A trust access control method based on user behavior profile

Abstract

The present invention belongs to the technical field of network security, and specifically relates to a trust access control method based on user behavior archives, comprising: constructing an electric power communication network system and initializing the system; the user sends an authentication request to the electric power communication system, and the system verifies the identity information of the user after receiving the request. If the verification fails, the system fails to authorize the user; if the verification succeeds, the system uses an initial role and permission allocation algorithm to allocate an initial role and initial permission to the user and sets a target permission value; the environment in which the user device is currently located is detected to obtain an environment value; the initial permission is updated according to the current environment value, the initial role and the target permission value to obtain the current authorization level of the user; and access control is performed on the system according to the current authorization level of the user; the present invention realizes fine-grained dynamic permission allocation through a TBAC model.

Description

Technical Field

[0001] The present invention belongs to the technical field of network security, and in particular relates to a trust access control method based on user behavior profiles. Background Art

[0002] With the continuous development of the power grid system, the scale of the power grid has become larger and more complex. The number of users in the power grid is also increasing, and the management of the permissions of each user is becoming more and more complicated. Traditional access control models often have difficulty adapting to the access requirements of different users for resources in a multi-task environment. After the role-based access control (RBAC) model is applied to the power grid system, users entering the power grid system will be assigned specific roles, so that users have certain permissions to manage and allocate resources in the power grid. However, no security-related attribute information is considered. For example, once an attacker steals the account and password of a power grid user, he can easily obtain access rights to the system and data and steal data resources. This security risk has a particularly large impact on accounts with high access rights. However, the RBAC model is static and coarse-grained, and it is impossible to achieve dynamic and fine-grained management of user permissions. Once the RBAC model completes the allocation of permissions and roles, the user's permissions will not change until the administrator reassigns them. The reason why the RBAC model cannot dynamically allocate user permissions. When the scale of the system changes or the business logic changes, it is also necessary to maintain the allocation relationship of a large number of users and roles, which is also prone to explosion problems. Summary of the invention

[0003] In order to solve the problems existing in the above-mentioned prior art, the present invention proposes a trust access control method based on user behavior profiles, which includes: building an electric power communication network system and initializing the system; the user sends an authentication request to the electric power communication system, and the system verifies the identity information of the user after receiving the request. If the verification fails, the system fails to authorize the user; if the verification succeeds, the system uses an initial role and permission allocation algorithm to allocate an initial role and initial permission to the user and sets a target permission value; the current environment of the user's device is detected to obtain an environment value; the initial permission is updated according to the current environment value, the initial role and the target permission value to obtain the user's current authorization level; and access control is performed on the system according to the user's current authorization level.

[0004] Beneficial effects of the present invention:

[0005] The present invention proposes a Trust-Based Access Control (TBAC) algorithm based on user behavior profiles in smart grids, which dynamically adjusts the authorization status by evaluating the user behavior value and the environmental value of the permission in the smart grid. The user's behavior value is obtained by predicting the user's possible behavior based on the user's historical data and the deviation between the current behavior. The environmental value is set according to the current environment of the resource. Through TBAC, fine-grained dynamic permission allocation can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Figure 1 It is the overall framework structure diagram of the TBAC model of the present invention;

[0007] Figure 2 It is a user access control authorization flow chart of the present invention;

[0008] Figure 3 An evaluation diagram after the rights of the present invention are allocated;

[0009] Figure 4 This is a flowchart of the initial allocation of role permissions of the present invention. DETAILED DESCRIPTION

[0010] 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.

[0011] A trust access control method based on user behavior profiles, the method comprising: constructing an electric power communication network system and initializing the system; the user sends an authentication request to the electric power communication system, the system verifies the user's identity information after receiving the request, if the verification fails, the system fails to authorize the user, if the verification succeeds, the system uses a role authority initial allocation algorithm to allocate an initial role and initial authority to the user and sets a target authority value; the environment in which the user's device is currently located is detected to obtain an environment value; the initial authority is updated according to the current environment value, the initial role and the target authority value to obtain the user's current authorization level; and access control is performed on the system according to the user's current authorization level.

