A multi-level adaptive identity verification method and system based on dynamic risk scoring

By adjusting identity authentication strategies through dynamic risk scoring, the problem of fixed strategies in traditional authentication systems is solved. This enables multi-level user adaptive verification and continuous monitoring, improving security and user experience. It is suitable for scenarios such as internet government affairs, financial services, and enterprise office work.

CN123122967APending Publication Date: 2026-11-17TRAFFIC MANAGEMENT RES INST OF THE MIN OF PUBLIC SECURITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611441179.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-16
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing identity authentication systems have fixed authentication policies that cannot adapt to multi-level user permission systems and real-time risk changes. They also lack continuous monitoring and dynamic response mechanisms, resulting in security vulnerabilities and poor user experience.

Method used

By adopting a multi-level adaptive identity verification method based on dynamic risk scoring, multi-dimensional identity verification information is collected, authentication strategies are dynamically adjusted, user behavior and environment are continuously monitored, and supplementary authentication is triggered according to the risk score range, so as to achieve the matching of authentication strength and business risk.

Benefits of technology

It achieves a match between authentication strength and risk, supports multi-level user adaptive verification, continuously defends against persistent attacks, reduces management costs, meets legal compliance requirements, and is suitable for various application scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122967A_ABST
    Figure CN123122967A_ABST
Patent Text Reader

Abstract

This invention relates to a multi-level adaptive identity verification method and system based on dynamic risk scoring. The invention includes: responding to a user's business operation request by collecting multi-dimensional identity verification information; calculating an initial risk score; constructing an authentication strategy mapping relationship; determining the corresponding verification confidence score; weighted fusion of the corresponding verification confidence scores based on the weight coefficients of each authentication factor to obtain a comprehensive authentication score; determining the initial identity authentication result; determining the behavioral anomaly score, environmental change risk increment, and session duration risk increment, and updating the initial risk score accordingly to obtain the current risk score; triggering supplementary authentication or reducing the number of authentication factors required for subsequent business operations and / or reducing the trigger frequency of supplementary authentication. This invention can improve the security and efficiency of identity verification in continuous session scenarios.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of information security and identity authentication technology, and in particular to a multi-level adaptive identity verification method and system based on dynamic risk scoring. Background Technology

[0002] With the widespread adoption of internet-based government services and online business processing, identity verification, as the first line of defense for ensuring business security, is becoming increasingly important. This is especially true in traffic management scenarios for organizational users, where there are multiple user levels (three levels of access: legal representative, administrator, and ordinary user), complex permissions, and diverse business types. Traditional unified identity authentication methods face the following prominent problems: First, static authentication methods cannot adapt to changes in risk. Existing identity authentication systems typically employ fixed authentication strategies, using the same combination of authentication factors for all users and all business scenarios. This makes it impossible to dynamically adjust authentication strength based on factors such as business risk level, user behavior characteristics, and operating environment. This "one-size-fits-all" approach either leads to over-authentication for low-risk businesses and a poor user experience, or under-authentication for high-risk businesses and obvious security vulnerabilities. Second, there is a lack of adaptive support for multi-level user permission systems. Under the three-level permission system of legal persons, administrators, and ordinary users, the operational permissions and risk exposure of users at different levels differ significantly, but existing systems struggle to dynamically configure differentiated verification strategies based on user roles. Third, there is a lack of continuous monitoring and dynamic response mechanisms after authentication. Traditional authentication methods only perform identity verification once upon login or business initiation, granting full operational permissions after successful authentication. This fails to continuously assess risk changes during user operations and lacks effective defense against persistent attacks such as session hijacking and account impersonation. Fourth, the combination of authentication factors is rigid. Although existing multi-factor authentication systems support multiple authentication factors, the factor combination methods are pre-fixed, making it impossible to dynamically adjust the type, number, and weight of authentication factors based on real-time risk scores.

[0003] In recent years, relevant national departments have continuously promoted the construction of a trusted identity authentication system for "Internet and government services," explicitly requiring the establishment of a hierarchical, dynamically adaptive, and fully traceable identity authentication mechanism. However, existing technical solutions have not yet formed a complete adaptive identity verification system capable of dynamically adjusting authentication strategies based on multi-dimensional real-time risk scoring and supporting multi-level user permission systems. This system fails to address the aforementioned industry pain points across the entire chain, including risk assessment, strategy matching, continuous monitoring, and dynamic response. Therefore, researching an adaptive identity verification method and system that can dynamically match authentication strength with business risks, support differentiated verification for multi-level users, and possess continuous monitoring and dynamic response capabilities has become an urgent need for the construction of current Internet business security authentication systems and the security protection of actual business operations. Summary of the Invention

[0004] Therefore, the technical problem to be solved by the present invention is to overcome the problems in the existing identity authentication methods, such as fixed authentication strategies, difficulty in adaptively configuring authentication strategies for multi-level users, lack of continuous risk monitoring after authentication, and inability to dynamically select authentication factors for specific risk sources.

[0005] To address the aforementioned technical problems, this invention provides a multi-level adaptive identity verification method based on dynamic risk scoring, comprising: In response to a user's business operation request, multi-dimensional identity verification information is collected, including user static attribute information, real-time environmental context information, business scenario information, and real-time behavioral feature information. User role risk score and historical behavior risk score are determined based on the user static attribute information, environmental risk score is determined based on the real-time environmental context information, business risk score is determined based on the business scenario information, and an initial risk score is calculated based on the user role risk score, the environmental risk score, the business risk score, and the historical behavior risk score. A certification strategy mapping relationship is constructed, which represents the mapping relationship between different risk score intervals and certification strategies. Each certification strategy includes a combination of certification factors and a certification pass threshold. The corresponding risk score interval is determined according to the initial risk score, and the certification strategy is determined according to the risk score interval. The verification of each certification factor is performed according to the combination of certification factors corresponding to the certification strategy, and the corresponding verification confidence score is determined according to the certification type and verification result of each certification factor. The corresponding verification confidence scores are weighted and fused based on the weight coefficients of each authentication factor to obtain a comprehensive authentication score; the initial identity authentication result is determined by comparing the comprehensive authentication score with the authentication pass threshold corresponding to the initial authentication strategy. During the user session after initial identity authentication, the user's operation behavior, operation environment and session duration are continuously monitored based on the real-time behavioral feature information. The abnormal behavior score, environmental change risk increment and session duration risk increment are determined according to the monitoring results, and the initial risk score is updated accordingly to obtain the current risk score. The current risk score is compared with the risk score range in the authentication policy mapping table. When the current risk score falls into a higher risk score range, supplementary authentication is triggered. When the current risk score falls back to a lower risk score range, the number of authentication factors required for subsequent business operations is reduced and / or the frequency of supplementary authentication is reduced.

