[0003] Traditional amusement equipment safety risk assessment and monitoring mainly rely on manual inspections and regular maintenance, which is not only inefficient but also prone to omissions and misjudgments due to human factors. In addition, the state of amusement equipment changes in real time during operation, and traditional monitoring methods are difficult to respond in real time and make accurate predictions.

[0004] In order to solve the above problems, there is an urgent need for an amusement equipment safety risk assessment and monitoring system based on artificial intelligence. Summary of the invention

[0005] The purpose of the present invention is to provide an artificial intelligence-based amusement equipment safety risk assessment and monitoring system: at least to solve the problems of low safety risk assessment accuracy and difficulty in real-time response to amusement equipment monitoring caused by relying on manual inspections and regular maintenance to conduct safety risk assessments on amusement equipment in existing solutions.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] An artificial intelligence-based amusement equipment safety risk assessment and monitoring system, the system comprising a first risk assessment module, a second risk assessment module and a monitoring module;

[0008] The first risk assessment module is used to obtain a data set D of amusement equipment, and use the gain of the structural score to split the tree on the data set D, and form multiple trees by fitting the residuals, and use logistic regression for each leaf node of each tree to obtain N logistic regression models, where N is the number of amusement equipment information, and each amusement equipment corresponds to one amusement equipment information;

[0009] The second risk assessment module is used to substitute N logistic regression models into the Adaboost algorithm to obtain risk assessment results;

[0010] The monitoring module is used to analyze the risk assessment results, obtain the analysis results, and issue early warnings based on the analysis results.

[0011] Furthermore, before obtaining the data set D of the amusement equipment, the system preprocesses the amusement equipment information to generate the data set D:

[0012] Features of amusement equipment information are selected through data cleaning and principal component analysis. Special labels are used as markers for some missing values ​​of amusement equipment information, and the average value is used to supplement a small amount of missing amusement equipment information. For label data, features with more than 80% missing are deleted, and four features are obtained: equipment manufacturing level, equipment maintenance frequency, equipment insurance type, and equipment maintenance personnel experience level. The data types of equipment manufacturing level, equipment maintenance frequency, equipment insurance type, and equipment maintenance personnel experience level are defined: equipment manufacturing level and equipment maintenance personnel experience level are percentage type, equipment maintenance frequency is floating point type, and equipment insurance type is positive integer type. Based on the defined data types, data set D is generated.

[0013] Furthermore, the dataset D uses the gain of the structural score to split the tree, and forms multiple trees by fitting the residuals. Logistic regression is used for each leaf node of each tree to obtain N logistic regression models, which specifically include the following process:

[0014] K trees are obtained by iterative training of the training data set D;

[0015] Taking Taylor’s formula as the approximation of the objective function, the objective function for the t-th tree is obtained by the second-order expansion of Taylor’s formula, where t∈K:

[0016] The cross entropy loss function is used as the loss function l, and the objective function L is obtained based on the loss function l:

[0017]

[0018] in, Among them, μ and γ are preset parameters, and T is the number of leaf nodes;

[0019] Calculate the first-order gradient g of the loss function of the i-th sample in the t-th tree i :

[0020]

[0021] Among them, l(y i ,y t-1 ) represents the value of y i ,y t-1 The function expression of f k (x i ) is the weight of the i-th sample in the k-th tree, y t-1 Represents the reference value of the t-1th tree;

[0022] Calculate the second-order gradient h of the loss function of the i-th sample in the t-th tree i :

[0023]

[0024] The first-order gradient g i , second-order gradient h i Substituting the objective function L into the final objective function:

[0025]

[0026] For the final objective function ω j The derivative of gets the structure score of the tree, using G j replace H j replace

[0027]

[0028] The difference in structure score is used to evaluate the gain of the tree before and after the split. The tree will split in a way that increases the structure score the most. The formula for calculating the structure score gain is as follows:

[0029]

[0030] Among them, G L , G R Represents the sum of the first-order derivatives of the left and right subtrees after splitting, H L , H R Represents the sum of the second-order derivatives of the left and right subtrees after splitting;

[0031] Split K trees in a way that maximizes Gain to obtain N logistic regression models.