[0012] System initialization includes initializing users, roles, and permissions respectively; initializing roles includes: analyzing user roles, which include basic roles, intermediate roles, advanced roles, and management roles; analyzing user historical behaviors, including user historical behavior data, device usage information, and login status.

[0013] The system verifies the user's identity including: the user sends an authentication request to the system, where the authentication request includes the user name, user password, fingerprint information and facial information; the system compares the identity information sent by the user with the identity information in the system database. If the comparison is successful, the user's identity authentication is successful, otherwise the user's identity authentication fails.

[0014] In this embodiment, in order to realize dynamic access control of users in the power grid, a trust access control (TBAC) based on user behavior profile is proposed. The overall architecture of the model is as follows: Figure 1 Compared with the RBAC model, this model adds trust and authorization adjustment nodes when allocating roles and permissions. This model dynamically allocates corresponding permissions to users based on user behavior trust and environmental values. User behavior trust is obtained by predicting the deviation between possible behaviors and current behaviors based on user historical data. Environmental values ​​are set according to the current environment.

[0015] The elements and allocation relationships of the TBAC model are:

[0016] User: A collection of all users. A user is the subject of accessing resources.

[0017] Role: A collection of all roles. Role is a bridge between users and permissions.

[0018] Object: A collection of resources. Resources are objects that are accessed by users.

[0019] Operation collection: An operation is an action that uses a resource.

[0020] Permission collection: A permission is a binary tuple consisting of an operation and an object, representing the TBAC model authorization flow qualification for applying the operation to a resource.

[0021] User behavior value: The system predicts the user's possible behavior based on historical behavior and the comprehensive trust evaluation of the current behavior. It is represented by UV.

[0022] Environmental value: A comprehensive consideration of the current environment of the resource. It is the minimum limit for users to operate on the object. The environmental value changes according to the changes in the environment. The environment here includes the current power load, power resource cost, power resource sensitivity level, and time changes. Changes in the environment will cause changes in the importance and security level of the resource. The environmental value is expressed as EV.

[0023] Authorization adjustment: In order to achieve dynamic access control and authorization, when UV≥EV, the system can adjust the user to the appropriate permissions.

[0024] In this embodiment, the access control model of TABC evaluates the user's behavior value based on the user's credit profile. The system can dynamically adjust the authorization relationship based on the user's behavior value and environmental value. The user's behavior value is evaluated based on the deviation between the current behavior of the power grid user and his historical behavior. The environmental value will be determined based on the power load of the corresponding resource in the power grid, the power resource cost, the sensitivity of the power resource, and the change over time. The specific authorization process is as follows: Figure 2 As shown; the specific process includes:

[0025] Step 1: Initialization: The power communication network system will initialize the role allocation, corresponding authority allocation, user behavior value and environment value for the user.

[0026] Step 2: Authentication: The system will authenticate the user and detect the user's device environment.

[0027] Step 3: Initial role and authority allocation: When the user passes the identity authentication of the power communication network system, the initial role and authority will be allocated according to the initial role and authority allocation algorithm, and the environment value EV of the current authorized authority will be obtained. cur And the target authority EV tar .

[0028] Step 4: Role behavior value evaluation: After the user obtains the role and corresponding permissions assigned by the power grid system, he / she performs a series of operations on the resources in the power grid. In this process, the power grid system calculates the degree of deviation between the user's current behavior and its predicted behavior. This paper uses this deviation value (user behavior value) to measure the degree of abnormality of user behavior.

[0029] The authorization algorithm is as follows: after a period of time after the user obtains the permission, the user's behavior value is evaluated; if UV k ≥EV cur , then proceed to the next step, otherwise directly deny authorization; if UV k <EV tar , the permissions of the user are adjusted, and the user behavior value is continued to be evaluated after a period of time. Otherwise, the target permissions of the user are adjusted and the authorization is terminated.

[0030] In order to dynamically adjust user permissions, it is necessary to periodically evaluate user behavior values ​​and then adjust user permissions. Figure 3 shown.

