CN112069726A - Risk analysis and evaluation method and device based on Bayesian network - Google Patents
Risk analysis and evaluation method and device based on Bayesian network Download PDFInfo
- Publication number
- CN112069726A CN112069726A CN202010839318.4A CN202010839318A CN112069726A CN 112069726 A CN112069726 A CN 112069726A CN 202010839318 A CN202010839318 A CN 202010839318A CN 112069726 A CN112069726 A CN 112069726A
- Authority
- CN
- China
- Prior art keywords
- bayesian network
- network model
- risk
- probability
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/08—Probabilistic or stochastic CAD
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computational Mathematics (AREA)
- Algebra (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Mathematical Analysis (AREA)
- Probability & Statistics with Applications (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The application discloses a Bayesian network-based risk analysis and assessment method and device, wherein the method comprises the following steps: determining risk factors of a system to be evaluated, and constructing a Bayesian network model for risk evaluation according to the risk factors; acquiring operation data of the system to be evaluated, and optimizing the Bayesian network model according to the operation data and/or preset expert experience data to obtain an optimized Bayesian network model; and acquiring risk factor data of the system to be evaluated, and performing online risk evaluation on the system to be evaluated according to the risk factor data and the optimized Bayesian network model to obtain a risk evaluation result. The risk assessment method and the risk assessment system solve the technical problems that in the prior art, the process of the risk assessment method is complex and the applicability is poor.
Description
Technical Field
The present application relates to the field of risk assessment technologies, and in particular, to a risk analysis and assessment method and apparatus based on a bayesian network.
Background
The system is generally formed by combining a plurality of subsystems, the plurality of subsystems inevitably have faults in the operation process, and when any subsystem appears, the performance of the whole system is influenced, for example, the influence on a risk sensitive system is particularly obvious, so that risk assessment on the system is one of the vital auxiliary means for ensuring the performance of the system.
At present, the existing risk assessment method comprises the following processes: determining a risk research object, constructing a risk index system, and establishing a risk model for risk evaluation; for a complex system, the incidence relation of each hierarchy of an evaluation object is complex, and monitoring parameters are mutually cross-linked, so that the existing risk evaluation system has a complex modeling process, is difficult to represent the propagation relation of risks among the hierarchies, and has poor applicability.
Disclosure of Invention
The technical problem that this application was solved is: aiming at the problems that the process of a risk assessment method in the prior art is complex, the propagation relation of risks among various levels is difficult to characterize, and the applicability is poor, a risk analysis assessment method and a risk analysis assessment device based on a Bayesian network are provided, in the scheme provided by the embodiment of the application, on one hand, the complexity of risk assessment is reduced because the modeling of a Bayesian network model is simple and convenient, and the risk assessment is carried out through the Bayesian network model; on the other hand, the Bayesian network model is optimized through operation data and/or preset expert experience data, and risk assessment is carried out according to the optimized Bayesian network model, namely learning adjustment and parameter learning adjustment can be carried out on the Bayesian network model structure, so that a risk assessment model which is most suitable for practical conditions is obtained, and the applicability of the risk assessment method is further improved.
In a first aspect, an embodiment of the present application provides a risk analysis and assessment method based on a bayesian network, where the method includes:
determining risk factors of a system to be evaluated, and constructing a Bayesian network model for risk evaluation according to the risk factors;
acquiring operation data of the system to be evaluated, and optimizing the Bayesian network model according to the operation data and/or preset expert experience data to obtain an optimized Bayesian network model;
and acquiring risk factor data of the system to be evaluated, and performing online risk evaluation on the system to be evaluated according to the risk factor data and the optimized Bayesian network model to obtain a risk evaluation result.
In the scheme provided by the embodiment of the application, a risk factor of a system to be evaluated is determined, a Bayesian network model for risk evaluation is constructed according to the risk factor, operation data of the system to be evaluated is obtained, the Bayesian network model is optimized according to the operation data and/or preset expert experience data, an optimized Bayesian network model is obtained, risk factor data of the system to be evaluated is obtained, and online risk evaluation is performed on the system to be evaluated according to the risk factor data and the optimized Bayesian network model, so that a risk evaluation result is obtained. Therefore, in the scheme provided by the embodiment of the application, on one hand, because the Bayesian network model is simple and convenient to model, the risk assessment is carried out through the Bayesian network model, and the complexity of the risk assessment is reduced; on the other hand, the Bayesian network model is optimized through operation data and/or preset expert experience data, and risk assessment is carried out according to the optimized Bayesian network model, namely learning adjustment and parameter learning adjustment can be carried out on the Bayesian network model structure, so that a risk assessment model which is most suitable for practical conditions is obtained, and the applicability of the risk assessment method is further improved.
Optionally, constructing a bayesian network model for risk assessment from the risk factors comprises:
determining an incidence relation between any two risk factors, and setting conditional probabilities of any two risk factors;
and constructing the Bayesian network model by taking the risk factors as nodes of the Bayesian network model and the conditional probability as the mutual influence degree of any two risk factors.