[0006] In one embodiment of the present invention, a user role risk score and a historical behavior risk score are determined based on the user's static attribute information, an environmental risk score is determined based on the real-time environmental context information, a business risk score is determined based on the business scenario information, and an initial risk score is calculated based on the user role risk score, the environmental risk score, the business risk score, and the historical behavior risk score, including: User role risk score is determined based on the role hierarchy in the user's static attribute information. The role hierarchy includes legal entities, administrators, and ordinary users, and historical behavior risk scores are determined based on historical authentication records in the user's static attribute information. ; Environmental risk is determined based on at least one of the following: geographical location, IP address, network environment, operation time, and device fingerprint, from the real-time environmental context information. ; Business risk is determined based on at least one of the following: business type, business sensitivity level, and operational impact scope from the business scenario information: ; The initial risk score is calculated using a weighted summation model, and the calculation formula is as follows: ; in, For the initial risk score, , , , These are the weighting coefficients corresponding to the user role risk score, the environment risk score, the business risk score, and the historical behavior risk score, respectively.

[0007] In one embodiment of the present invention, determining a corresponding risk score range based on the initial risk score, and determining an authentication strategy based on the risk score range, includes: The risk scoring interval includes a first risk scoring interval, a second risk scoring interval, and a third risk scoring interval; When the initial risk score is lower than the first risk threshold but greater than 0, it is determined that the initial risk score is within the first risk score range, and a low-risk authentication strategy using single-factor authentication is matched. When the initial risk score is greater than or equal to the first risk threshold and lower than the second risk threshold, it is determined that the initial risk score is in the second risk score range, and a medium-risk authentication strategy using two-factor authentication is matched. When the initial risk score is greater than or equal to the second risk threshold, the initial risk score is determined to be within the third risk score range, and a high-risk authentication strategy using three or more authentication factors is matched.

[0008] In one embodiment of the present invention, determining the initial identity authentication result based on a comparison between the comprehensive authentication score and the authentication pass threshold corresponding to the initial authentication strategy includes: Certification factors are divided into binary judgment type certification factors and continuous scoring type certification factors; For the binary decision-type authentication factor, the verification result is determined based on the matching result between the authentication information and the preset verification conditions. When the verification passes, the corresponding verification confidence score is assigned. The score is set at 100. If the verification fails, the corresponding verification confidence score will be adjusted. The score is set to 0. The binary judgment authentication factor includes at least one of password verification, SMS verification code verification, digital certificate verification, and dynamic password verification. For the continuously rated authentication factor, the matching score output by the verification algorithm of the continuously rated authentication factor is determined as the corresponding verification confidence score. The matching score ranges from 0 to 100, and the continuous scoring authentication factor includes face recognition. The verification confidence scores of each authentication factor involved in the current initial identity authentication are weighted and fused according to the following formula to obtain the comprehensive authentication score. : ; in, To achieve a comprehensive certification score, Let i be the weight of the i-th authentication factor. Let be the verification confidence score of the i-th authentication factor, and n be the number of authentication factors participating in the current authentication strategy; When a pre-set authentication factor in the authentication strategy is not involved in the current initial identity authentication, the weight coefficient corresponding to that authentication factor is proportionally allocated to other authentication factors involved in the current initial identity authentication. When the comprehensive authentication score C0 is greater than or equal to the authentication pass threshold corresponding to the authentication strategy, the initial identity authentication is determined to be successful. When the overall authentication score C0 is less than the authentication pass threshold, it is determined that the initial identity authentication has failed.

[0009] In one embodiment of the present invention, the abnormal behavior score, environmental change risk increment, and session duration risk increment are determined based on monitoring results, and the initial risk score is updated accordingly to obtain the current risk score, including: calculating the current risk score according to the following formula: ; in, Assess the current risk level. Scoring for behavioral abnormalities For the initial risk score, For the incremental risk of environmental change, This represents an increase in the risk associated with session duration.

[0010] In one embodiment of the present invention, behavioral anomaly scoring Calculation methods include: Within a unit time window, outlier operation frequency is calculated based on the user's historical average number of operations, the standard deviation of historical operation counts, and the current actual number of operations. The calculation formula is as follows: ; in, This represents the actual number of operations performed within the current time window. This represents the user's historical average number of operations within a given time window. The standard deviation of the historical number of operations; When a user skips the normal steps and directly accesses sensitive functions, the operation path will be an anomaly. Set to 30, otherwise set to 0; When the current operation time falls within the preset non-working hours or statutory holidays, the operation time will be considered an abnormal value. Set to 20, otherwise set to 0; The formula for calculating the behavioral abnormality score is as follows: ; in, A score is given for abnormal behavior, with a maximum score of 40.

[0011] In one embodiment of the present invention, the environmental change risk increment The calculation methods include: Construct an environmental feature vector consisting of IP address range, geolocation code, device fingerprint, and network type, and compare the current session's environmental feature vector with the user's historical trusted environmental feature vector. When the city corresponding to an IP address range changes, the incremental risk of IP address change is determined. When a device fingerprint changes, the incremental risk of device change is determined; When a sudden change in geographical location occurs, the incremental risk of geographical location change should be determined. Determine the risk increment of network change when the network type changes; The formula for calculating the incremental environmental change risk is as follows: ; in, For the incremental risk of environmental change, This is to mitigate the increased risk associated with IP address changes. This is to account for the increased risk associated with equipment changes. The risk of increased risk due to changes in geographical location The risk increment is determined by network changes, with an upper limit of 50 points for environmental change risks.

[0012] In one embodiment of the present invention, the session duration risk increment The calculation methods include: Set session security duration threshold Maximum session duration threshold Based on the duration of the current session Determine the incremental risk of session duration ;in: hour, ; hour, ; hour, .

[0013] In one embodiment of the present invention, when the current risk score enters a higher risk score range, supplementary authentication is triggered, including: The types of certification factors to be added are determined based on the types of risk sources that cause the current risk score to rise, including: Based on at least one of the sources of risk from environmental change risk increment, the sources of risk from abnormal behavior scoring, and the business risk score corresponding to the current business scenario, candidate authentication factors corresponding to the risk source type are determined from the authentication factor library; Exclude authentication factors that have already been verified in the current session, and determine at least one supplementary authentication factor from the remaining candidate authentication factors based on the number of authentication factors required for the current risk level. If the current risk score, after supplementary authentication, is still higher than the maximum risk threshold, the current session will be terminated and the user account will be locked. Wherein, when the incremental risk of environmental change reaches a preset environmental anomaly threshold, and the incremental risk of environmental change is generated by changes in device fingerprint, at least one of digital certificate authentication and facial recognition authentication is determined as a candidate authentication factor. When the incremental risk of environmental change is caused by changes in IP address or geographic location, at least one of face recognition and liveness detection will be identified as a candidate authentication factor. When the behavior anomaly score reaches a preset behavior anomaly threshold, candidate authentication factors are determined based on the behavior anomaly type that generated the behavior anomaly score. Specifically, when the behavior anomaly type is an operation frequency anomaly, dynamic password authentication is determined as a candidate authentication factor, and when the behavior anomaly type is an operation path anomaly, facial recognition authentication is determined as a candidate authentication factor. When the business risk score corresponding to the current business scenario reaches the preset business risk threshold, and no abnormal environmental changes or behavioral abnormalities are detected, digital certificate authentication will be determined as a candidate authentication factor.