[0031] After the user obtains the target permission, in order to understand the abnormal situation of user permission in real time, it is still necessary to periodically evaluate the user's behavior value. k <EV tar, indicating that the current user has an abnormal situation, the authorization of the user's rights is cancelled; otherwise, the original rights are kept unchanged, and the user behavior value is evaluated again after a time period T.

[0032] In this embodiment, the initial allocation of role permissions includes: different types of roles in the power grid have different degrees of influence on the normal operation of the power grid. Before the power grid initializes the allocation of roles to users, it is necessary to classify the roles according to the degree of influence on the power grid, and then further allocate permissions according to the roles of the users. If all roles in the power grid are initially allocated equally, not only is the allocation form single and cannot meet the differentiated management of role permissions, but it may also cause improper data access, operational errors, improper resource access, increased security risks and other adverse effects on the power grid. In order to avoid the above problems, reasonable initial permission allocation is crucial to the security and reliability of the power grid system. By grading according to the role type and the degree of influence on the operation of the power grid, and allocating permissions according to actual business needs and security considerations, it can ensure that each role has appropriate permissions to perform its specific duties and tasks. This helps maintain the security, availability and compliance of the system.

[0033] After the user passes the system's identity authentication, the role permissions will be initially allocated. Since the user will operate the resources in the power grid, he will apply for the corresponding operation permissions. However, the system will make considerations based on the permissions applied by the user, so as to achieve the purpose of reasonably allocating user permissions. The specific operation process is as follows Figure 4 shown.

[0034] The system performs initial role assignment. First, it analyzes the user's role, which includes the following two steps:

[0035] (1) Level Analysis Different roles in the power grid have different impacts on the power grid, which requires the administrators in the power grid to classify the roles in the power grid according to the actual situation of the power grid. According to the survey, the role levels in the power grid can be divided into: basic roles, such as electricity consumers; intermediate roles, such as power equipment engineers, power market participants; advanced roles, such as system administrators, power operation and maintenance experts, etc.; management roles, such as power network administrators, etc.

[0036] (2) The user historical behavior analysis system needs to analyze the user's historical behavior, including user historical behavior data, device usage information, login status, etc., and prepare for the subsequent permission initialization allocation.

[0037] Based on the role behavior analysis, develop a role permission matrix that lists the operations that each role should be authorized to perform. This can include different levels of operations such as viewing, downloading, adding, deleting, and modifying.

[0038] Before authorization, the user will apply for the target operation permission. The system will evaluate the user's operation permission and then reasonably allocate the initial permission. The specific details are as follows: First, based on the previous system analysis of the user's historical behavior, the corresponding permission matrix of the user role is judged. Within a certain time range (the time range can be set by the power network administrator), in the data of the user's historical behavior analysis: if the user has used permissions outside the non-role permission matrix, regardless of the role level, the authorization will be directly revoked; if the user's permission has a record of unauthorized jump, regardless of the role level, the initial permission will be allocated; if the user has an abnormal operation behavior record within the permission, if the role is a basic role, the next level of permission will be allocated to the user; if the role level is above the basic role, the next two levels of permission will be allocated to the user; if the user's historical behavior is normal after analysis, the user will be allocated the target operation permission; the initial allocation of user permissions is completed.

[0039] In this embodiment, the access control and authorization system in the power grid will make a preliminary assessment based on the type and attributes of the user, analyze the security level of the user's behavior from multiple dimensions, and regularly calculate the user's behavior value. Based on the user's behavior value and the current resource environment value, the authorization relationship between the user and the authority is dynamically changed.

[0040] There are many types of roles in the power grid, and different types of roles have different permissions and types of resources that can be operated in the power grid. According to the survey, the roles in the power grid include: generator operator, transmission system operator (TSO), distribution system operator (DSO), power market participants (Market Participants), end users (End Users, etc. The roles, permissions and operational resources in these power grids are described in the following Table 1:

[0041] Table 1 Power grid role table

[0042]

[0043]

[0044] The access control and authorization system in the power grid has established a user behavior profile by accumulating and analyzing a large amount of user behavior and attribute data. The user behavior profile covers multiple characteristics of the user's historical behavior. By studying the degree of deviation between the user's current behavior and its predicted behavior, the system can determine whether the current behavior is abnormal.