Optionally, optimizing the bayesian network model according to the operating data and/or preset expert experience data to obtain an optimized bayesian network model, including:
optimizing the structure and parameters of the Bayesian network model according to a preset iterative algorithm to obtain a structurally optimized Bayesian network model;
determining a parameter which enables a likelihood function value to be maximum from the Bayesian network model after the structure optimization according to a preset maximum likelihood estimation algorithm;
determining prior probability corresponding to the parameters, and performing conditional probability learning according to the prior probability and the operation data to obtain posterior probability corresponding to the parameters;
and adjusting the conditional probability corresponding to each node in the Bayesian network model after the structure optimization according to the posterior probability and/or the preset expert experience data to obtain the optimized Bayesian network model.
Optionally, the optimizing the structure and the parameters of the bayesian network model according to a preset iterative algorithm to obtain a structure-optimized bayesian network model includes:
iteration is carried out from any one Bayesian network in the Bayesian network model, and the current optimal Bayesian network is obtained after t iterations;
completing the data set of the Bayesian network model according to the current optimal Bayesian network and a preset EM algorithm to obtain a completed data set;
and optimizing the structure and parameters of the Bayesian network model according to the supplemented data set to obtain the structurally optimized Bayesian network model.
Optionally, determining a parameter that maximizes the likelihood function value from the structurally optimized bayesian network model according to a preset maximum likelihood estimation algorithm, includes:
the parameters are determined by the following formula:
L(θ|D,G)＝P(D|θ,G)
θ*＝argmaxL(θ|D,G)＝logP(D|θ,G)
wherein, theta*Representing the parameter; l (theta | D, G) represents a likelihood function corresponding to a node in the Bayesian network model; p (D | θ, G) represents the probability of occurrence of node D under a particular bayesian network structure G and parameter θ.
Optionally, performing conditional probability learning according to the prior probability and the operating data to obtain a posterior probability corresponding to the parameter, including:
the posterior probability is calculated by the following formula:
wherein P (h | D) represents the probability of h occurring if D occurs; p (D) represents the probability of D occurring; p (D | h) represents the probability of D occurring if h occurs; p (h) represents the probability of h occurring; p (theta) represents the probability corresponding to the parameter theta; p (D | θ) represents the probability of D occurring at parameter θ.
Optionally, the method further comprises: and storing the risk assessment in a database, and displaying the risk assessment result in a visual form.
In a second aspect, an embodiment of the present application provides a bayesian network-based risk analysis and assessment apparatus, where the apparatus includes:
the system comprises a modeling unit, a risk evaluation unit and a risk evaluation unit, wherein the modeling unit is used for determining risk factors of a system to be evaluated and constructing a Bayesian network model for risk evaluation according to the risk factors;
the optimization unit is used for acquiring the operating data of the system to be evaluated, and optimizing the Bayesian network model according to the operating data and/or preset expert experience data to obtain an optimized Bayesian network model;
and the evaluation unit is used for acquiring risk factor data of the system to be evaluated and carrying out online risk evaluation on the system to be evaluated according to the risk factor data and the optimized Bayesian network model to obtain a risk evaluation result.
Optionally, the modeling unit is specifically configured to:
determining an incidence relation between any two risk factors, and setting conditional probabilities of any two risk factors;
and constructing the Bayesian network model by taking the risk factors as nodes of the Bayesian network model and the conditional probability as the mutual influence degree of any two risk factors.
Optionally, the optimization unit is specifically configured to:
optimizing the structure and parameters of the Bayesian network model according to a preset iterative algorithm to obtain a structurally optimized Bayesian network model;
determining a parameter which enables a likelihood function value to be maximum from the Bayesian network model after the structure optimization according to a preset maximum likelihood estimation algorithm;
determining prior probability corresponding to the parameters, and performing conditional probability learning according to the prior probability and the operation data to obtain posterior probability corresponding to the parameters;
and adjusting the conditional probability corresponding to each node in the Bayesian network model after the structure optimization according to the posterior probability and/or the preset expert experience data to obtain the optimized Bayesian network model.
Optionally, the optimization unit is specifically configured to:
iteration is carried out from any one Bayesian network in the Bayesian network model, and the current optimal Bayesian network is obtained after t iterations;
completing the data set of the Bayesian network model according to the current optimal Bayesian network and a preset EM algorithm to obtain a completed data set;
and optimizing the structure and parameters of the Bayesian network model according to the supplemented data set to obtain the structurally optimized Bayesian network model.
Optionally, the optimization unit is specifically configured to: the parameters are determined by the following formula:
L(θ|D,G)＝P(D|θ,G)
θ*＝argmaxL(θ|D,G)＝logP(D|θ,G)
wherein, theta*Representing the parameter; l (theta | D, G) represents a likelihood function corresponding to a node in the Bayesian network model; p (D | θ, G) represents the probability of occurrence of node D under a particular bayesian network structure G and parameter θ.