[0014] This invention also provides a multi-level adaptive identity verification system based on dynamic risk scoring, comprising: The multi-dimensional information collection module is used to collect multi-dimensional identity verification information in response to user business operation requests. The multi-dimensional identity verification information includes user static attribute information, real-time environmental context information, business scenario information, and real-time behavioral feature information. The risk scoring calculation module is used to determine the user role risk score and historical behavior risk score based on the user's static attribute information, determine the environment risk score based on the real-time environment context information, determine the business risk score based on the business scenario information, and calculate the initial risk score based on the user role risk score, the environment risk score, the business risk score, and the historical behavior risk score. The authentication strategy mapping and decision-making module is used to construct authentication strategy mapping relationships, which represent the mapping relationship between different risk score intervals and authentication strategies. Each authentication strategy includes a combination of authentication factors and an authentication pass threshold. The module determines the corresponding risk score interval based on the initial risk score and determines the authentication strategy based on the risk score interval. It then performs the verification of each authentication factor according to the authentication factor combination corresponding to the authentication strategy and determines the corresponding verification confidence score based on the authentication type and verification result of each authentication factor. The initial identity authentication result acquisition module is used to perform weighted fusion of the corresponding verification confidence scores based on the weight coefficients of each authentication factor to obtain a comprehensive authentication score; and to determine the initial identity authentication result based on the comparison between the comprehensive authentication score and the authentication pass threshold corresponding to the initial authentication strategy. The continuous behavior monitoring module is used to continuously monitor user operation behavior, operation environment and session duration based on the real-time behavior feature information during the user session after the initial identity authentication is passed. Based on the monitoring results, it determines the behavior abnormality score, the risk increment of environmental change and the risk increment of session duration, and updates the initial risk score accordingly to obtain the current risk score. The adjustment module is used to compare the current risk score with the risk score range in the authentication policy mapping table. When the current risk score falls into a higher risk score range, supplementary authentication is triggered. When the current risk score falls back to a lower risk score range, the number of authentication factors required for subsequent business operations is reduced and / or the trigger frequency of supplementary authentication is reduced.

[0015] The technical solution of the present invention has the following advantages compared with the prior art: This invention discloses a multi-level adaptive identity verification method and system based on dynamic risk scoring. Firstly, it matches authentication strength with risk, balancing security and efficiency. The invention dynamically adjusts the authentication strategy based on multi-dimensional real-time risk scoring. Low-risk services use lightweight authentication to improve user experience, while high-risk services automatically upgrade authentication strength to ensure security, solving the problem of balancing security and efficiency inherent in traditional fixed authentication strategies. Secondly, it supports adaptive verification for multi-level users. By constructing a three-level role model (legal entity, administrator, and ordinary user), with different roles corresponding to different basic risk values ​​and permission boundaries, the invention achieves adaptive identity verification support for a multi-level user system. Thirdly, it provides continuous monitoring and dynamic response to defend against persistent attacks. After initial authentication, the invention continuously monitors user behavior and environmental changes. Upon detecting anomalies, it immediately triggers risk reassessment and authentication upgrades, effectively defending against persistent security threats such as session hijacking and account impersonation. Fourthly, it automatically adapts authentication strategies, reducing management costs. Through an authentication strategy mapping table and automated decision-making mechanism, the invention achieves dynamic matching and automatic upgrading and downgrading of authentication strategies without manual intervention, significantly reducing system management costs. Fifth, the entire process is auditable and traceable. This invention writes the entire identity verification process to a blockchain or tamper-proof audit log, supporting post-event security audits and compliance checks, meeting the requirements of laws and regulations such as the Cybersecurity Law and the Personal Information Protection Law for identity authentication and operational auditing. Sixth, it is highly scalable and applicable to a wide range of scenarios. This invention is not only applicable to traffic management business processing scenarios for organizational users, but can also be extended to various application scenarios requiring hierarchical identity authentication, such as e-government, financial services, enterprise office, and telemedicine. Attached Figure Description

[0016] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0017] Figure 1 This is an overall flowchart of the multi-level adaptive identity verification method based on dynamic risk scoring of the present invention.

[0018] Figure 2 This is a schematic diagram illustrating the structure of the multi-dimensional information collection method of the present invention.

[0019] Figure 3 This is a module architecture diagram of the multi-level adaptive identity verification system based on dynamic risk scoring according to the present invention.

[0020] Figure 4 This is a flowchart illustrating the continuous monitoring and dynamic upgrade authentication strategy of this invention. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0022] In this invention, "several" means one or more, "multiple" means two or more, "greater than," "less than," "exceeding," etc., are understood to exclude the stated number; "above," "below," "within," etc., are understood to include the stated number. In the description of this invention, the terms "first" and "second" are used only to distinguish technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0023] Example 1 Reference Figures 1 to 4 As shown, this embodiment provides a multi-level adaptive identity verification method based on dynamic risk scoring. This method can be applied to scenarios where unit users handle Internet traffic management business, as well as other online business systems with hierarchical permissions, differences in business sensitivity levels, and continuous session risk control requirements.

[0024] Before performing identity verification, a multi-level user role model and permission mapping relationship can be pre-built. The role hierarchy for organizational users includes legal representatives, administrators, and ordinary users. Different role levels correspond to different sets of basic permissions, permission boundaries, and risk exposures. Legal representatives can perform operations such as organizational account management, business approval, and authorization; administrators can perform operations such as daily business processing and personnel management; and ordinary users can perform business operations corresponding to their personal permissions. The system can also pre-establish the mapping relationship between role levels and business operation types, and configure corresponding business sensitivity levels and basic risk parameters for different business types. The aforementioned role model, risk thresholds, weight coefficients, and authentication strategy parameters can all be configured and optimized by the system management module.

[0025] The steps are as follows: S1. In response to the user's business operation request, collect multi-dimensional identity verification information, including user static attribute information, real-time environmental context information, business scenario information and real-time behavioral feature information.

[0026] It should be noted that user static attribute information may include user ID, role level, historical authentication records, bound device information, associated mobile phone number and certificate information; real-time environmental context information may include geographical location, IP address, network environment, operation time and device fingerprint, the device fingerprint may be generated by device model, operating system version, screen resolution and other information; business scenario information may include the business type of the current request, business sensitivity level and operation impact range; real-time behavioral characteristic information may include the operation frequency of the current session, mouse or touch trajectory, input rhythm and page dwell time.