[0032] Furthermore, the second risk assessment module is used to substitute N logistic regression models into the Adaboost algorithm to obtain the risk assessment result, which specifically includes the following process:

[0033] For the Sth logistic regression model, an initial weight is set, the selected weak classifier set is X, the current iteration number V is obtained, and the weak classifier with the lowest error rate is selected from the weak classifier set X as the Vth basic classifier. The error value of the weight distribution of the training samples of the weak classifier in the round of iteration number V training is obtained. The error value is the sum of the weights of the samples misclassified by the weak classifier e v , the weight β of the weak classifier in the strong classifier is calculated by the error value, where the expression of β is: Update the weights of all logistic regression models according to the weight β and the initial weight, and compare the risk assessment result list according to the weight of the logistic regression model to obtain the risk assessment result, where the risk assessment result list is a preset set of risk assessment results, each risk assessment result corresponds to a weight interval, and the risk assessment result is determined based on the weight and the weight interval.

[0034] Furthermore, splitting the tree in a way that maximizes Gain to obtain N logistic regression models specifically includes the following process:

[0035] For the tth tree, traverse all features and calculate the splitting threshold corresponding to the maximum Gain value in the current feature, and determine whether the feature has been used as a split feature in the ancestor node of the node. If it has been used, select the feature corresponding to the largest Gain value other than this feature as the split feature. When the maximum depth of each tree is reached or there is no split that can obtain a larger structural score, the growth of the tree is terminated, the residual of the tree is input into the next tree, and the samples on each leaf node are trained for logistic regression to obtain N logistic regression models.

[0036] Furthermore, the monitoring module is used to analyze the risk assessment results, obtain the analysis results, and issue an early warning based on the analysis results. Specifically, the following processes are included:

[0037] The monitoring module receives the risk assessment results and determines the safety status of the current amusement equipment by comparing historical data, industry standards and expert opinions, and identifies the existing risk points. For high-risk points, the early warning mechanism is triggered to alert managers and operators through sound and light alarms and voice prompts.

[0038] Compared with the existing solutions, the present invention achieves the following beneficial effects:

[0039] The system of the present invention obtains a data set D of amusement equipment, and uses the gain of the structural score to split the tree for the data set D, and forms multiple trees by fitting residuals, and uses logistic regression for each leaf node of each tree to obtain N logistic regression models, wherein N is the number of amusement equipment information, and each amusement equipment corresponds to one amusement equipment information; the N logistic regression models are substituted into the Adaboost algorithm to obtain risk assessment results; the risk assessment results are analyzed to obtain analysis results, and early warnings are issued based on the analysis results.

[0040] On the one hand, using deep learning algorithms to process and analyze amusement equipment information can get rid of the safety risks brought by manual inspections and regular maintenance of amusement equipment safety risk assessments, and improve the accuracy of safety risk assessments.

[0041] On the other hand, the present invention can monitor and accurately predict amusement equipment through real-time monitoring, greatly improving the efficiency of safety management. The system can automatically identify and warn of potential safety hazards, reducing omissions and misjudgments caused by human factors; finally, the system can also provide data support for managers to help them formulate more scientific and reasonable maintenance and management strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0043] Figure 1 It is a system block diagram of an artificial intelligence-based amusement equipment safety risk assessment and monitoring system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0044] 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.