Optionally, the optimizing unit 402 is specifically configured to: the posterior probability is calculated by the following formula:
wherein P (h | D) represents the probability of h occurring if D occurs; p (D) represents the probability of D occurring; p (D | h) represents the probability of D occurring if h occurs; p (h) represents the probability of h occurring; p (theta) represents the probability corresponding to the parameter theta; p (D | θ) represents the probability of D occurring at parameter θ.
Optionally, the evaluation unit is further configured to: and storing the risk assessment in a database, and displaying the risk assessment result in a visual form.
In a third aspect, the present application provides a computer device, comprising:
a memory for storing instructions for execution by at least one processor;
a processor for executing instructions stored in a memory to perform the method of the first aspect.
In a fourth aspect, the present application provides a computer readable storage medium having stored thereon computer instructions which, when run on a computer, cause the computer to perform the method of the first aspect.
Drawings
Fig. 1 is a schematic flowchart of a risk analysis and evaluation method based on a bayesian network according to an embodiment of the present application;
fig. 2 is a schematic structural diagram of a bayesian network model according to an embodiment of the present application;
fig. 3 is a schematic diagram of a bayesian network model design according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of a risk analysis and evaluation device based on a bayesian network according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of a computer device according to an embodiment of the present application.
Detailed Description
In the solutions provided in the embodiments of the present application, the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The bayesian network-based risk analysis and assessment method provided by the embodiment of the present application is further described in detail below with reference to the drawings in the specification, and a specific implementation manner of the method may include the following steps (a method flow is shown in fig. 1):
step 101, determining risk factors of a system to be evaluated, and constructing a Bayesian network model for risk evaluation according to the risk factors.
In the scheme provided by the embodiment of the application, the computer equipment identifies and analyzes the risk factors in the system to be evaluated through methods such as accident trees, main logic analysis and the like. For example, if the system to be evaluated is a beidou satellite system, determining the risk factors influencing the stable operation of the beidou satellite system mainly comprises: the method comprises the steps of controlling risks of technical state while running and joint debugging, reliably running risks of domestic components in an on-orbit long term, abnormal on-orbit risks of satellites, operational service risks of inter-satellite links, on-orbit operational maintenance risks of the third satellite, on-orbit health management risks of multiple satellites and the like.
Further, after determining the risk factors of the system to be evaluated, the computer device constructs a bayesian network model for risk evaluation according to the risk factors. Specifically, there are various ways to construct the bayesian network model, and a preferred way is described as an example below.
In one possible implementation, constructing a bayesian network model for risk assessment from the risk factors comprises: determining an incidence relation between any two risk factors, and setting conditional probabilities of any two risk factors; and constructing the Bayesian network model by taking the risk factors as nodes of the Bayesian network model and the conditional probability as the mutual influence degree of any two risk factors.
Specifically, any risk factor in the system to be evaluated may be present independently, or may be present under specific conditions, for example, a risk factor a and a risk factor B are present in the system to be evaluated, where the risk factor a is generated on the premise that the risk factor B is present, that is, a certain association relationship exists between the risk factor a and the risk factor B. The risk factors are used as nodes of the bayesian network model, any two related risk factors are connected through a connecting line according to the incidence relation between the risk factors, the conditional probabilities of any two risk factors are set, the conditional probabilities are used as the mutual influence degree of the risk factors, all the risk factors are drawn in a directed graph to obtain the bayesian network model, specifically referring to the structure of the bayesian network model shown in fig. 2, the bayesian network model comprises multiple layers, for example, the multiple layers are risk events, risk states and risk factors from top to bottom in sequence.
Further, in the solution provided in the embodiment of the present application, in order to ensure the rationality of the constructed bayesian network model, the computer device needs to evaluate the quality of the bayesian network model in the process of constructing the bayesian network model.
Referring to fig. 3, an exemplary structure of a risk assessment system based on a bayesian network model is provided in the embodiment of the present application. In fig. 3, the risk assessment system includes a view layer, a data layer, and a model layer, where the view layer includes a drawing window, a model management window, a parameter setting window, and a result display window; the data layer comprises XML model data and database data; the model layer comprises a parameter learning module, a structure learning module and an inference module, wherein the parameter learning module comprises complete data learning and incomplete data learning, the complete data is learned through maximum likelihood estimation and Bayesian estimation, the incomplete data is supplemented into complete data through an Expectation optimization (EM) algorithm, and then the incomplete data is learned through maximum likelihood estimation and Bayesian estimation; the Structure learning also includes complete data learning and incomplete data learning, the complete data is learned through a search algorithm and a score function, the incomplete data is firstly supplemented into the complete data through a Structure expectation optimization (SEM) algorithm, and then is learned through the search algorithm and the score function, for example, the search algorithm includes a PC algorithm, an exhaustive search, a hill climbing algorithm, K2 (Cooper-herskovers), an SEM algorithm, and the like, and the score function includes a Bayesian Dirichlet equivalent paradigm (Bdeu), a Bayesian Information Criterion (BIC), and K2.