[0027] S2. Determine the user role risk score and historical behavior risk score based on the user's static attribute information, determine the environment risk score based on the real-time environment context information, determine the business risk score based on the business scenario information, and calculate an initial risk score based on the user role risk score, the environment risk score, the business risk score, and the historical behavior risk score. Specifically, this includes: User role risk score is determined based on the role hierarchy in the user's static attribute information. The role hierarchy includes legal entities, administrators, and ordinary users, and historical behavior risk scores are determined based on historical authentication records in the user's static attribute information. ; Environmental risk is determined based on at least one of the following: geographical location, IP address, network environment, operation time, and device fingerprint, from the real-time environmental context information. ; Business risk is determined based on at least one of the following: business type, business sensitivity level, and operational impact scope from the business scenario information: ; The initial risk score is calculated using a weighted summation model, and the calculation formula is as follows: ; in, For the initial risk score, , , , These are the weighting coefficients corresponding to the user role risk score, the environment risk score, the business risk score, and the historical behavior risk score, respectively. Each weighting coefficient can be adjusted based on business security requirements, historical statistical results, or policy configuration. In addition to the weighted summation model, if a risk score usable for authentication policy mapping can be output, a trained machine learning model can also be used to score the multi-dimensional identity verification information.

[0028] S3. Construct an authentication strategy mapping relationship, which represents the mapping relationship between different risk score intervals and authentication strategies. Each authentication strategy includes a combination of authentication factors and an authentication pass threshold. Determine the corresponding risk score interval based on the initial risk score, and determine the authentication strategy based on the risk score interval. Perform the verification of each authentication factor according to the authentication factor combination corresponding to the authentication strategy, and determine the corresponding verification confidence score based on the authentication type and verification result of each authentication factor.

[0029] The process of determining a corresponding risk score range based on the initial risk score, and determining an authentication strategy based on the risk score range, includes: The risk scoring intervals include a first risk scoring interval, a second risk scoring interval, and a third risk scoring interval, which correspond to low-risk authentication strategies, medium-risk authentication strategies, and high-risk authentication strategies, respectively.

[0030] When the initial risk score is lower than the first risk threshold but greater than 0, it is determined that the initial risk score is within the first risk score range, and a low-risk authentication strategy using single-factor authentication is matched. When the initial risk score is greater than or equal to the first risk threshold and lower than the second risk threshold, it is determined that the initial risk score is in the second risk score range, and a medium-risk authentication strategy using two-factor authentication is matched. When the initial risk score is greater than or equal to the second risk threshold, the initial risk score is determined to be within the third risk score range, and a high-risk authentication strategy using three or more authentication factors is matched.

[0031] As an example parameter configuration, when 0 ≤ R0 < 40, the initial risk score is assigned to the first risk score interval, using single-factor authentication. The authentication factor can be a password or SMS verification code, and the authentication pass threshold can be set to 60 points. When 40 ≤ R0 < 75, the initial risk score is assigned to the second risk score interval, using two-factor authentication. The authentication factors can be a password or facial recognition, and the authentication pass threshold can be set to 80 points. When 75 ≤ R0 ≤ 100, the initial risk score is assigned to the third risk score interval, employing a high-risk authentication strategy with three or more authentication factors. The authentication factors can be a password, facial recognition, or digital certificate, and the authentication pass threshold can be set to 90 points. The values ​​of 40, 75, and the authentication pass thresholds mentioned above are example values ​​and can be configured according to the business security level in actual deployment.

[0032] S4. The corresponding verification confidence scores are weighted and fused based on the weight coefficients of each authentication factor to obtain a comprehensive authentication score; the initial identity authentication result is determined based on the comparison between the comprehensive authentication score and the authentication pass threshold corresponding to the initial authentication strategy.

[0033] Specifically, it includes: Certification factors are divided into binary judgment type certification factors and continuous scoring type certification factors; For the binary decision-type authentication factor, the verification result is determined based on the matching result between the authentication information and the preset verification conditions. When the verification passes, the corresponding verification confidence score is assigned. The score is set at 100. If the verification fails, the corresponding verification confidence score will be adjusted. The score is set to 0. The binary judgment authentication factor includes at least one of password verification, SMS verification code verification, digital certificate verification, and dynamic password verification. For the continuously rated authentication factor, the matching score output by the verification algorithm of the continuously rated authentication factor is determined as the corresponding verification confidence score. The matching score ranges from 0 to 100, and the continuous scoring authentication factor includes face recognition. The verification confidence scores of each authentication factor involved in the current initial identity authentication are weighted and fused according to the following formula to obtain the comprehensive authentication score. : ; in, To achieve a comprehensive certification score, Let i be the weight of the i-th authentication factor. Let be the verification confidence score of the i-th authentication factor, and n be the number of authentication factors participating in the current authentication strategy. When a certain authentication factor preset in the authentication strategy does not participate in the current initial identity authentication due to strategy configuration, terminal capability or other reasons, the authentication factor will not participate in the comprehensive authentication score calculation, and its corresponding weight coefficient will be proportionally allocated to other authentication factors that actually participate in the authentication, so as to keep the sum of the weights of the authentication factors that actually participate in the authentication consistent.

[0034] When the comprehensive authentication score C0 is greater than or equal to the authentication pass threshold corresponding to the authentication strategy, the initial identity authentication is determined to be successful, and the user is allowed to proceed with subsequent business operations. When the comprehensive authentication score C0 is less than the authentication pass threshold, it is determined that the initial identity authentication has failed. The current business operation request can be rejected, or the process can be transferred to manual review according to the business management rules.

[0035] S5. During the user session after the initial identity authentication is passed, the user's operation behavior, operation environment and session duration are continuously monitored based on the real-time behavioral feature information. The abnormal behavior score, environmental change risk increment and session duration risk increment are determined according to the monitoring results, and the initial risk score is updated accordingly to obtain the current risk score.

[0036] Specifically, it includes: The current risk score is calculated using the following formula: ; in, Assess the current risk level. Scoring for behavioral abnormalities For the initial risk score, For the incremental risk of environmental change, This represents an increase in the risk associated with session duration.

[0037] By limiting the upper limit of each risk increment, we can avoid the risk score from being abnormally amplified by extreme values ​​in a single dimension, while enabling the current risk score to comprehensively represent the overall risk level of the current session.

[0038] Specifically, using a unit time window T window For example, T can be... window Set to 60 seconds to count the average number of historical operations performed by the user within this time window. Standard deviation of historical operation count and the current number of actual operations And calculate the outlier of the operation frequency according to the following formula. : ; in, This represents the actual number of operations performed within the current time window. This represents the user's historical average number of operations within a given time window. The standard deviation of the historical number of operations; The formula for calculating the behavioral abnormality score is as follows: ; in, A score is given for abnormal behavior, with a maximum score of 40.

[0039] It should be noted that when the current actual number of operations does not exceed the sum of the historical average number of operations and twice the standard deviation, the operation frequency anomaly value is 0; when it exceeds the above baseline, the greater the excess, the higher the operation frequency anomaly value. If a user skips the normal operation steps and directly accesses sensitive functions, such as calling the sensitive business submission interface without entering the business application page, then the operation path anomaly value will be... Set to 30 otherwise, set to 0. If the current operation time falls within preset non-working hours, such as 22:00 to 06:00 the next day, or on a statutory holiday, then the operation time will be considered an abnormal value. Set to 20, otherwise set to 0.