[0045] In addition, the described features, structures or characteristics may be combined in one or more example embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. may be adopted. In other cases, well-known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0046] This embodiment provides an amusement equipment safety risk assessment and monitoring system based on artificial intelligence. Figure 1 is a system block diagram of an artificial intelligence-based amusement equipment safety risk assessment and monitoring system according to an embodiment of the present invention, such as Figure 1 As shown, the system includes a first risk assessment module, a second risk assessment module and a monitoring module;

[0047] The first risk assessment module is used to obtain a data set D of amusement equipment, and use the gain of the structural score to split the tree on the data set D, and form multiple trees by fitting the residuals, and use logistic regression for each leaf node of each tree to obtain N logistic regression models, where N is the number of amusement equipment information, and each amusement equipment corresponds to one amusement equipment information;

[0048] The second risk assessment module is used to substitute N logistic regression models into the Adaboost algorithm to obtain risk assessment results;

[0049] The monitoring module is used to analyze the risk assessment results, obtain the analysis results, and issue early warnings based on the analysis results.

[0050] In summary, the system of the present invention obtains a data set D of amusement equipment, and uses the gain of the structural score to split the tree for the data set D, and forms multiple trees by fitting residuals, and uses logistic regression for each leaf node of each tree to obtain N logistic regression models, wherein N is the number of amusement equipment information, and each amusement equipment corresponds to one amusement equipment information; the N logistic regression models are substituted into the Adaboost algorithm to obtain risk assessment results; the risk assessment results are analyzed to obtain analysis results, and warnings are issued based on the analysis results.

[0051] On the one hand, using deep learning algorithms to process and analyze amusement equipment information can get rid of the safety risks brought by manual inspections and regular maintenance of amusement equipment safety risk assessments, and improve the accuracy of safety risk assessments.

[0052] On the other hand, the present invention can monitor and accurately predict amusement equipment through real-time monitoring, greatly improving the efficiency of safety management. The system can automatically identify and warn of potential safety hazards, reducing omissions and misjudgments caused by human factors; finally, the system can also provide data support for managers to help them formulate more scientific and reasonable maintenance and management strategies.

[0053] In some embodiments, before obtaining the data set D of the amusement device, the system preprocesses the amusement device information to generate the data set D:

[0054] Features of amusement equipment information are selected through data cleaning and principal component analysis. Special labels are used as markers for some missing values ​​of amusement equipment information, and the average value is used to supplement a small amount of missing amusement equipment information. For label data, features with more than 80% missing are deleted, and four features are obtained: equipment manufacturing level, equipment maintenance frequency, equipment insurance type, and equipment maintenance personnel experience level. The data types of equipment manufacturing level, equipment maintenance frequency, equipment insurance type, and equipment maintenance personnel experience level are defined: equipment manufacturing level and equipment maintenance personnel experience level are percentage type, equipment maintenance frequency is floating point type, and equipment insurance type is positive integer type. Based on the defined data types, data set D is generated.

[0055] Furthermore, the dataset D uses the gain of the structural score to split the tree, and forms multiple trees by fitting the residuals. Logistic regression is used for each leaf node of each tree to obtain N logistic regression models, which specifically include the following process:

[0056] K trees are obtained by iterative training of the training data set D;

[0057] Taking Taylor’s formula as the approximation of the objective function, the objective function for the t-th tree is obtained by the second-order expansion of Taylor’s formula, where t∈K:

[0058] The cross entropy loss function is used as the loss function l, and the objective function L is obtained based on the loss function l:

[0059]

[0060] in, Among them, μ and γ are preset parameters, and T is the number of leaf nodes;

[0061] Calculate the first-order gradient g of the loss function of the i-th sample in the t-th tree i :

[0062]

[0063] Among them, l(y i ,y t-1 ) represents the value of y i ,y t-1 The function expression of f k (x i ) is the weight of the i-th sample in the k-th tree, y t-1 Represents the reference value of the t-1th tree;

[0064] Calculate the second-order gradient h of the loss function of the i-th sample in the t-th tree i :

[0065]

[0066] The first-order gradient g i , second-order gradient h i Substituting the objective function L into the final objective function:

[0067]

[0068] For the final objective function ω j The derivative of G is used to get the structure of the tree. j replace H j replace

[0069]