In the scheme provided by the embodiment of the application, the computer device can judge the quality of the model through the Bdeu, Bayesian information criterion BIC, K2 and other scoring functions, obtain a reasonable Bayesian network model by combining with expert experience, and present the probability value and the inference value of each node in the Bayesian network model by using a visual interface. In the structure shown in fig. 3, the inference module infers through inference algorithms and inference functions, for example, the inference algorithms include Variable inference (VE), Belief Propagation (BP), Maximum multiplicative linear programming (MPLP), and Dynamic Bayesian (DB), and the inference functions include inference and diagnosis.
And 102, acquiring the operation data of the system to be evaluated, and optimizing the parameters of the Bayesian network model according to the operation data and/or preset expert experience data to obtain an optimized Bayesian network model.
In the scheme provided by the embodiment of the application, after the computer device constructs the bayesian network model, the computer device obtains the operation data of the system to be evaluated in a periodic triggering mode, a conditional triggering mode or a real-time triggering mode, and then optimizes the parameters of the bayesian network model according to the operation data and/or preset expert experience data to obtain the optimized bayesian network model. Specifically, there are various ways to optimize parameters of the bayesian network model according to the operating data and/or the preset expert experience data to obtain the optimized bayesian network model, and a preferred way is described below as an example.
In a possible implementation manner, optimizing the bayesian network model according to the operating data and/or preset expert experience data to obtain an optimized bayesian network model, including: optimizing the structure and parameters of the Bayesian network model according to a preset iterative algorithm to obtain a structurally optimized Bayesian network model; determining a parameter which enables a likelihood function value to be maximum from the Bayesian network model after the structure optimization according to a preset maximum likelihood estimation algorithm; determining prior probability corresponding to the parameters, and performing conditional probability learning according to the prior probability and the operation data to obtain posterior probability corresponding to the parameters; and adjusting the conditional probability corresponding to each node in the Bayesian network model after the structure optimization according to the posterior probability and/or the preset expert experience data to obtain the optimized Bayesian network model.
Further, in a possible implementation manner, optimizing the structure and parameters of the bayesian network model according to a preset iterative algorithm to obtain a structurally optimized bayesian network model, the method includes: iteration is carried out from any one Bayesian network in the Bayesian network model, and the current optimal Bayesian network is obtained after t iterations; completing the data set of the Bayesian network model according to the current optimal Bayesian network and a preset EM algorithm to obtain a completed data set; and optimizing the structure and parameters of the Bayesian network model according to the supplemented data set to obtain the structurally optimized Bayesian network model.
In the solution provided in the embodiment of the present application, a bayesian network in the bayesian network model is a network with an incomplete data set, and there are various methods for optimizing and adjusting the structure, parameters, and conditional probabilities corresponding to nodes of the bayesian network model, for example, an SEM algorithm.
In order to facilitate understanding of the process of optimizing the bayesian network model in the solution provided in the embodiments of the present application, the following briefly describes the process of the SEM algorithm. Specifically, the SEM algorithm proceeds as follows:
step 1, iteration is carried out from any Bayesian network B0 (G0, theta 0) in the Bayesian network model, and the current best Bayesian network Bt (Gt, theta t) is obtained after t iterations.
Step 2, carrying out t +1 iteration based on the current best Bayesian network, specifically comprising the following two steps:
and step 21, complementing the data set D of the Bayesian network model by using an EM (effective electromagnetic) algorithm based on the current best Bayesian network Bt (Gt, theta t), so that the data set D is complete and a complete data set Dt is obtained.
And step 22, further optimizing the bayesian network model and the parameters based on the data set Dt to obtain Bt +1 ═ (Gt +1, θ t + 1).
Further, in a possible implementation manner, determining a parameter that maximizes the likelihood function value from the structurally optimized bayesian network model according to a preset maximum likelihood estimation algorithm includes:
the parameters are determined by the following formula:
L(θ|D,G)＝P(D|θ,G)
θ*＝argmaxL(θ|D,G)＝logP(D|θ,G)
wherein, theta*Representing the parameter; l (theta)| D, G) represents a likelihood function corresponding to a node in the Bayesian network model; p (D | θ, G) represents the probability of occurrence of node D under a particular bayesian network structure G and parameter θ.
Further, in a possible implementation manner, performing conditional probability learning according to the prior probability and the operation data to obtain a posterior probability corresponding to the parameter includes:
the posterior probability is calculated by the following formula:
wherein P (h | D) represents the probability of h occurring if D occurs; p (D) represents the probability of D occurring; p (D | h) represents the probability of D occurring if h occurs; p (h) represents the probability of h occurring; p (theta) represents the probability corresponding to the parameter theta; p (D | θ) represents the probability of D occurring at parameter θ.