[0040] When a user skips the normal steps and directly accesses sensitive functions, the operation path will be an anomaly. Set to 30, otherwise set to 0; When the current operation time falls within the preset non-working hours or statutory holidays, the operation time will be considered an abnormal value. Set to 20, otherwise set to 0.

[0041] Specifically, the increase in environmental change risk The calculation methods include: Construct an environmental feature vector consisting of IP address range, geolocation code, device fingerprint, and network type, and compare the current session's environmental feature vector with the user's historical trusted environmental feature vector. When the city corresponding to an IP address range changes, the incremental risk of IP address change is determined. When a device fingerprint changes, the incremental risk of device change is determined; When a sudden change in geographical location occurs, the incremental risk of geographical location change should be determined. Determine the risk increment of network change when the network type changes; The formula for calculating the incremental environmental change risk is as follows: ; in, For the incremental risk of environmental change, This is to mitigate the increased risk associated with IP address changes. This is to account for the increased risk associated with equipment changes. The risk of increased risk due to changes in geographical location The risk increment is determined by network changes, with an upper limit of 50 points for environmental change risks.

[0042] It should be noted that the risk of environmental change is increasing. This is used to reflect the degree of change in the current operating environment relative to the user's historical trusted environment. An environment feature vector V can be constructed as (IP address range, geolocation code, device fingerprint, network type), where the environment feature vector for the current session is... The user's historical trusted environment feature vector is .when and When there is a perfect match, ΔE can be determined to be 0; when the city corresponding to the IP address range changes, the IP address change risk increment can be calculated. Set to 25; when the device fingerprint changes, i.e., when a new device is used for the first time, the device change risk increment can be increased. Set to 25; when there are sudden changes in geographical location, such as switching from one city to another in a short period of time, the risk increment of geographical location change can be increased. Set to 30; this increases the network change risk increment when the network type changes, such as switching from workplace Wi-Fi to mobile data. Set to 10. The environmental change risk increment ΔE can be calculated using the following formula: Specifically, the risk increment of session duration The calculation methods include: Set session security duration threshold Maximum session duration threshold Based on the duration of the current session Determine the incremental risk of session duration ;in: hour, ; hour, ; hour, .

[0043] S6. Compare the current risk score with the risk score range in the authentication strategy mapping table. When the current risk score falls into a higher risk score range, trigger supplementary authentication; when the current risk score falls back to a lower risk score range, reduce the number of authentication factors required for subsequent business operations and / or reduce the frequency of supplementary authentication triggering. For example, if the current authentication requirement falls from three high-risk authentication factors to the medium-risk range, subsequent business operations can execute two authentication factors according to the medium-risk strategy; if it falls from the medium-risk range to the low-risk range, subsequent business operations can execute a single authentication factor according to the low-risk strategy. The downgrade only applies to the authentication requirements of subsequent business operations and does not change the existing audit records.

[0044] Specifically, when the current risk score enters a higher risk score range, supplementary authentication is triggered, including: The types of certification factors to be added are determined based on the types of risk sources that cause the current risk score to rise, including: Based on at least one of the sources of risk from environmental change risk increment, the sources of risk from abnormal behavior scoring, and the business risk score corresponding to the current business scenario, candidate authentication factors corresponding to the risk source type are determined from the authentication factor library; Exclude authentication factors that have already been verified in the current session, and determine at least one supplementary authentication factor from the remaining candidate authentication factors based on the number of authentication factors required for the current risk level. If the current risk score, after supplementary authentication, is still higher than the maximum risk threshold, the current session will be terminated and the user account will be locked. Wherein, when the incremental risk of environmental change reaches a preset environmental anomaly threshold, and the incremental risk of environmental change is generated by changes in device fingerprint, at least one of digital certificate authentication and facial recognition authentication is determined as a candidate authentication factor. When the incremental risk of environmental change is caused by changes in IP address or geographic location, at least one of face recognition and liveness detection will be identified as a candidate authentication factor. When the behavior anomaly score reaches a preset behavior anomaly threshold, candidate authentication factors are determined based on the behavior anomaly type that generated the behavior anomaly score. Specifically, when the behavior anomaly type is an operation frequency anomaly, dynamic password authentication is determined as a candidate authentication factor, and when the behavior anomaly type is an operation path anomaly, facial recognition authentication is determined as a candidate authentication factor. When the business risk score corresponding to the current business scenario reaches the preset business risk threshold, and no abnormal environmental changes or behavioral abnormalities are detected, digital certificate authentication will be determined as a candidate authentication factor.

[0045] It should be noted that supplementary authentication determines the number of authentication factors to be added not only based on the risk score but also based on the type of risk source that caused the current risk score to rise. The system can select candidate authentication factors from the authentication factor library based on at least one of the following: the source of the increased risk from environmental changes, the source of the abnormal behavior score, and the business risk score corresponding to the current business scenario. It excludes authentication factors that have already been verified in the current session and then determines at least one supplementary authentication factor based on the number of authentication factors required for the current risk level, thereby avoiding the repeated execution of authentications with the same verification attributes.

[0046] When the incremental risk of environmental change reaches a preset environmental anomaly threshold, and the risk source is a change in device fingerprint, at least one of digital certificate authentication and facial recognition authentication, which have a weak binding relationship with the current device, can be identified as a candidate authentication factor to avoid continuing to rely on the changed device fingerprint for identity verification. When the incremental risk of environmental change is caused by changes in IP address or geographical location, at least one of facial recognition and liveness detection can be identified as a candidate authentication factor to confirm the consistency between the current operator and the account holder. When the behavior anomaly score reaches a preset behavior anomaly threshold, candidate authentication factors are determined according to the type of behavior anomaly. Specifically, dynamic password authentication can be given priority when the operation frequency is abnormal, and facial recognition authentication can be given priority when the operation path is abnormal. When the business risk score corresponding to the current business scenario reaches a preset business risk threshold, and no environmental change anomalies or behavior anomalies are detected, digital certificate authentication can be identified as a candidate authentication factor.

[0047] When at least two risk sources exist simultaneously, the candidate authentication factor sets corresponding to each risk source can be merged and deduplicated, with priority given to retaining authentication factors with complementary verification attributes. For example, when both device fingerprint changes and high-risk business scenarios correspond to digital certificate authentication, the digital certificate is executed only once; if there is also a change in geographical location, facial recognition or liveness detection can be added to form a supplementary authentication strategy that can verify different risk sources simultaneously.