[0070] The difference in structure score is used to evaluate the gain of the tree before and after the split. The tree will split in a way that increases the structure score the most. The formula for calculating the structure score gain is as follows:

[0071]

[0072] Among them, G L , G R Represents the sum of the first-order derivatives of the left and right subtrees after splitting, H L , H R Represents the sum of the second-order derivatives of the left and right subtrees after splitting;

[0073] Split K trees in a way that maximizes Gain to obtain N logistic regression models:

[0074] For the tth tree, traverse all features and calculate the splitting threshold corresponding to the maximum Gain value in the current feature, and determine whether the feature has been used as a split feature in the ancestor node of the node. If it has been used, select the feature corresponding to the largest Gain value other than this feature as the split feature. When the maximum depth of each tree is reached or there is no split that can obtain a larger structural score, the growth of the tree is terminated, the residual of the tree is input into the next tree, and the samples on each leaf node are trained for logistic regression to obtain N logistic regression models.

[0075] In some embodiments, the second risk assessment module is used to substitute N logistic regression models into the Adaboost algorithm to obtain a risk assessment result, specifically including the following process:

[0076] For the Sth logistic regression model, an initial weight is set, the selected weak classifier set is X, the current iteration number V is obtained, and the weak classifier with the lowest error rate is selected from the weak classifier set X as the Vth basic classifier. The error value of the weight distribution of the training samples of the weak classifier in the round of iteration number V training is obtained. The error value is the sum of the weights of the samples misclassified by the weak classifier e v , the weight β of the weak classifier in the strong classifier is calculated by the error value, where the expression of β is: Update the weights of all logistic regression models according to the weight β and the initial weight, and compare the risk assessment result list according to the weight of the logistic regression model to obtain the risk assessment result, where the risk assessment result list is a preset set of risk assessment results, each risk assessment result corresponds to a weight interval, and the risk assessment result is determined based on the weight and the weight interval.

[0077] The monitoring module is used to analyze the risk assessment results, obtain the analysis results, and issue warnings based on the analysis results. The specific process also includes the following:

[0078] The monitoring module receives the risk assessment results and determines the safety status of the current amusement equipment by comparing historical data, industry standards and expert opinions, and identifies the existing risk points. For high-risk points, the early warning mechanism is triggered to alert managers and operators through sound and light alarms and voice prompts.

[0079] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0080] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0081] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0082] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0083] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0084] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0085] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. The security risk assessment and monitoring system of amusement equipment based on artificial intelligence is characterized by: The system includes a first risk assessment module, a second risk assessment module and a monitoring module; The first risk assessment module is used to obtain a data set D of amusement equipment, and use the gain of the structural score to split the tree on the data set D, and form multiple trees by fitting the residuals, and use logistic regression for each leaf node of each tree to obtain N logistic regression models, where N is the number of amusement equipment information, and each amusement equipment corresponds to one amusement equipment information; The second risk assessment module is used to substitute N logistic regression models into the Adaboost algorithm to obtain risk assessment results; The monitoring module is used to analyze the risk assessment results, obtain the analysis results, and issue early warnings based on the analysis results.

2. The artificial intelligence-based amusement equipment safety risk assessment and monitoring system according to claim 1 is characterized in that: Before obtaining the dataset D of the amusement equipment, the system preprocesses the amusement equipment information to generate the dataset D: Features of amusement equipment information are selected through data cleaning and principal component analysis. Special labels are used as markers for some missing values ​​of amusement equipment information, and the average value is used to supplement a small amount of missing amusement equipment information. For label data, features with more than 80% missing are deleted, and four features are obtained: equipment manufacturing level, equipment maintenance frequency, equipment insurance type, and equipment maintenance personnel experience level. The data types of equipment manufacturing level, equipment maintenance frequency, equipment insurance type, and equipment maintenance personnel experience level are defined: equipment manufacturing level and equipment maintenance personnel experience level are percentage type, equipment maintenance frequency is floating point type, and equipment insurance type is positive integer type. Based on the defined data types, data set D is generated.