In order to facilitate understanding of the above-mentioned process of obtaining the posterior probability corresponding to the parameter through the conditional probability learning, the following briefly introduces the process. The method comprises the following specific steps:
1) and selecting the prior probability P (theta) corresponding to the Bayesian network parameter theta.
Specifically, in the process of selecting the prior probability P (θ) corresponding to the parameter θ, if there is no knowledge for determining the prior probability P (θ), a conjugate distribution family may be selected as the prior probability P (θ) of the parameter θ, that is, the prior probability P (θ) and the posterior probability P (θ) of the parameter θ are satisfied and belong to the same class of distribution. Common conjugate distributions include: binomial distribution, multinomial distribution, normal distribution, Dirichlet distribution, with Dirichlet distribution being the most common.
2) And calculating the posterior probability of the parameter theta according to a Bayesian formula, and deducing unknown parameters.
The Bayesian method integrates the prior information and the sample information of the unknown parameter for the estimation of the unknown parameter,
103, acquiring risk factor data of the system to be evaluated, and performing online risk evaluation on the system to be evaluated according to the risk factor data and the optimized Bayesian network model to obtain a risk evaluation result.
Further, in order to facilitate the user to view the risk assessment result, in the solution provided in this embodiment, after step 103, the method further includes:
and storing the risk assessment in a database, and displaying the risk assessment result in a visual form.
In the scheme provided by the embodiment of the application, a risk factor of a system to be evaluated is determined, a Bayesian network model for risk evaluation is constructed according to the risk factor, operation data of the system to be evaluated is obtained, the Bayesian network model is optimized according to the operation data and/or preset expert experience data, an optimized Bayesian network model is obtained, risk factor data of the system to be evaluated is obtained, and online risk evaluation is performed on the system to be evaluated according to the risk factor data and the optimized Bayesian network model, so that a risk evaluation result is obtained. Therefore, in the scheme provided by the embodiment of the application, on one hand, because the Bayesian network model is simple and convenient to model, the risk assessment is carried out through the Bayesian network model, and the complexity of the risk assessment is reduced; on the other hand, the Bayesian network model is optimized through operation data and/or preset expert experience data, and risk assessment is carried out according to the optimized Bayesian network model, namely learning adjustment and parameter learning adjustment can be carried out on the Bayesian network model structure, so that a risk assessment model which is most suitable for practical conditions is obtained, and the applicability of the risk assessment method is further improved.
Based on the same inventive concept as the method shown in fig. 1, an embodiment of the present application provides a risk analysis and evaluation device based on a bayesian network, see fig. 4, the device includes:
the modeling unit 401 is configured to determine risk factors of a system to be evaluated, and construct a bayesian network model for risk evaluation according to the risk factors;
an optimizing unit 402, configured to obtain operation data of the system to be evaluated, and optimize the bayesian network model according to the operation data and/or preset expert experience data to obtain an optimized bayesian network model;
the evaluation unit 403 is configured to obtain risk factor data of the system to be evaluated, and perform online risk evaluation on the system to be evaluated according to the risk factor data and the optimized bayesian network model to obtain a risk evaluation result.
Optionally, the modeling unit 401 is specifically configured to:
determining an incidence relation between any two risk factors, and setting conditional probabilities of any two risk factors;
and constructing the Bayesian network model by taking the risk factors as nodes of the Bayesian network model and the conditional probability as the mutual influence degree of any two risk factors.
Optionally, the optimizing unit 402 is specifically configured to:
optimizing the structure and parameters of the Bayesian network model according to a preset iterative algorithm to obtain a structurally optimized Bayesian network model;
determining a parameter which enables a likelihood function value to be maximum from the Bayesian network model after the structure optimization according to a preset maximum likelihood estimation algorithm;
determining prior probability corresponding to the parameters, and performing conditional probability learning according to the prior probability and the operation data to obtain posterior probability corresponding to the parameters;
and adjusting the conditional probability corresponding to each node in the Bayesian network model after the structure optimization according to the posterior probability and/or the preset expert experience data to obtain the optimized Bayesian network model.
Optionally, the optimizing unit 402 is specifically configured to:
iteration is carried out from any one Bayesian network in the Bayesian network model, and the current optimal Bayesian network is obtained after t iterations;
completing the data set of the Bayesian network model according to the current optimal Bayesian network and a preset EM algorithm to obtain a completed data set;
and optimizing the structure and parameters of the Bayesian network model according to the supplemented data set to obtain the structurally optimized Bayesian network model.
Optionally, the optimizing unit 402 is specifically configured to: the parameters are determined by the following formula:
L(θ|D,G)＝P(D|θ,G)
θ*＝argmaxL(θ|D,G)＝logP(D|θ,G)
wherein, theta*Representing the parameter; l (theta | D, G) represents a likelihood function corresponding to a node in the Bayesian network model; p (D | θ, G) represents the probability of occurrence of node D under a particular bayesian network structure G and parameter θ.