[0048] Supplementary authentication can be performed as follows: First, analyze the sources of risk that lead to an increased risk score. Based on the current risk score's risk level range, select candidate authentication factors from the authentication factor library. After excluding authentication factors already verified in the current session, a supplementary authentication strategy is formed, and a supplementary authentication challenge is generated and pushed to the user's terminal. A response time limit of, for example, 180 seconds can be set. Second, execute the verification process for each supplementary authentication factor sequentially on the user's terminal. If a certain authentication factor fails verification, retry is allowed, but the number of retries cannot exceed a preset limit, such as 3 times. After all supplementary authentication factors are completed, the verification confidence scores of each authentication factor are weighted and fused to obtain a comprehensive supplementary authentication score. This comprehensive score is then compared with a preset supplementary authentication pass threshold.

[0049] When supplementary authentication passes, the high-confidence verification result of the supplementary authentication factor can be used as a risk reduction factor to reduce the current risk score and re-determine the risk level. If the recalculated current risk score falls below the security threshold, the current session is maintained and the user is allowed to continue performing business operations. If the current risk score is still higher than the security threshold, the next round of supplementary authentication can be determined based on the new risk source and the current risk level. When supplementary authentication fails, including when the user fails to complete authentication within the response time limit, fails after reaching the preset retry limit, or the supplementary authentication comprehensive score is lower than the preset threshold, or when the recalculated current risk score after supplementary authentication still continues to be higher than the highest risk threshold, the current session is terminated and the user account is locked, and a security alert is pushed to the system administrator.

[0050] S7. Furthermore, generate tamper-proof audit logs throughout the entire identity verification process. Audit content may include anonymized user identity identifiers, calculated values ​​and basis for each risk score, authentication policy matching and switching records, verification results and confidence levels of each authentication factor, triggering reasons and times for policy upgrades or downgrades, and the final authentication result.

[0051] Audit logs can employ a chained hash storage structure. Each log entry includes at least a unique log identifier, session identifier, operation type, current risk score, authentication policy identifier, verification results for each authentication factor, the hash value `prev_hash` of the previous record, and the hash value `hash` of the current record. The hash of the current record is obtained by concatenating the business fields of the current record with a string and performing a SHA-256 hash operation on `prev_hash`, where the `prev_hash` of the first log entry uses a preset initial value. When writing a log, the business fields of the current record are first generated, the hash of the previous log entry is read from storage as `prev_hash`, the current hash is calculated, and then written to the complete record. The current hash is also cached for use by the next record. During post-audit, the hashes of each record are recalculated sequentially, and the `prev_hash` of each record is checked to ensure consistency with the hash of the previous record, thereby determining the integrity of the log chain. If any historical record is modified, subsequent hash chain verifications will show inconsistencies.

[0052] The following describes the operation of the above method using a specific business scenario. A unit administrator initiates a traffic violation transfer in the 12123 unit version of the traffic management application. The system collects information that the user's role is administrator, historical authentication records are normal, the operating environment is a historically trusted environment, the business type is a high-risk business, and the current operation is normal. The risk score calculation module obtains an initial risk score of 72 points based on the R0 calculation formula. Taking a first risk threshold of 40 and a second risk threshold of 75 as an example, 72 points falls within the second risk score range. Therefore, a two-factor authentication strategy is matched, executing password verification and facial recognition. The authentication pass threshold can be set to 80 points.

[0053] After initial authentication, the continuous behavior monitoring module continues to monitor user actions, operating environment, and session duration. If a subsequent change in the device fingerprint is detected and other increments are 0, the parameters described in the example above will be applied. =25, the current risk score can be updated to 97, thus entering the third risk score range and triggering supplementary authentication. Since password verification and facial recognition have already been completed in the current session, the system can select digital certificate authentication as a supplementary authentication factor from candidate authentication factors based on the risk source of device fingerprint change. After the digital certificate verification is successful, the system reassesses the risk based on the supplementary authentication result; if the risk score falls below the security threshold, the session continues; if the supplementary authentication fails or the risk remains above the highest risk threshold, the session is terminated and the account is locked. The audit log storage module synchronously records the initial score, authentication policy, environmental changes, supplementary authentication process, and final handling result.

[0054] Example 2 This embodiment uses the example of a unit administrator handling illegal transfer business on the "Traffic Management 12123" platform to explain in detail the implementation process of the present invention.

[0055] Step 1: As Figure 1 As shown, the process begins with the user initiating a business operation request. The unit administrator logs into the unit version of the APP, initiates an illegal transfer business operation, and the system starts the identity verification process.

[0056] Step 2: As Figure 2 As shown, the multi-dimensional information collection module collects four types of information: User static attributes: Role is "Administrator", historical authentication records are normal; Environment context: The operating IP is located in Wuxi, Jiangsu Province; the device fingerprint matches the historically bound device; the operation time is 10:30 on a weekday. Business scenario: The business type is "illegal transfer", and the preset risk level is "high"; Real-time behavioral characteristics: Normal operation frequency.

[0057] Step 3: The risk scoring calculation module uses the formula Calculate the initial risk score. Since the business type is high-risk (illegal transfer involves liability determination) and the role is administrator, the initial risk score R0 = 72 points (out of 100).

[0058] Step 4: The authentication strategy mapping relationship used in this embodiment is shown in Table 1.

[0059] Table 1: Authentication Policy Mapping Relationship R0=72 falls within the medium-risk range of [40, 75), therefore a "medium-risk authentication strategy"—i.e., two-factor authentication (password + facial recognition)—is matched, and the authentication passes the threshold C. th2 =80 points. That is, users need to complete password verification and facial recognition verification in sequence, and the weighted combined score of the two factors must reach 80 points or above to pass the initial identity authentication.

[0060] Step 5: The multi-factor authentication module sequentially performs password verification and facial recognition. Upon successful initial authentication, the user enters the service processing interface.

[0061] Step 6: As Figure 4 As shown, the continuous behavior monitoring module monitors in real time during the user session. When the user is about to submit an application, the system detects that the current IP address suddenly changes to Shanghai, indicating an abnormal geographical location (environmental change risk increment ΔE=18).

[0062] Step 7: Update the risk score to =72 + 18 = 90 points. (For example...) Figure 4 As shown in the upgrade branch, Entering the high-risk range of [75, 100] triggers an authentication upgrade. The system requires supplementary authentication—namely, three-factor authentication (adding digital certificate verification).

[0063] Step 8: The user inserts their digital certificate to complete verification. If the supplementary authentication passes, the risk score decreases, and the session is maintained. If the supplementary authentication fails, then... Figure 4 As shown, the session is forcibly terminated and the account is locked.

[0064] Step 9: As Figure 3 As shown in the underlying architecture, the audit log storage module packages the entire authentication process (initial scoring, policy switching, environmental anomalies, and supplementary authentication results) and writes it into the blockchain storage to generate tamper-proof audit credentials.

[0065] Example 3 Based on the same inventive concept, this embodiment provides a multi-level adaptive identity verification system based on dynamic risk scoring. The principle of solving the problem is similar to that of the multi-level adaptive identity verification method based on dynamic risk scoring, and the repetitions will not be repeated.