[0045] The user behavior profile contains historical behavior data, device information, power usage patterns, security information and other information. Each category of the user behavior profile describes the user's behavior characteristics. By comparing the user behavior profile with the current behavior, the degree of deviation between the user's current behavior and the predicted behavior can be found, so as to dynamically adjust the trust level. The details are shown in Table 2 below:

[0046] Table 2 User behavior profile

[0047]

[0048]

[0049] According to the changes in the security requirements of the power grid system, the behavior categories contained in the user behavior profile can be expanded or deleted. At the same time, different types of data and data accuracy can also be used.

[0050] In this embodiment, each category in the user behavior file corresponds to a trust evaluation factor and weight of the user behavior value. The trust evaluation factor has three evaluation values: trust (corresponding to value 1), partial trust (corresponding to value 0.5) and distrust (corresponding to value 0). The trust evaluation factor of each category of the user behavior file is determined based on the degree of deviation between the current behavior data and the predicted behavior.

[0051] Each category in the user behavior profile corresponds to a trust evaluation factor and weight of the user behavior value. The trust evaluation factor has three evaluation values: trust (corresponding to value 1), partial trust (corresponding to value 0.5) and distrust (corresponding to value 0). The trust evaluation factor of each category in the user behavior profile is determined based on the degree of deviation between the current behavior data and the predicted data of the category by the ARIMA (AutoregressiveIntegrated Moving Average) model.

[0052] The RIMA model can be used to predict data for a period of time in the future. The ARIMA model (autoregressive integrated moving average model) is a time series forecasting model that predicts future data based on past data. Specifically, the ARIMA model uses the autocorrelation and partial autocorrelation properties in time series data to establish a mathematical model that describes the changes in data over time, and then uses the model to predict future data.

[0053] Specifically, calculating the user's behavior value includes: selecting a certain category of user historical data in the user behavior file, using the ARIMA model to predict the obtained category user data, and obtaining the prediction result; calculating the data prediction deviation x between the predicted data and the actual data b , set the maximum upper limit of the deviation degree Δx t2 and a general upper limit Δxt1 , if the current data prediction deviation x b ≤Δx t1 , then the deviation is small, and the corresponding trust evaluation factor is 1; if the current data prediction deviation is Δx t1 <x b ≤Δx t2 , then the current user behavior deviates from the predicted behavior, and the corresponding trust evaluation factor is 0.5; if the current data prediction deviation x b >Δx t2 , the current user behavior deviates too much from the predicted behavior, and the trust evaluation factor is 0; the user's behavior value is calculated based on the trust evaluation factor.

[0054] The ARIMA model is a classic time series analysis method, which consists of three parts: autoregression (AR), difference (I) and moving average (MA). The mathematical expression of the ARIMA model is as follows:

[0055] Given a time series {x t}, where t = 1, 2, 3, ..., T, the ARIMA (p, d, q) model is expressed as: (1-φ 1 B-φ 2 B 2 -…-φ p B p )(1-B) d X t =(1+θ 1 B+θ 2 B 2 +…+θ q B q )ε t

[0056] Among them, B is the backward movement operator (lag operator); φ 1 ,φ 2 ,…,φ p is the autoregressive coefficient, which indicates the linear relationship between the current value and the previous p values; θ 1 ,θ 2 ,…,θ q is the moving average coefficient, which indicates the linear relationship between the current value and the previous q residuals; d is the order of time series difference, which is used to transform the non-stationary time series into a stationary series; ε t is a white noise series, representing random fluctuations in the time series that are not explained by the model.

[0057] In this embodiment, using the ARIMA model to predict user behavior data includes: preprocessing the user behavior data and dividing the data into a training set and a prediction set; using the finite difference method to perform smooth processing on the preprocessed data; using the autocorrelation function and the partial autocorrelation function to determine the order of the ARIMA model; according to the order of the ARIMA model, the parameters of the model are determined, and the smoothed training data is input into the ARIMA model for training, so that the historical data fits the parameters of the ARIMA model, and the test data set is used for testing, and the root mean square error and the mean absolute error are used to evaluate the model; the data of the current user is input into the ARIMA model for prediction to obtain the final prediction result.

[0058] Specifically include:

[0059] 1) First, you need to clean the data, including processing missing values ​​and outliers, and perform feature selection, delete irrelevant features, and perform corresponding feature conversion. Finally, you need to divide the data set into a training set and a test set. After processing the data, if you want to use the ARIMA model to predict the data, you need to ensure the stability of the data set. If it is unstable, perform data differential processing to keep the data stable.

[0060] There are many ways to make data stationary. One common method is to use the difference method, which is to perform finite differences on the data points. t Indicates the static trend by Z t Therefore, Z t Equal to Y t Differences between consecutive values:

[0061] Z t =Y t -Y t-1

[0062] If the non-stationary data set also exhibits seasonal characteristics, seasonal differencing is performed on the data set.

[0063] Z t =Y t -Y t-m

[0064] Differencing is very effective and therefore recommended because in most cases, non-stationary data can be easily transformed into stationary data after the first differencing, so no further transformation is needed. Other transformation methods can achieve stationarity after multiple transformations on the same data point. A single root test (ADF test) is usually used to determine the order of differencing required, that is, the value of d.

[0065] 2) Determination of model order.

[0066] The autocorrelation function (ACF) and partial autocorrelation function (PACF) are used to determine the order of the ARIMA model, namely AR(p) and MA(q).

[0067] The autocorrelation function (ACF) describes the linear correlation between the observed values ​​of the time series and its past observations. The calculation formula is as follows:

[0068]

[0069] Where k represents the lag period, and its value range is [-1,1].

[0070] For a stationary AR(p) model, we can find the autocorrelation coefficient ρ at lag k. k After that, what is actually obtained is not a simple correlation between x(t) and x(tk). Because x(t) is also affected by the k-1 random variables in the middle, and these k-1 random variables are all correlated with x(tk), so the autocorrelation coefficient is actually mixed with the influence of other variables on x(t) and x(tk). In order to simply measure the influence of x(tk) on x(t), the concept of partial autocorrelation coefficient (PACF) is introduced. The specific expression is as follows:

[0071]

[0072] The specific ARIMA model is determined by ACF and PACF. The determined model is shown in Table 3:

[0073] Table 3 Determine the specific ARIMA model table

[0074] ACF tailing ACF truncation PACF tailing ARMA(p,q) MA(q) PACF censoring AR(p) Model not applicable

[0075] After determining the order of the ARIMA model through the above steps, according to the mathematical expression of the ARIMA model, the determined order is used to fit the parameters of the ARIMA model, namely φ and θ, using historical data. The parameters are estimated using the maximum likelihood estimation method, which estimates the parameters of the model by maximizing the log-likelihood function.

[0076] 3) After determining the parameters of the model, use the trained ARIMA model to predict the value at a certain point in time (t+h) in the future. The expression of the predicted value is:

[0077]

[0078] in: Represents the predicted value of the time series XX at time t+h; c is a constant term, representing the intercept of the model; φ 1 ,φ 2 ,…,φ pis the autoregressive coefficient, which indicates the linear relationship between the current value and the previous p values; X t+h-1 ,X t+h-2 ,.....,X t+h-q is the lag term of the time series at the prediction time t+h; θ 1 ,θ 2 ,…,θ q is the moving average coefficient, which indicates the linear relationship between the current value and the previous q residuals; θ 1 e t+h-1 ,θ 2 e t+h-2 ,...,θ q e t+h-q It is the residual term of the time series at the prediction time t+h.

[0079] 4) After obtaining the predicted value, it is necessary to evaluate the accuracy of the predicted value, using the root mean square error (RMSE) and mean absolute error (MAE) as evaluation indicators. By calculating the prediction results of the test set, the root mean square error (RMSE) and mean absolute error (MAE) are expressed as follows:

[0080]

[0081] Among them, n is the number of test samples, is the real observation of the test samples, is the prediction at the corresponding time point. RMSE is the standard deviation of the difference between the predicted value and the true value. The smaller the RMSE, the smaller the difference between the predicted value and the true value of the model, and the better the prediction effect. MAE is the average of the absolute errors between the predicted value and the true value. Similar to RMSE, the smaller the MAE, the better the prediction effect. Generally speaking, the two indicators of RMSE and MAE are considered comprehensively to comprehensively evaluate the prediction effect of the model. If the RMSE and MAE of the model are both small, it means that the model has good prediction ability. Otherwise, it is necessary to further improve the model or find other methods to improve the prediction accuracy.

[0082] 5) Predict the behavior data of each category of the current user. After obtaining the predicted data, normalize the deviation between the predicted result and the actual result. Define the data prediction deviation as:

[0083]

[0084] Set the maximum upper limit of the prediction deviation Δx t2 and a general upper limit Δx t1 Δx t1 It is the general upper limit of the normal forecast data deviation. Δx t2 is the maximum upper limit of the normal forecast data deviation. If the current data forecast deviation x b ≤Δxt1 , then the recognition bias is small, and the corresponding trust evaluation factor is 1. If the current data prediction deviation Δx t1 <x b ≤Δx t2 , indicating that the current user behavior deviates from the predicted behavior within a certain range, and the corresponding trust evaluation factor is 0.5. b >Δx t2 It means that the current user behavior deviates too much from the predicted behavior, and the trust evaluation factor is 0.

[0085] The system will evaluate the user's trust based on the degree of deviation between the user's current behavior and the predicted behavior. If the deviation is small, the trust is high (factor is 1), the trust is moderate within a certain range (factor is 0.5), and the trust is low if the deviation is too large (factor is 0).

[0086] The user behavior value is generated by adding the trust evaluation factors of each category of the user behavior profile according to the weight ratio. Its expression is:

[0087]

[0088] Among them, variable n is the category of behavioral characteristics. Variable s i is the trust evaluation factor of each category of user behavior profile. i is the weight of each category of user behavior profile. The weight is dynamically adjusted according to the changes in the system security environment. Since the user behavior value is generated regularly, the user's historical behavior value will have a certain impact on the current access behavior. Therefore, the time decay factor γ is introduced. The user's final behavior value UV in the current cycle k Based on the behavior value UV of the previous cycle k-1 and the currently calculated behavior value UV cur Calculated. Final behavior value UV k The expression is:

[0089] UV k =γUV k-1 +(1-γ)UV cur

[0090] In this embodiment, different resources of the system have different operation permissions, and different operation permissions have different environmental values. When the user's behavior value is greater than a certain environmental value, the system will assign and adjust certain permissions to the user. For example, a resource has a variety of different operations, and the corresponding environmental values ​​are different. This article takes the power equipment of the power generator as an example to set the operation permissions of the power equipment under different environmental values, as shown in Table 4 below.

[0091] Table 4 Power equipment environment value settings and operation permissions

[0092] Permission Level Environmental Values Operation Permissions Level 0 0.0-0.2 No permission, no operation can be performed Level 1 0.2-0.4 Can only view device status and obtain basic information Level 2 0.4-0.6 Ability to start / stop the device and adjust basic settings Level 3 0.6-0.8 Advanced adjustment of device parameters and performance optimization can be performed Level 4 0.8-0.9 Allows for equipment maintenance and diagnostics Level 5 0.9-1.0 Full control of the device, firmware / software updates

[0093] If the current user's behavior value is 0.1, the user has no operation authority over the power equipment. If the current user's behavior value is 0.5, the user can start or stop the equipment and make basic settings for the equipment. When the user's behavior value is 0.95, the user can fully control the equipment.

[0094] Different power resources are located in different geographical environments, and the factors used to evaluate environmental values ​​also vary from environment to environment. Therefore, administrators are required to set the operating permissions of power resources based on the specific environment in which the power resources are located.

[0095] The access control and authorization system in the power grid regularly calculates the behavior value of the access user. At the same time, the environmental value of each permission is dynamically adjusted according to the changes in the environment to achieve fine-grained dynamic access control.

[0096] The above embodiments further illustrate the purpose, technical solutions and advantages of the present invention in detail. It should be understood that the above embodiments are only preferred implementation modes of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made to the present invention within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A trust access control method based on user behavior profile, characterized in that: include: Build the power communication network system and initialize the system; The user sends an authentication request to the power communication system. After receiving the request, the system verifies the user's identity information. If the verification fails, the system fails to authorize the user. If the verification succeeds, the system uses the role permission initial allocation algorithm to assign the user an initial role and initial permission and set the target permission value; detect the current environment of the user's device to obtain the environment value; update the initial permission according to the current environment value, initial role and target permission value to obtain the user's current authorization level; Control access to the system based on the user's current authorization level; Obtaining the final behavior value of the user includes: selecting a certain category of user historical data in the user behavior file, using the ARIMA model to predict the obtained category user data, and obtaining the prediction result; calculating the data prediction deviation between the predicted data and the actual data x b , obtain the trust evaluation factor according to the data prediction deviation; calculate the user's behavior value according to the trust evaluation factor; the user's final behavior value is: , ;in, γ is the time decay factor, n is the category of the behavior feature, s i is the trust evaluation factor for each category of user behavior profile, w i is the weight of each category of user behavior profile; The initial role and permissions are assigned to users using the role and permissions initial allocation algorithm, including: Step 1: Get the user's final behavior value UV k , set the target permission environment value EV tar ; Step 2: Compare the user's final behavior value with the current authorization permission environment value EV cur For comparison, if UV k ≥ EV cur , then execute step 3, otherwise user authorization fails; Step 3: Compare the user's final behavior value with the target permission environment value EV tar For comparison, if UV k < EV tar , the permissions of the user are adjusted, and the final behavior value of the user is evaluated after a period of time. Otherwise, the target permissions of the user are adjusted and the authorization is terminated.

2. A trust access control method based on user behavior profile according to claim 1, characterized in that: Initializing the system includes initializing users, roles, and permissions respectively; The initial analysis of roles includes: analyzing the user's roles, which include basic roles, intermediate roles, advanced roles, and management roles; Analyze the user's historical behavior, including user historical behavior data, device usage information, and login status.

3. A trust access control method based on user behavior profile according to claim 1, characterized in that: The system verifies the user's identity including: the user sends an authentication request to the system, where the authentication request includes the user name, user password, fingerprint information and facial information; the system compares the identity information sent by the user with the identity information in the system database. If the comparison is successful, the user's identity authentication is successful, otherwise the user's identity authentication fails.

4. A trust access control method based on user behavior profile according to claim 1, characterized in that: The use of the ARIMA model to predict user behavior data includes: preprocessing the user behavior data and dividing the data into a training set and a prediction set; using the finite difference method to perform smooth processing on the preprocessed data; using the autocorrelation function and the partial autocorrelation function to determine the order of the ARIMA model; determining the parameters of the model according to the order of the ARIMA model, and inputting the smoothed training data into the ARIMA model for training, so that the historical data fits the parameters of the ARIMA model, using the test data set for testing, and using the root mean square error and mean absolute error to evaluate the model; inputting the current user's data into the ARIMA model for prediction to obtain the final prediction result.

5. A trust access control method based on user behavior profile according to claim 4, characterized in that: The autocorrelation function is: ; where k represents the lag period, ρ k is the autocorrelation function, Cov is the covariance, x t is the data at the current time t, x t-k The data for time tk, Var is the variance.

6. A trust access control method based on user behavior profile according to claim 4, characterized in that: The partial autocorrelation function is: ;in, is the mean of the data before time t, is the mean of the data before time tz.

7. A trust access control method based on user behavior profile according to claim 4, characterized in that: The ARIMA model predicts user behavior data including: ; where c is a constant term, φ p is the autoregression coefficient, X t+h-p is the lag term of the time series at the prediction time t+h, θ q is the moving average coefficient, e t+h-p It is the residual term of the time series at the prediction time t+h.

Images