Optionally, the optimizing unit 402 is specifically configured to: the posterior probability is calculated by the following formula:
wherein P (h | D) represents the probability of h occurring if D occurs; p (D) represents the probability of D occurring; p (D | h) represents the probability of D occurring if h occurs; p (h) represents the probability of h occurring; p (theta) represents the probability corresponding to the parameter theta; p (D | θ) represents the probability of D occurring at parameter θ.
Optionally, the evaluation unit 403 is further configured to: and storing the risk assessment in a database, and displaying the risk assessment result in a visual form.
Referring to fig. 5, the present application provides a computer device comprising:
a memory 501 for storing instructions for execution by at least one processor;
a processor 502 for executing instructions stored in memory to perform the method described in fig. 1.
A computer-readable storage medium having stored thereon computer instructions which, when executed on a computer, cause the computer to perform the method of fig. 1.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.
Claims (10)
1. A risk analysis and assessment method based on a Bayesian network is characterized by comprising the following steps:
determining risk factors of a system to be evaluated, and constructing a Bayesian network model for risk evaluation according to the risk factors;
acquiring operation data of the system to be evaluated, and optimizing the Bayesian network model according to the operation data and/or preset expert experience data to obtain an optimized Bayesian network model;
and acquiring risk factor data of the system to be evaluated, and performing online risk evaluation on the system to be evaluated according to the risk factor data and the optimized Bayesian network model to obtain a risk evaluation result.
2. The method of claim 1, wherein constructing a bayesian network model for risk assessment based on the risk factors comprises:
determining an incidence relation between any two risk factors, and setting conditional probabilities of any two risk factors;
and constructing the Bayesian network model by taking the risk factors as nodes of the Bayesian network model and the conditional probability as the mutual influence degree of any two risk factors.
3. The method of claim 1, wherein optimizing the bayesian network model based on the operational data and/or pre-determined expert empirical data to obtain an optimized bayesian network model comprises:
optimizing the structure and parameters of the Bayesian network model according to a preset iterative algorithm to obtain a structurally optimized Bayesian network model;
determining a parameter which enables a likelihood function value to be maximum from the Bayesian network model after the structure optimization according to a preset maximum likelihood estimation algorithm;
determining prior probability corresponding to the parameters, and performing conditional probability learning according to the prior probability and the operation data to obtain posterior probability corresponding to the parameters;
and adjusting the conditional probability corresponding to each node in the Bayesian network model after the structure optimization according to the posterior probability and/or the preset expert experience data to obtain the optimized Bayesian network model.
4. The method of claim 3, wherein optimizing the structure and parameters of the Bayesian network model according to a predetermined iterative algorithm to obtain a structurally optimized Bayesian network model comprises:
iteration is carried out from any one Bayesian network in the Bayesian network model, and the current optimal Bayesian network is obtained after t iterations;
completing the data set of the Bayesian network model according to the current optimal Bayesian network and a preset EM algorithm to obtain a completed data set;
and optimizing the structure and parameters of the Bayesian network model according to the supplemented data set to obtain the structurally optimized Bayesian network model.
5. The method of claim 4, wherein determining the parameter that maximizes the likelihood function value from the structurally optimized Bayesian network model according to a predetermined maximum likelihood estimation algorithm comprises:
the parameters are determined by the following formula:
L(θ|D，G)＝P(D|θ，G)
θ*＝argmaxL(θ|D，G)＝log P(D|θ，G)
wherein, theta*Representing the parameter; l (theta | D, G) represents a likelihood function corresponding to a node in the Bayesian network model; p (D | θ, G) represents the probability of occurrence of node D under a particular bayesian network structure G and parameter θ.
6. The method of claim 3, wherein performing conditional probability learning based on the prior probabilities and the operational data to obtain posterior probabilities corresponding to the parameters comprises:
the posterior probability is calculated by the following formula:
wherein P (h | D) represents the probability of h occurring if D occurs; p (D) represents the probability of D occurring; p (D | h) represents the probability of D occurring if h occurs; p (h) represents the probability of h occurring; p (theta) represents the probability corresponding to the parameter theta; p (D | θ) represents the probability of D occurring at parameter θ.
7. The method of any one of claims 1 to 6, further comprising:
and storing the risk assessment in a database, and displaying the risk assessment result in a visual form.
8. A Bayesian network-based risk analysis and evaluation device, comprising:
the system comprises a modeling unit, a risk evaluation unit and a risk evaluation unit, wherein the modeling unit is used for determining risk factors of a system to be evaluated and constructing a Bayesian network model for risk evaluation according to the risk factors;
the optimization unit is used for acquiring the operating data of the system to be evaluated, and optimizing the Bayesian network model according to the operating data and/or preset expert experience data to obtain an optimized Bayesian network model;
and the evaluation unit is used for acquiring risk factor data of the system to be evaluated and carrying out online risk evaluation on the system to be evaluated according to the risk factor data and the optimized Bayesian network model to obtain a risk evaluation result.