[0066] This embodiment provides a multi-level adaptive identity verification system based on dynamic risk scoring, including: The multi-dimensional information collection module is used to collect multi-dimensional identity verification information in response to user business operation requests. The multi-dimensional identity verification information includes user static attribute information, real-time environmental context information, business scenario information, and real-time behavioral feature information. The risk scoring calculation module is used to determine the user role risk score and historical behavior risk score based on the user's static attribute information, determine the environment risk score based on the real-time environment context information, determine the business risk score based on the business scenario information, and calculate the initial risk score based on the user role risk score, the environment risk score, the business risk score, and the historical behavior risk score. The authentication strategy mapping and decision-making module is used to construct authentication strategy mapping relationships, which represent the mapping relationship between different risk score intervals and authentication strategies. Each authentication strategy includes a combination of authentication factors and an authentication pass threshold. The module determines the corresponding risk score interval based on the initial risk score and determines the authentication strategy based on the risk score interval. It then performs the verification of each authentication factor according to the authentication factor combination corresponding to the authentication strategy and determines the corresponding verification confidence score based on the authentication type and verification result of each authentication factor. The initial identity authentication result acquisition module is used to perform weighted fusion of the corresponding verification confidence scores based on the weight coefficients of each authentication factor to obtain a comprehensive authentication score; and to determine the initial identity authentication result based on the comparison between the comprehensive authentication score and the authentication pass threshold corresponding to the initial authentication strategy. The continuous behavior monitoring module is used to continuously monitor user operation behavior, operation environment and session duration based on the real-time behavior feature information during the user session after the initial identity authentication is passed. Based on the monitoring results, it determines the behavior abnormality score, the risk increment of environmental change and the risk increment of session duration, and updates the initial risk score accordingly to obtain the current risk score. The adjustment module is used to compare the current risk score with the risk score range in the authentication policy mapping table. When the current risk score falls into a higher risk score range, supplementary authentication is triggered. When the current risk score falls back to a lower risk score range, the number of authentication factors required for subsequent business operations is reduced and / or the trigger frequency of supplementary authentication is reduced.

[0067] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0068] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can 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, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0069] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The function specified in one or more boxes.

[0070] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0071] Finally, it should be noted that the above specific embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to examples, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A multi-level adaptive identity verification method based on dynamic risk scoring, characterized in that, include: In response to a user's business operation request, multi-dimensional identity verification information is collected, including user static attribute information, real-time environmental context information, business scenario information, and real-time behavioral feature information. User role risk score and historical behavior risk score are determined based on the user static attribute information, environmental risk score is determined based on the real-time environmental context information, business risk score is determined based on the business scenario information, and an initial risk score is calculated based on the user role risk score, the environmental risk score, the business risk score, and the historical behavior risk score. Construct a certification strategy mapping relationship, which represents the mapping relationship between different risk score intervals and certification strategies. Each certification strategy includes a combination of certification factors and a certification pass threshold. The corresponding risk score range is determined based on the initial risk score, and the certification strategy is determined based on the risk score range. The verification of each certification factor is performed according to the certification factor combination corresponding to the certification strategy, and the corresponding verification confidence score is determined based on the certification type and verification result of each certification factor. The corresponding verification confidence scores are weighted and fused based on the weight coefficients of each certification factor to obtain the comprehensive certification score; The initial identity authentication result is determined based on the comparison between the comprehensive authentication score and the authentication pass threshold corresponding to the initial authentication strategy. During the user session after initial identity authentication, the user's operation behavior, operation environment and session duration are continuously monitored based on the real-time behavioral feature information. The abnormal behavior score, environmental change risk increment and session duration risk increment are determined according to the monitoring results, and the initial risk score is updated accordingly to obtain the current risk score. The current risk score is compared with the risk score range in the authentication policy mapping table. When the current risk score falls into a higher risk score range, supplementary authentication is triggered. When the current risk score falls back to a lower risk score range, the number of authentication factors required for subsequent business operations is reduced and / or the frequency of supplementary authentication is reduced.

2. The multi-level adaptive identity verification method based on dynamic risk scoring according to claim 1, characterized in that, User role risk score and historical behavior risk score are determined based on the user's static attribute information; environmental risk score is determined based on the real-time environmental context information; business risk score is determined based on the business scenario information; and an initial risk score is calculated based on the user role risk score, the environmental risk score, the business risk score, and the historical behavior risk score, including: User role risk score is determined based on the role hierarchy in the user's static attribute information. The role hierarchy includes legal entities, administrators, and ordinary users, and historical behavior risk scores are determined based on historical authentication records in the user's static attribute information. ; Environmental risk is determined based on at least one of the following: geographical location, IP address, network environment, operation time, and device fingerprint, from the real-time environmental context information. ; The business risk score is determined based on at least one of the following: business type, business sensitivity level, and operational impact scope from the business scenario information: ; The initial risk score is calculated using a weighted summation model, and the calculation formula is as follows: ; in, For the initial risk score, , , , These are the weighting coefficients corresponding to the user role risk score, the environment risk score, the business risk score, and the historical behavior risk score, respectively.

3. The multi-level adaptive identity verification method based on dynamic risk scoring according to claim 1, characterized in that, Based on the initial risk score, a corresponding risk score range is determined, and based on the risk score range, an authentication strategy is determined, including: The risk scoring interval includes a first risk scoring interval, a second risk scoring interval, and a third risk scoring interval; When the initial risk score is lower than the first risk threshold but greater than 0, it is determined that the initial risk score is within the first risk score range, and a low-risk authentication strategy using single-factor authentication is matched. When the initial risk score is greater than or equal to the first risk threshold and lower than the second risk threshold, it is determined that the initial risk score is in the second risk score range, and a medium-risk authentication strategy using two-factor authentication is matched. When the initial risk score is greater than or equal to the second risk threshold, the initial risk score is determined to be within the third risk score range, and a high-risk authentication strategy using three or more authentication factors is matched.

4. The multi-level adaptive identity verification method based on dynamic risk scoring according to claim 1, characterized in that, The initial identity authentication result is determined based on the comparison between the comprehensive authentication score and the authentication pass threshold corresponding to the initial authentication strategy, including: Certification factors are divided into binary judgment type certification factors and continuous scoring type certification factors; For the binary decision-type authentication factor, the verification result is determined based on the matching result between the authentication information and the preset verification conditions. When the verification passes, the corresponding verification confidence score is assigned. The score is set at 100. If the verification fails, the corresponding verification confidence score will be adjusted. The score is set to 0. The binary judgment authentication factor includes at least one of password verification, SMS verification code verification, digital certificate verification, and dynamic password verification. For the continuously rated authentication factor, the matching score output by the verification algorithm of the continuously rated authentication factor is determined as the corresponding verification confidence score. The matching score ranges from 0 to 100, and the continuous scoring authentication factor includes face recognition. The verification confidence scores of each authentication factor involved in the initial identity authentication are weighted and fused according to the following formula to obtain the comprehensive authentication score. : ; in, To achieve a comprehensive certification score, Let i be the weight of the i-th authentication factor. Let be the verification confidence score of the i-th authentication factor, and n be the number of authentication factors participating in the current authentication strategy; When a pre-set authentication factor in the authentication strategy is not involved in the current initial identity authentication, the weight coefficient corresponding to that authentication factor is proportionally allocated to other authentication factors involved in the current initial identity authentication. When the comprehensive authentication score C0 is greater than or equal to the authentication pass threshold corresponding to the authentication strategy, the initial identity authentication is determined to be successful. When the overall authentication score C0 is less than the authentication pass threshold, it is determined that the initial identity authentication has failed.

5. The multi-level adaptive identity verification method based on dynamic risk scoring according to claim 1, characterized in that, Based on the monitoring results, determine the behavioral anomaly score, environmental change risk increment, and session duration risk increment, and update the initial risk score accordingly to obtain the current risk score, including: calculating the current risk score according to the following formula: ; in, Assess the current risk level. Scoring for behavioral abnormalities For the initial risk score, As an increase in environmental change risk, This represents an increase in the risk associated with session duration.

6. The multi-level adaptive identity verification method based on dynamic risk scoring according to claim 5, characterized in that, Behavioral abnormality score Calculation methods include: Within a unit time window, outlier operation frequencies are calculated based on the user's historical average number of operations, the standard deviation of historical operation counts, and the current actual number of operations. The calculation formula is as follows: ; in, This represents the actual number of operations performed within the current time window. This represents the user's historical average number of operations within a given time window. The standard deviation of the historical number of operations; When a user skips the normal steps and directly accesses sensitive functions, the operation path will be an anomaly. Set to 30, otherwise set to 0; When the current operation time falls within the preset non-working hours or statutory holidays, the operation time will be considered an abnormal value. Set to 20, otherwise set to 0; The formula for calculating the behavioral abnormality score is as follows: ; in, A score is given for abnormal behavior, with a maximum score of 40.

7. The multi-level adaptive identity verification method based on dynamic risk scoring according to claim 5, characterized in that, Increased risk of environmental change The calculation methods include: Construct an environmental feature vector consisting of IP address range, geolocation code, device fingerprint, and network type, and compare the current session's environmental feature vector with the user's historical trusted environmental feature vector. When the city corresponding to an IP address range changes, the incremental risk of IP address change is determined. When a device fingerprint changes, the incremental risk of device change is determined; When a sudden change in geographical location occurs, the incremental risk of geographical location change should be determined. Determine the risk increment of network change when the network type changes; The formula for calculating the incremental environmental change risk is as follows: ; in, As an increase in environmental change risk, This is to mitigate the increased risk associated with IP address changes. This is to account for the increased risk associated with equipment changes. The risk of changes in geographical location increases. The risk increment is determined by network changes, with an upper limit of 50 points for environmental change risks.

8. The multi-level adaptive identity verification method based on dynamic risk scoring according to claim 5, characterized in that, Incremental risk of session duration The calculation methods include: Set session security duration threshold Maximum session duration threshold Based on the duration of the current session Determine the incremental risk of session duration ;in: hour, ; hour, ; hour, .

9. The multi-level adaptive identity verification method based on dynamic risk scoring according to claim 1, characterized in that, When the current risk score enters a higher risk score range, supplementary authentication is triggered, including: The types of certification factors to be added are determined based on the types of risk sources that cause the current risk score to rise, including: Based on at least one of the sources of risk from environmental change risk increment, the sources of risk from abnormal behavior scoring, and the business risk score corresponding to the current business scenario, candidate authentication factors corresponding to the risk source type are determined from the authentication factor library; Exclude authentication factors that have already been verified in the current session, and determine at least one supplementary authentication factor from the remaining candidate authentication factors based on the number of authentication factors required for the current risk level. If the current risk score, after supplementary authentication, is still higher than the maximum risk threshold, the current session will be terminated and the user account will be locked. Wherein, when the incremental risk of environmental change reaches a preset environmental anomaly threshold, and the incremental risk of environmental change is generated by changes in device fingerprint, at least one of digital certificate authentication and facial recognition authentication is determined as a candidate authentication factor. When the incremental risk of environmental change is caused by changes in IP address or geographical location, at least one of face recognition and liveness detection will be identified as a candidate authentication factor. When the behavior anomaly score reaches a preset behavior anomaly threshold, candidate authentication factors are determined based on the behavior anomaly type that generated the behavior anomaly score. Specifically, when the behavior anomaly type is an operation frequency anomaly, dynamic password authentication is determined as a candidate authentication factor, and when the behavior anomaly type is an operation path anomaly, facial recognition authentication is determined as a candidate authentication factor. When the business risk score corresponding to the current business scenario reaches the preset business risk threshold, and no abnormal environmental changes or behavioral abnormalities are detected, digital certificate authentication will be determined as a candidate authentication factor.

10. A multi-level adaptive identity verification system based on dynamic risk scoring, characterized in that, include: The multi-dimensional information collection module is used to collect multi-dimensional identity verification information in response to user business operation requests. The multi-dimensional identity verification information includes user static attribute information, real-time environmental context information, business scenario information, and real-time behavioral feature information. The risk scoring calculation module is used to determine the user role risk score and historical behavior risk score based on the user's static attribute information, determine the environment risk score based on the real-time environment context information, determine the business risk score based on the business scenario information, and calculate the initial risk score based on the user role risk score, the environment risk score, the business risk score, and the historical behavior risk score. The certification strategy mapping and decision module is used to construct certification strategy mapping relationships, which represent the mapping relationship between different risk score intervals and certification strategies. Each certification strategy includes a combination of certification factors and a certification pass threshold. The corresponding risk score range is determined based on the initial risk score, and the certification strategy is determined based on the risk score range. The verification of each certification factor is performed according to the certification factor combination corresponding to the certification strategy, and the corresponding verification confidence score is determined based on the certification type and verification result of each certification factor. The initial identity authentication result acquisition module is used to perform weighted fusion of the corresponding verification confidence scores based on the weight coefficients of each authentication factor to obtain the comprehensive authentication score; The initial identity authentication result is determined based on the comparison between the comprehensive authentication score and the authentication pass threshold corresponding to the initial authentication strategy. The continuous behavior monitoring module is used to continuously monitor user operation behavior, operation environment and session duration based on the real-time behavior feature information during the user session after the initial identity authentication is passed. Based on the monitoring results, it determines the behavior abnormality score, the risk increment of environmental change and the risk increment of session duration, and updates the initial risk score accordingly to obtain the current risk score. The adjustment module is used to compare the current risk score with the risk score range in the authentication policy mapping table. When the current risk score falls into a higher risk score range, supplementary authentication is triggered. When the current risk score falls back to a lower risk score range, the number of authentication factors required for subsequent business operations is reduced and / or the trigger frequency of supplementary authentication is reduced.