3. The artificial intelligence-based amusement equipment safety risk assessment and monitoring system according to claim 2 is characterized in that: Dataset D uses the gain of structural score to split the tree, and forms multiple trees by fitting residuals. Logistic regression is used for each leaf node of each tree to obtain N logistic regression models. The specific process includes the following: K trees are obtained by iterative training of the training data set D; Taking Taylor’s formula as the approximation of the objective function, the objective function for the t-th tree is obtained by the second-order expansion of Taylor’s formula, where t∈K: The cross entropy loss function is used as the loss function l, and the objective function L is obtained based on the loss function l: in, Among them, μ and γ are preset parameters, and T is the number of leaf nodes; Calculate the first-order gradient g of the loss function of the i-th sample in the t-th tree i : Among them, l(y i ,y t-1 0 means about y i ,y t-1 The function expression of f k (x i ) is the weight of the i-th sample in the k-th tree, y t-1 Represents the reference value of the t-1th tree; Calculate the second-order gradient h of the loss function of the i-th sample in the t-th tree i : The first-order gradient g i , second-order gradient h i Substituting the objective function L into the final objective function: For the final objective function ω j The derivative of G is used to get the structure of the tree. j replace H j replace The difference in structure score is used to evaluate the gain of the tree before and after the split. The tree will split in a way that increases the structure score the most. The formula for calculating the structure score gain is as follows: Among them, G L , G R Represents the sum of the first-order derivatives of the left and right subtrees after splitting, H L , H R Represents the sum of the second-order derivatives of the left and right subtrees after splitting; Split K trees in a way that maximizes Gain to obtain N logistic regression models.

4. The artificial intelligence-based amusement equipment safety risk assessment and monitoring system according to claim 3 is characterized in that: The second risk assessment module is used to substitute N logistic regression models into the Adaboost algorithm to obtain risk assessment results. The process includes: For the Sth logistic regression model, an initial weight is set, the selected weak classifier set is X, the current iteration number V is obtained, and the weak classifier with the lowest error rate is selected from the weak classifier set X as the Vth basic classifier. The error value of the weight distribution of the training samples of the weak classifier in the round of iteration number V training is obtained. The error value is the sum of the weights of the samples misclassified by the weak classifier e v , the weight β of the weak classifier in the strong classifier is calculated by the error value, where the expression of β is: Update the weights of all logistic regression models according to the weight β and the initial weight, and compare the risk assessment result list according to the weight of the logistic regression model to obtain the risk assessment result, where the risk assessment result list is a preset set of risk assessment results, each risk assessment result corresponds to a weight interval, and the risk assessment result is determined based on the weight and the weight interval.

5. The artificial intelligence-based amusement equipment safety risk assessment and monitoring system according to claim 3 is characterized in that: The specific process of splitting the tree to obtain N logistic regression models in a way that maximizes Gain is as follows: For the tth tree, traverse all features and calculate the splitting threshold corresponding to the maximum Gain value in the current feature, and determine whether the feature has been used as a split feature in the ancestor node of the node. If it has been used, select the feature corresponding to the largest Gain value other than this feature as the split feature. When the maximum depth of each tree is reached or there is no split that can obtain a larger structural score, the growth of the tree is terminated, the residual of the tree is input into the next tree, and the samples on each leaf node are trained for logistic regression to obtain N logistic regression models.

6. The artificial intelligence-based amusement equipment safety risk assessment and monitoring system according to claim 1 is characterized in that: The monitoring module is used to analyze the risk assessment results, obtain the analysis results, and issue warnings based on the analysis results. The process includes: The monitoring module receives the risk assessment results and determines the safety status of the current amusement equipment by comparing historical data, industry standards and expert opinions, and identifies the existing risk points. For high-risk points, the early warning mechanism is triggered to alert managers and operators through sound and light alarms and voice prompts.

Images