9. The apparatus of claim 8, wherein the modeling unit is specifically configured to:
determining an incidence relation between any two risk factors, and setting conditional probabilities of any two risk factors;
and constructing the Bayesian network model by taking the risk factors as nodes of the Bayesian network model and the conditional probability as the mutual influence degree of any two risk factors.
10. The apparatus according to claim 9, wherein the optimization unit is specifically configured to:
optimizing the structure and parameters of the Bayesian network model according to a preset iterative algorithm to obtain a structurally optimized Bayesian network model;
determining a parameter which enables a likelihood function value to be maximum from the Bayesian network model after the structure optimization according to a preset maximum likelihood estimation algorithm;
determining prior probability corresponding to the parameters, and performing conditional probability learning according to the prior probability and the operation data to obtain posterior probability corresponding to the parameters;
and adjusting the conditional probability corresponding to each node in the Bayesian network model after the structure optimization according to the posterior probability and/or the preset expert experience data to obtain the optimized Bayesian network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010839318.4A CN112069726A (en) | 2020-08-19 | 2020-08-19 | Risk analysis and evaluation method and device based on Bayesian network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010839318.4A CN112069726A (en) | 2020-08-19 | 2020-08-19 | Risk analysis and evaluation method and device based on Bayesian network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112069726A true CN112069726A (en) | 2020-12-11 |
Family
ID=73661363
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010839318.4A Pending CN112069726A (en) | 2020-08-19 | 2020-08-19 | Risk analysis and evaluation method and device based on Bayesian network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112069726A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112686532A (en) * | 2020-12-29 | 2021-04-20 | 中国航天标准化研究所 | Passive operation risk analysis and evaluation method and device based on Bayesian network model |
| CN113159571A (en) * | 2021-04-20 | 2021-07-23 | 中国农业大学 | Cross-border foreign species risk level determination and intelligent identification method and system |
| CN113222358A (en) * | 2021-04-23 | 2021-08-06 | 北京国信云服科技有限公司 | Logistics risk assessment model method, device, equipment and medium based on dynamic Bayesian network |
| CN113537757A (en) * | 2021-07-13 | 2021-10-22 | 北京交通大学 | Method for analyzing uncertain operation risk of rail transit system |
| CN113536678A (en) * | 2021-07-19 | 2021-10-22 | 中国人民解放军国防科技大学 | XSS risk analysis method and device based on Bayesian network and STRIDE model |
| CN113919225A (en) * | 2021-10-12 | 2022-01-11 | 上海海事大学 | Environmental test chamber reliability assessment method and system |
| CN114580874A (en) * | 2022-02-24 | 2022-06-03 | 哈尔滨工业大学 | Multidimensional distributed data analysis system suitable for highway risk assessment |
| CN114792209A (en) * | 2022-05-11 | 2022-07-26 | 保利长大工程有限公司 | Method, equipment and storage medium for engineering construction risk assessment |
| CN115545477A (en) * | 2022-10-08 | 2022-12-30 | 广东电力交易中心有限责任公司 | Power transmission line blocking risk probability assessment method and product based on incremental interpolation |
| CN116822965A (en) * | 2023-08-28 | 2023-09-29 | 中铁七局集团电务工程有限公司武汉分公司 | Subway construction risk early warning method and system |
| CN117670068A (en) * | 2024-02-02 | 2024-03-08 | 青岛哈尔滨工程大学创新发展中心 | AUV real-time risk assessment system and method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107480895A (en) * | 2017-08-19 | 2017-12-15 | 中国标准化研究院 | A kind of reliable consumer goods methods of risk assessment based on Bayes enhancing study |
| CN111178713A (en) * | 2019-12-19 | 2020-05-19 | 中国航天标准化研究所 | Ka phased array antenna on-orbit reliability assessment method based on Bayesian network |
-
2020
- 2020-08-19 CN CN202010839318.4A patent/CN112069726A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107480895A (en) * | 2017-08-19 | 2017-12-15 | 中国标准化研究院 | A kind of reliable consumer goods methods of risk assessment based on Bayes enhancing study |
| CN111178713A (en) * | 2019-12-19 | 2020-05-19 | 中国航天标准化研究所 | Ka phased array antenna on-orbit reliability assessment method based on Bayesian network |
Non-Patent Citations (1)
| Title |
|---|
| 毛子骏 等: "基于贝叶斯网络的智慧城市信息安全风险评估研究", 现代情报, vol. 40, no. 5, pages 19 - 26 * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112686532A (en) * | 2020-12-29 | 2021-04-20 | 中国航天标准化研究所 | Passive operation risk analysis and evaluation method and device based on Bayesian network model |
| CN113159571A (en) * | 2021-04-20 | 2021-07-23 | 中国农业大学 | Cross-border foreign species risk level determination and intelligent identification method and system |
| CN113222358A (en) * | 2021-04-23 | 2021-08-06 | 北京国信云服科技有限公司 | Logistics risk assessment model method, device, equipment and medium based on dynamic Bayesian network |
| CN113537757A (en) * | 2021-07-13 | 2021-10-22 | 北京交通大学 | Method for analyzing uncertain operation risk of rail transit system |
| CN113537757B (en) * | 2021-07-13 | 2024-02-09 | 北京交通大学 | Analysis method for uncertain risk of rail transit system operation |
| CN113536678A (en) * | 2021-07-19 | 2021-10-22 | 中国人民解放军国防科技大学 | XSS risk analysis method and device based on Bayesian network and STRIDE model |
| CN113536678B (en) * | 2021-07-19 | 2022-04-19 | 中国人民解放军国防科技大学 | XSS risk analysis method and device based on Bayesian network and STRIDE model |
| CN113919225B (en) * | 2021-10-12 | 2024-02-02 | 上海海事大学 | Environmental test box reliability assessment method and system |
| CN113919225A (en) * | 2021-10-12 | 2022-01-11 | 上海海事大学 | Environmental test chamber reliability assessment method and system |
| CN114580874A (en) * | 2022-02-24 | 2022-06-03 | 哈尔滨工业大学 | Multidimensional distributed data analysis system suitable for highway risk assessment |
| CN114792209A (en) * | 2022-05-11 | 2022-07-26 | 保利长大工程有限公司 | Method, equipment and storage medium for engineering construction risk assessment |
| CN114792209B (en) * | 2022-05-11 | 2022-11-22 | 保利长大工程有限公司 | Method, equipment and storage medium for engineering construction risk assessment |
| CN115545477B (en) * | 2022-10-08 | 2023-07-14 | 广东电力交易中心有限责任公司 | Power transmission line blocking risk probability assessment method and product based on incremental interpolation |
| CN115545477A (en) * | 2022-10-08 | 2022-12-30 | 广东电力交易中心有限责任公司 | Power transmission line blocking risk probability assessment method and product based on incremental interpolation |
| CN116822965A (en) * | 2023-08-28 | 2023-09-29 | 中铁七局集团电务工程有限公司武汉分公司 | Subway construction risk early warning method and system |
| CN116822965B (en) * | 2023-08-28 | 2023-11-21 | 中铁七局集团电务工程有限公司武汉分公司 | Subway construction risk early warning method and system |
| CN117670068A (en) * | 2024-02-02 | 2024-03-08 | 青岛哈尔滨工程大学创新发展中心 | AUV real-time risk assessment system and method |
| CN117670068B (en) * | 2024-02-02 | 2024-04-12 | 青岛哈尔滨工程大学创新发展中心 | AUV real-time risk assessment system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112069726A (en) | Risk analysis and evaluation method and device based on Bayesian network | |
| US10877863B2 (en) | Automatic prediction system for server failure and method of automatically predicting server failure | |
| Pearl | On evidential reasoning in a hierarchy of hypotheses | |
| US20220405641A1 (en) | Method for recommending information, recommendation server, and storage medium | |
| JP6176979B2 (en) | Project management support system | |
| CN113031983B (en) | Intelligent software upgrading method and device based on deep reinforcement learning | |
| CN109389518A (en) | Association analysis method and device | |
| CN115705322A (en) | Database management system, data processing method and equipment | |
| KR101741248B1 (en) | Method and apparatus for estimating causality among variables | |
| CN115373888A (en) | Fault positioning method and device, electronic equipment and storage medium | |
| Kirchhof et al. | Root cause analysis in lithium-ion battery production with fmea-based large-scale bayesian network | |
| Sandhu et al. | A comparative analysis of conjugate gradient algorithms & PSO based neural network approaches for reusability evaluation of procedure based software systems | |
| Zhang et al. | A novel QoS prediction approach for cloud service based on bayesian networks model | |
| US20140052431A1 (en) | Supporting proactive decision-making in event-driven applications | |
| Bayarri et al. | A Bayesian sequential look at u-control charts | |
| CN117666546A (en) | Distributed control system fault diagnosis method and device | |
| Smirnov | Intelligent decision support system for the control of complex technical systems | |
| Saeed et al. | Non-functional requirements trade-off in self-adaptive systems | |
| CN115712874A (en) | Thermal energy power system fault diagnosis method and device based on time series characteristics | |
| US20160110641A1 (en) | Determining a level of risk for making a change using a neuro fuzzy expert system | |
| Mohammadi et al. | Machine learning assisted stochastic unit commitment: A feasibility study | |
| KR20080087571A (en) | Context prediction system and method thereof | |
| CN113064812A (en) | Project development process quality defect prediction method, device and medium | |
| Rawat et al. | A review on software reliability: metrics, models and tools. | |
| KR100818493B1 (en) | Method for analyzing and predicting error generated in broadcasting system and apparatus therefor |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |