Clinical data acquisition and management system

Abstract

The invention relates to the technical field of data acquisition and management, and discloses a clinical data acquisition and management system which comprises a data acquisition and integration module, a data verification and cleaning module, a security storage and management module, a report generation and export module, a user authority and audit module and a training and support module. The data acquisition and integration module is responsible for collecting clinical data and supporting multiple data input modes and equipment integration; and the data verification and cleaning module verifies and cleans the collected data, and identifies and processes abnormal data. The Bayesian network is introduced into the data acquisition and integration module for data consistency and correctness verification, and the Kalman filtering algorithm is used for fusion processing of equipment data, so that the accuracy and consistency of clinical data are effectively improved, data entry errors are reduced, and the accuracy and accuracy of the clinical data are improved. And abnormal data detection is carried out in the data verification and cleaning module by using a support vector machine algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of data collection management, and in particular to a clinical data collection management system. Background Art

[0002] In modern medical and clinical research, the collection, effective management and secure storage of clinical data are the key to ensuring research quality and patient safety. However, the existing clinical data collection and management systems limit the overall performance and reliability of the system in terms of data collection, verification, storage, report generation and user support:

[0003] Data input and equipment integration rely on manual entry or simple automation tools, lacking effective fusion processing algorithms, resulting in data redundancy, conflict, consistency and accuracy issues, and a high data entry error rate;

[0004] Traditional data verification and cleaning methods cannot effectively handle abnormal data and high-dimensional data, lack advanced algorithm support, and lack verification of data sequence relationships, which affects data quality;

[0005] Traditional systems lack dynamic permission adjustment mechanisms and transparent audit tracking, and are unable to effectively prevent permission abuse and unauthorized access. The report generation process relies on manual operations or semi-automated tools, which is inefficient and prone to content inconsistencies and errors. There is also a lack of automated compliance checking mechanisms.

[0006] Therefore, those skilled in the art provide a clinical data collection and management system to solve the problems raised in the above background technology. Summary of the invention

[0007] In view of the deficiencies of the prior art, the present invention provides a clinical data collection and management system to solve the problems raised in the above background technology.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: a clinical data collection and management system, including a data collection and integration module, a data verification and cleaning module, a security storage and management module, a report generation and export module and a user authority and audit module;

[0009] The data acquisition and integration module is responsible for collecting clinical data and supports multiple data input methods and equipment integration;

[0010] The data verification and cleaning module verifies and cleans the collected data, identifies and processes abnormal data;

[0011] The secure storage and management module securely stores clinical data, manages data access permissions, and maintains data integrity and availability;

[0012] The report generation and export module generates clinical reports that meet the specifications and supports multiple data export formats to complete data sharing and analysis;

[0013] The user rights and audit module manages user access rights and records and tracks data operations;

[0014] The training and support module provides training and technical support for system operation, so that users can use the system effectively and solve problems quickly.

[0015] Preferably, the data acquisition and integration module includes a data input unit, a device integration unit and a data verification unit; the data input unit supports multiple data input modes; the device integration unit integrates and manages different medical devices; the data verification unit performs preliminary verification during the data input process to prevent data entry errors;

[0016] The data verification and cleaning module includes a data verification unit, a data cleaning unit and an anomaly detection unit; the data verification verifies the collected data; the data cleaning unit cleans incomplete, duplicate or abnormal data, and identifies and processes abnormal items in the data through rules and algorithms; the anomaly detection unit uses advanced algorithms to detect abnormal patterns in the data, providing a basis for data cleaning;

[0017] The secure storage and management module includes a data storage unit, a permission management unit and a permission management unit; the data storage unit is responsible for securely storing clinical data, using encryption technology to protect data and prevent unauthorized access; the permission management unit manages user access rights to data; the data backup unit regularly backs up data and provides data recovery.

[0018] Preferably, the report generation and export module includes a report generation unit, a data export unit and a compliance check unit; the report generation unit generates a clinical research report that complies with industry and regulatory standards, and summarizes and displays analysis results; the data export unit supports multiple data export formats; the compliance check unit performs compliance checks during the report generation process;

[0019] The user authority and audit module includes a user management unit, an audit log unit and an authority review unit; the user management unit creates and manages system users and their roles, assigns corresponding authorities, and controls users' access to different functions and data of the system; the audit log unit records user operations and data modifications and provides detailed audit logs; the authority review unit regularly reviews and adjusts user authorities;

[0020] The training and support module includes a training unit, a technical support unit and a user feedback unit; the training unit provides training materials and online courses for system operation; the technical support unit provides multi-channel technical support to help users solve technical problems during use; the user feedback unit collects user feedback, provides improvement suggestions, and helps the system development team optimize system functions and user experience.

[0021] Preferably, in the data input unit, during the data input process, a Bayesian network is used to verify the consistency and correctness of the data. The Bayesian network formula is:

[0022]

[0023] Among them, X i is the current data point, Parents(X i ) is the same as X i Other data points associated with P(X i |Parents(X i )) is the conditional probability of the current data point given the relevant data points;

[0024] In the device integration unit, Kalman filtering can be used for data fusion. The Kalman filtering formula is as follows:

[0025] Prediction update step:

[0026]

[0027] P k|k-1 =AP k-1|k-1 A T +Q,

[0028] Correction update steps:

[0029] K k =P k|k-1 H T (HP k|k-1 H T +R) -1 ,

[0030]

[0031] P k|k =(IK k H)P k|k-1 ,

[0032] in, is the predicted state vector, P k|k-1 is the predicted covariance matrix, A is the state transfer matrix, B is the control input matrix, u kis the control input vector, Q is the process noise covariance, R is the measurement noise covariance, K k is the Kalman gain, H is the observation matrix, z k is the actual measured value.

[0033] Preferably, in the data verification unit, a support vector machine is used for anomaly detection, and the support vector machine anomaly detection formula is as follows:

[0034] Objective function:

[0035]

[0036] Constraints:

[0037] y i (w T φ(x i )+b)≥1-ξ i ,ξ i ≥0,i=1,…,n,

[0038] Among them, w is the weight vector, b is the bias term, ξ i is the slack variable, C is the penalty parameter, φ(x i ) is the kernel function that maps the input data to a high-dimensional space.

[0039] Preferably, in the data verification unit, a hidden Markov model is used to verify the consistency and order relationship of the data. The hidden Markov model formula is:

[0040] State transition probability:

[0041] P(S t =s j |S t-1 =s i )=A ij ,

[0042] Observation probability:

[0043] P(O t =o k |S t =s j )=B jk ,

[0044] Initial state probability:

[0045] P(S 1 =s i )=π i ,

[0046] Among them, S t represents the hidden state at time t, O t represents the observed value at time t, Aij is the state i To state s j The transition probability, B jk is in state s j Observed below o k The probability of i The initial state is s i The probability of

[0047] In the data cleaning unit, the Lagrange interpolation method is used to deal with the data missing problem, and the missing data is filled by interpolation. The Lagrange interpolation formula is:

[0048] For a given n-bit data point (x 0 ,y 0 ),(x 1 ,y 1 ),…,(x n ,y n ), the interpolation polynomial P(x) is expressed as:

[0049]

[0050] The Lagrangian basis function L i (x) is defined as:

[0051]

[0052] This formula is used to calculate the interpolation polynomial, which is used to predict the missing values.

[0053] Preferably, in the permission management unit, the user permissions are managed based on the role-based access control model, and the mathematical expression of RBAC usually involves the relationship mapping between users, roles and permissions. The basic formula of RBAC is:

[0054] User and role mapping:

[0055]

[0056] Among them, U is the user set, R is the role set, and UA is the mapping relationship between users and roles;

[0057] Role and permission mapping:

[0058]

[0059] Among them, P is the permission set, PA is the mapping relationship between roles and permissions;

[0060] User permissions derivation:

[0061]

[0062] This formula represents the permission set P that user u obtains through his role set R.

[0063] Preferably, in the user management unit, an attribute-based access control model is used to manage the user's access rights. The access control model determines the access rights by combining user attributes, environment attributes, and resource attributes. The access control model formula is:

[0064] The decision rule of the access control model is expressed as:

[0065] Permit=f(U,R,E),

[0066] Among them, U is the user attribute set, R is the resource attribute set, E is the environment attribute set, and f is the decision function;

[0067] The specific access decision is evaluated using the following rules:

[0068]

[0069] The permission control of the access control model dynamically determines the user's access rights by combining multiple attributes.

[0070] Preferably, in the permission review unit, a dynamic permission review algorithm is used to adjust user permissions based on user behavior analysis and risk scoring, and a machine learning algorithm is combined to evaluate the user's permission usage and dynamically adjust permissions. The risk scoring algorithm formula is as follows:

[0071] Behavioral deviation calculation:

[0072]

[0073] Among them, A ij is the behavior data of user i on operation j, is the average behavior of users on operation j, D i is the behavioral deviation of user i;

[0074] Risk score calculation:

[0075]

[0076] Among them, R i is the risk score of user i, max(D) is the maximum value of the user behavior deviation;

[0077] Dynamic permission adjustment:

[0078]

[0079] Among them, P i is the permission of user i, T 1 and T2 is the risk threshold for adjusting permissions.

[0080] Preferably, in the training unit, a personalized learning path recommendation algorithm is used to dynamically recommend the most suitable learning content and path according to the user's learning behavior and history. The personalized learning path recommendation formula is as follows:

[0081] User similarity calculation:

[0082]

[0083] Among them, Sim(u i ,u j ) is user u i and u j The similarity between i,k Is user u i The score of learning content k, Is user u i The average rating of user u i and u j A collection of commonly rated content;

[0084] Recommended content:

[0085]

[0086] Among them, P(C i |U i ) is a given user u i Time Content C i The recommendation probability, P(U i |C i ) is user u i Learning Content C i The conditional probability, P(C i ) is the content C i The prior probability, P(U i ) is user U i The probability distribution of .

[0087] The present invention provides a clinical data collection and management system. It has the following beneficial effects:

[0088] 1. The present invention effectively improves the accuracy and consistency of clinical data and reduces data entry errors by introducing Bayesian network in the data acquisition and integration module to verify data consistency and correctness, and uses Kalman filter algorithm for fusion processing of equipment data. In addition, support vector machine algorithm is used in the data verification and cleaning module to detect abnormal data, and the sequential relationship of data is verified in combination with hidden Markov model to ensure accurate identification and processing of abnormal data, thereby providing a solid foundation for data cleaning.

[0089] 2. The present invention adopts advanced encryption technology to protect data in the security storage and management module, and manages user rights based on the role-based access control model. The system can effectively prevent unauthorized access and ensure the security and compliance of data storage. The user rights and audit module is combined with the audit log storage method and dynamic rights review algorithm of blockchain technology, which can record user operations in detail, regularly review and dynamically adjust user rights, and ensure the transparency and security of system operations.

[0090] 3. In the report generation and export module, the present invention uses a natural language generation algorithm to automatically generate clinical research reports that meet industry and regulatory standards, so as to ensure the compliance of the report content through a compliance check unit, improve the efficiency and quality of report generation, and at the same time introduce a personalized learning path recommendation algorithm to dynamically recommend learning content based on the user's learning behavior, thereby improving the targeted nature of training, quickly locating and solving technical problems encountered by users during use, and improving the user support capabilities of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 is a system diagram of the present invention;

[0092] Figure 2 It is a schematic diagram of the data acquisition and integration module of the present invention;

[0093] Figure 3 It is a schematic diagram of the data verification and cleaning module of the present invention;

[0094] Figure 4 It is a schematic diagram of the security storage and management module of the present invention;

[0095] Figure 5 A schematic diagram of a report generation and export module of the present invention;

[0096] Figure 6 It is a schematic diagram of the user authority and audit module of the present invention;

[0097] Figure 7 It is a schematic diagram of the training and support module of the present invention. DETAILED DESCRIPTION

[0098] In order to make the technical personnel in the technical field understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in combination with the drawings in the embodiment of the present invention. Obviously, the described embodiment is a partial embodiment of the present invention, not a complete embodiment. Based on the embodiment of the present invention, other embodiments obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the present invention.

[0099] The present invention is described in detail below in conjunction with the accompanying drawings:

[0100] Example:

[0101] Please refer to the attached Figure 1 - Attachment Figure 7 The embodiment of the present invention provides a clinical data collection and management system, including a data collection and integration module, a data verification and cleaning module, a security storage and management module, a report generation and export module, a user authority and audit module, and a training and support module; the data collection and integration module is responsible for collecting clinical data and supports multiple data input methods and equipment integration; the data verification and cleaning module verifies and cleans the collected data, identifies and processes abnormal data; the security storage and management module safely stores clinical data, manages data access rights, and maintains data integrity and availability; the report generation and export module generates clinical reports that meet specifications and supports multiple data export formats to complete data sharing and analysis; the user authority and audit module manages user access rights, records and tracks data operations; the training and support module provides system operation training and technical support, so that users can effectively use the system and quickly solve problems.

[0102] The benefits of the data collection and integration module are to improve the flexibility of data collection, integrate and manage a variety of medical devices, realize automated data collection, reduce manual input errors, and improve data collection efficiency and accuracy; the benefits of the data verification and cleaning module are to reduce the risk of data errors and maintain high data quality; the benefits of the secure storage and management module are to prevent unauthorized access and data leakage, flexibly manage user permissions, ensure the security and compliance of data access, and ensure the consistency and availability of data during storage and access; the benefits of the report generation and export module are to reduce manual operations, improve report generation efficiency, and ensure that the report content complies with relevant regulations and standards; the benefits of the user permission and audit module are to enhance the security of the system and prevent abuse of permissions; the benefits of the training and support module are to ensure that users can quickly master the use of the system, improve work efficiency, and enhance user experience.

[0103] The data acquisition and integration module includes a data input unit, an equipment integration unit and a data verification unit; the data input unit supports multiple data input methods; the equipment integration unit integrates and manages different medical devices; the data verification unit performs preliminary verification during the data input process to prevent data entry errors; the data verification and cleaning module includes a data verification unit, a data cleaning unit and anomaly detection unit; data verification verifies the collected data; the data cleaning unit cleans incomplete, duplicate or abnormal data, and identifies and processes abnormal items in the data through rules and algorithms; the anomaly detection unit uses advanced algorithms to detect abnormal patterns in the data, providing a basis for data cleaning; the secure storage and management module includes a data storage unit, a permission management unit and a data backup unit; the data storage unit is responsible for securely storing clinical data, using encryption technology to protect data and prevent unauthorized access; the permission management unit manages user access rights to data; the data backup unit regularly backs up data and provides data recovery.

[0104] The benefits of the data input unit are to reduce errors in the manual input process through automated input and improve the accuracy of data collection; the benefits of the equipment integration unit are to realize automated data collection, reduce data entry time and manual intervention, and avoid data conflicts and redundancy; the benefits of the data verification unit are to ensure the basic correctness of the data through preliminary verification and reduce the workload of subsequent data processing;

[0105] The benefits of the data verification unit are to reduce errors in the data and reduce compliance risks in the subsequent data processing and analysis; the benefits of the data cleaning unit are to automatically identify and process abnormal items in the data through rules and algorithms, reduce manual intervention, and improve work efficiency; the benefits of the anomaly detection unit are to accurately identify abnormal data and provide reliable abnormal data detection results for the data cleaning unit;

[0106] The benefits of the data storage unit are that it uses encryption technology to protect data, prevent unauthorized access and data leakage, ensure data confidentiality, and avoid data damage; the benefits of the permission management unit are that it prevents data abuse and enhances system compliance; the benefits of the data backup unit are that the system can quickly resume normal operation and ensure business continuity.

[0107] The report generation and export module includes a report generation unit, a data export unit and a compliance check unit; the report generation unit generates clinical research reports that comply with industry and regulatory standards, and summarizes and displays analysis results; the data export unit supports multiple data export formats; the compliance check unit performs compliance checks during the report generation process; the user authority and audit module includes a user management unit, an audit log unit and a authority review unit; the user management unit creates and manages system users and their roles, assigns corresponding authorities, and controls user access to different functions and data of the system; the audit log unit records user operations and data modifications, and provides detailed audit logs; the authority review unit regularly reviews and adjusts user authorities; the training and support module includes a training unit, a technical support unit and a user feedback unit; the training unit provides training materials and online courses for system operations; the technical support unit provides multi-channel technical support to help users solve technical problems during use; the user feedback unit collects user feedback, provides improvement suggestions, and helps the system development team optimize system functions and user experience.

[0108] Benefits of the report generation unit: Reduce manual operation time and improve report generation efficiency and accuracy; Benefits of the data export unit: Improve the flexibility of data processing and exchange; Benefits of the compliance inspection unit: Reduce compliance risks, reduce manual errors through automated inspections, and improve the accuracy and credibility of reports;

[0109] Benefits of the user management unit: flexibly allocate permissions, perform refined permission control according to user roles, and improve system security and operational efficiency; Benefits of the audit log unit: provide detailed audit logs to ensure the transparency and traceability of system operations; Benefits of the permission review unit: prevent the abuse of permissions, and timely discover and correct unreasonable permission settings through regular review of permission configurations, thereby reducing security risks;

[0110] The benefits of the training unit are to shorten the learning curve and ensure that users can continue to learn and master new functions of the system. The benefits of the technical support unit are that technical support can respond to user needs in a timely manner to ensure smooth operation of the system and user satisfaction. The benefits of the user feedback unit are to provide important reference for system optimization, help the system development team optimize system functions and user experience, and ensure that the system continues to meet user needs.

[0111] In the data input unit, during the data input process, the Bayesian network is used to verify the consistency and correctness of the data. The Bayesian network formula is:

[0112]

[0113] Among them, X i is the current data point, Parents(X i ) is the same as X i Other data points associated with P(Xi |Parents(X i )) is the conditional probability of the current data point given the relevant data points;

[0114] In the device integration unit, Kalman filtering can be used for data fusion. The Kalman filtering formula is as follows:

[0115] Prediction update step:

[0116]

[0117] P k|k-1 =AP k-1|k-1 A T +Q,

[0118] Correction update steps:

[0119] K k =P k|k-1 H T (HP k|k-1 H T +R) -1 ,

[0120]

[0121] P k|k =(IK k H)P k|k-1 ,

[0122] in, is the predicted state vector, P k|k-1 is the predicted covariance matrix, A is the state transfer matrix, B is the control input matrix, u k is the control input vector, Q is the process noise covariance, R is the measurement noise covariance, K k is the Kalman gain, H is the observation matrix, z k is the actual measured value.

[0123] The Bayesian network formula verifies the logical consistency of the data by calculating the conditional probability of the current data point and its associated data points to determine whether the data point conforms to the expected probability distribution. By analyzing the dependency relationship and conditional probability between the data points, the Bayesian network can identify abnormal or unreasonable data points to ensure the correctness of the data.

[0124] The function of the Kalman filter formula is that the prediction update step predicts the state at the next moment through the current state and control input, and provides a preliminary estimate of the equipment data. The update of the covariance matrix reflects the uncertainty of the prediction. Combined with the process noise and system model, the Kalman filter can dynamically adjust the accuracy of the prediction. Through the measurement update step, the actual measurement data is used to correct the predicted state. The Kalman gain matrix determines the weight of the measurement data and the predicted data, thereby achieving the best fusion of multi-source data and reducing the impact of noise on the results.

[0125] In the data verification unit, support vector machine is used for anomaly detection. The support vector machine anomaly detection formula is as follows:

[0126] Objective function:

[0127]

[0128] Constraints:

[0129] y i (w T φ(x i )+b)≥1-ξ i ,ξ i ≥0,i=1,…,n,

[0130] Among them, w is the weight vector, b is the bias term, ξ i is the slack variable, C is the penalty parameter, φ(x i ) is the kernel function that maps the input data to a high-dimensional space.

[0131] The support vector machine anomaly detection formula distinguishes normal data from abnormal data by finding a hyperplane that maximizes the inter-class interval in a high-dimensional space. When the data is noisy or overlapped, the slack variable is introduced by i , support vector machines can strike a balance between accurate classification and tolerance of errors. In anomaly detection, support vector machines are trained to distinguish normal data from abnormal data points. The objective function is to minimize the norm of the weight vector ∥w∥ 2 , making the model more stable and avoiding overfitting, while processing nonlinear boundaries through kernel functions to accurately identify abnormal data.

[0132] In the data verification unit, the hidden Markov model is used to verify the consistency and order of the data. The hidden Markov model formula is:

[0133] State transition probability:

[0134] P(S t =s j |S t-1 =s i )=A ij ,

[0135] Observation probability:

[0136] P(O t =o k |S t =s j )=B jk ,

[0137] Initial state probability:

[0138] P(S 1 =s i )=π i ,

[0139] Among them, S t represents the hidden state at time t, O t represents the observed value at time t, A ij is the state i To state s j The transition probability, B jk is in state s j Observed below o k The probability of i The initial state is s i probability;

[0140] In the data cleaning unit, the Lagrange interpolation method is used to deal with the problem of missing data. The missing data is filled by interpolation. The Lagrange interpolation formula is:

[0141] For a given n-bit data point (x 0 ,y 0 ),(x 1 ,y 1 ),…,(x n ,y n ), the interpolation polynomial P(x) is expressed as:

[0142]

[0143] The Lagrangian basis function L i (x) is defined as:

[0144]

[0145] This formula is used to calculate the interpolation polynomial, which is used to predict the missing values.

[0146] The role of the hidden Markov model in the data verification unit verifies whether the data changes in the expected sequence and ensures that the logical order of the data is correct. The HMM can infer whether the observed data conforms to the expected hidden state pattern, verify the correctness of the data, and ensure the coherence of the data from the beginning to the end;

[0147] The role of Lagrange interpolation in the data cleaning unit is to construct the interpolation polynomial P(x). Lagrange interpolation can predict the function value at the missing data points, fill the missing data, and construct a smooth polynomial curve through the existing data points to ensure that the interpolated data remains consistent and continuous with the original data.

[0148] In the permission management unit, the role-based access control model is used to manage user permissions. The mathematical expression of RBAC usually involves the relationship mapping between users, roles and permissions. The basic formula of RBAC is:

[0149] User and role mapping:

[0150]

[0151] Among them, U is the user set, R is the role set, and UA is the mapping relationship between users and roles;

[0152] Role and permission mapping:

[0153]

[0154] Among them, P is the permission set, PA is the mapping relationship between roles and permissions;

[0155] User permissions derivation:

[0156]

[0157] This formula represents the permission set P that user u obtains through his role set R.

[0158] User and role mapping formula simplifies the permission management process by mapping users and roles. Roles serve as the middle layer of permissions, allowing administrators to manage user access rights more efficiently.

[0159] Role and permission mapping formula function Through the mapping of roles and permissions, administrators only need to manage the relationship between roles and permissions, without the need for users to configure permissions, which reduces management complexity, improves management efficiency, and reduces the possibility of errors;

[0160] User rights derivation formula works as follows: User rights are derived from their roles, which means that the allocation and modification of user rights can be achieved by adjusting the user's role, thus realizing dynamic rights management.

[0161] In the user management unit, an attribute-based access control model is used to manage user access rights. The access control model determines access rights through a combination of user attributes, environment attributes, and resource attributes. The access control model formula is:

[0162] The decision rule of the access control model is expressed as:

[0163] Permit=f(U,R,E),

[0164] Among them, U is the user attribute set, R is the resource attribute set, E is the environment attribute set, and f is the decision function;

[0165] The specific access decision is evaluated using the following rules:

[0166]

[0167] The permission control of the access control model dynamically determines the user's access rights by combining multiple attributes.

[0168] By combining multiple attributes of users, resources, and the environment, the ABAC model can dynamically determine the user's access rights based on the context. The dynamic nature makes access control more flexible and refined, and can accurately limit or allow access to resources, thereby improving security.

[0169] The access decision formula ensures the clarity and consistency of permission decisions based on the result of the decision function f(U, R, E). The access decision is based on real-time attribute evaluation. The system can dynamically adjust permissions when user attributes, resource attributes or environmental attributes change, ensuring that the security policy is always consistent with the actual situation.

[0170] The overall role of the ABAC model is to determine access rights by combining multi-dimensional attributes. Compared with traditional role-based or other static rule-based models, it provides flexibility and administrators can define complex access rules based on business needs and security policies.

[0171] In the permission review unit, a dynamic permission review algorithm is used to adjust user permissions based on user behavior analysis and risk scoring, and a machine learning algorithm is combined to evaluate the user's permission usage and dynamically adjust permissions. The risk scoring algorithm formula is as follows:

[0172] Behavioral deviation calculation:

[0173]

[0174] Among them, A ij is the behavior data of user i on operation j, is the average behavior of users on operation j, D i is the behavioral deviation of user i;

[0175] Risk score calculation:

[0176]

[0177] Among them, R iis the risk score of user i, max(D) is the maximum value of the user behavior deviation;

[0178] Dynamic permission adjustment:

[0179]

[0180] Among them, P i is the permission of user i, T 1 and T 2 is the risk threshold for adjusting permissions.

[0181] The behavior deviation calculation formula calculates the deviation between the user's behavior data on a specific operation and the overall average behavior, evaluates the difference between the user's behavior and the normal operation mode, and then indicates potential security risks;

[0182] The risk score calculation formula is used to measure the user's risk level in a relative manner by standardizing the user's behavioral deviation. A high score indicates that the user's behavior is abnormal.

[0183] The dynamic permission adjustment formula dynamically adjusts the user's permissions according to the user's risk score. If the risk score is lower than the threshold, the user's permissions can be increased; if the risk score is higher than the threshold, the user's permissions are reduced or restricted. The dynamic adjustment mechanism helps to respond to changes in user behavior in a timely manner and enhance the security of the system.

[0184] In the training unit, a personalized learning path recommendation algorithm is used to dynamically recommend the most suitable learning content and path based on the user's learning behavior and history. The personalized learning path recommendation formula is as follows:

[0185] User similarity calculation:

[0186]

[0187] Among them, Sim(u i ,u j ) is user u i and u j The similarity between i,k Is user u i The score of learning content k, Is user u i The average rating of user u i and u j A collection of commonly rated content;

[0188] Recommended content:

[0189]

[0190] Among them, P(C i |Ui ) is a given user u i Time Content C i The recommendation probability, P(U i |C i ) is user u i Learning Content C i The conditional probability, P(C i ) is the content C i The prior probability, P(U i ) is user U i The probability distribution of .

[0191] The user similarity calculation formula evaluates the similarity of learning preferences by calculating the difference in users' ratings on common learning content. A high similarity indicates that the users' learning habits and interests are similar. User similarity is the basis of the personalized recommendation algorithm. By identifying users similar to the target user, the system can recommend learning content to the target user based on the learning history of similar users.

[0192] The content recommendation formula uses Bayesian reasoning to calculate the probability of a certain learning content being recommended given the user's specific attributes and historical behavior. In this way, the system can dynamically adjust the recommended content to match the user's needs and interests, and at the same time combine the user's personalized data to provide more targeted learning path recommendations.

[0193] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A clinical data collection and management system, characterized in that: It includes data collection and integration module, data verification and cleaning module, security storage and management module, report generation and export module, user authority and audit module and training and support module; The data acquisition and integration module is responsible for collecting clinical data and supports multiple data input methods and equipment integration; The data verification and cleaning module verifies and cleans the collected data, identifies and processes abnormal data; The secure storage and management module securely stores clinical data, manages data access permissions, and maintains data integrity and availability; The report generation and export module generates clinical reports that meet the specifications and supports multiple data export formats to complete data sharing and analysis; The user rights and audit module manages user access rights and records and tracks data operations; The training and support module provides training and technical support for system operation, so that users can use the system effectively and solve problems quickly.

2. A clinical data collection and management system according to claim 1, characterized in that: The data acquisition and integration module includes a data input unit, a device integration unit and a data verification unit; the data input unit supports multiple data input methods; the device integration unit integrates and manages different medical devices; the data verification unit performs preliminary verification during the data input process to prevent data entry errors; The data verification and cleaning module includes a data verification unit, a data cleaning unit and an anomaly detection unit; the data verification verifies the collected data; the data cleaning unit cleans incomplete, duplicate or abnormal data, and identifies and processes abnormal items in the data through rules and algorithms; the anomaly detection unit uses advanced algorithms to detect abnormal patterns in the data, providing a basis for data cleaning; The secure storage and management module includes a data storage unit, a permission management unit and a data backup unit; the data storage unit is responsible for securely storing clinical data, using encryption technology to protect data and prevent unauthorized access; the permission management unit manages user access rights to data; the data backup unit regularly backs up data and provides data recovery.

3. A clinical data collection and management system according to claim 1, characterized in that: The report generation and export module includes a report generation unit, a data export unit and a compliance check unit; the report generation unit generates a clinical research report that complies with industry and regulatory standards, and summarizes and displays analysis results; The data export unit supports multiple data export formats; The compliance checking unit performs compliance checking during the report generation process; The user authority and audit module includes a user management unit, an audit log unit and an authority review unit; The user management unit creates and manages system users and their roles, assigns corresponding permissions, and controls users' access to different functions and data of the system; the audit log unit records user operations and data modifications and provides detailed audit logs; the permission review unit regularly reviews and adjusts user permissions; The training and support module includes a training unit, a technical support unit and a user feedback unit; the training unit provides training materials and online courses for system operation; the technical support unit provides multi-channel technical support to help users solve technical problems during use; The user feedback unit collects user feedback, provides improvement suggestions, and helps the system development team optimize system functions and user experience.

4. A clinical data collection and management system according to claim 2, characterized in that: In the data input unit, during the data input process, the Bayesian network is used to verify the consistency and correctness of the data. The Bayesian network formula is: Among them, X i is the current data point, Parents(X i ) is the same as X i Other data points associated with P(X i |Parents(X i )) is the conditional probability of the current data point given the relevant data points; In the device integration unit, Kalman filtering can be used for data fusion. The Kalman filtering formula is as follows: Prediction update step: P k|k-1 =AP k-1|k-1 From T +Q, Correction update steps: K k =P k|k-1 H T (HP k|k-1 H T +R) -1 , P k|k =(I-K k H)P k|k-1 , in, is the predicted state vector, P k|k-1 is the predicted covariance matrix, A is the state transfer matrix, B is the control input matrix, u k is the control input vector, Q is the process noise covariance, R is the measurement noise covariance, K k is the Kalman gain, H is the observation matrix, z k is the actual measured value.

5. A clinical data collection and management system according to claim 2, characterized in that: In the data verification unit, a support vector machine is used for anomaly detection. The support vector machine anomaly detection formula is as follows: Objective function: Constraints: y i (w T φ(x i )+b)≥1-ξ i ,x i ≥0,i=1,…,n, Among them, w is the weight vector, b is the bias term, ξ i is the slack variable, C is the penalty parameter, φ(x i ) is the kernel function that maps the input data to a high-dimensional space.

6. A clinical data collection and management system according to claim 2, characterized in that: In the data verification unit, a hidden Markov model is used to verify the consistency and order of data. The hidden Markov model formula is: State transition probability: P(S t =s j |S t-1 =s i )=A ij , Observation probability: P(O t =o k |S t =s j )=B jk , Initial state probability: P(S1=s i )=π i , Among them, S t represents the hidden state at time t, O t represents the observed value at time t, A ij is the state i To state s j The transition probability, B jk is in state s j Observed below o k The probability of i The initial state is s i probability; In the data cleaning unit, the Lagrange interpolation method is used to deal with the data missing problem, and the missing data is filled by interpolation. The Lagrange interpolation formula is: For a given n-bit data point (x0, y0), (x1, y1), ..., (x n ,y n ), the interpolation polynomial P(x) is expressed as: The Lagrangian basis function L i (x) is defined as: This formula is used to calculate the interpolation polynomial, which is used to predict the missing values.

7. A clinical data collection and management system according to claim 2, characterized in that: In the permission management unit, the user permissions are managed based on the role-based access control model, and the mathematical expression of RBAC usually involves the relationship mapping between users, roles and permissions. The basic formula of RBAC is: User and role mapping: Among them, U is the user set, R is the role set, and UA is the mapping relationship between users and roles; Role and permission mapping: Among them, P is the permission set, PA is the mapping relationship between roles and permissions; User permissions derivation: This formula represents the permission set P that user u obtains through his role set R.

8. A clinical data collection and management system according to claim 3, characterized in that: In the user management unit, an attribute-based access control model is used to manage user access rights. The access control model determines access rights through a combination of user attributes, environment attributes, and resource attributes. The access control model formula is: The decision rule of the access control model is expressed as: Permit=f(U,R,E), Among them, U is the user attribute set, R is the resource attribute set, E is the environment attribute set, and f is the decision function; The specific access decision is evaluated using the following rules: The permission control of the access control model dynamically determines the user's access rights by combining multiple attributes.

9. A clinical data collection and management system according to claim 3, characterized in that: In the permission review unit, a dynamic permission review algorithm is used to adjust user permissions based on user behavior analysis and risk scoring, and a machine learning algorithm is combined to evaluate the user's permission usage and dynamically adjust permissions. The risk scoring algorithm formula is as follows: Behavioral deviation calculation: Among them, A ij is the behavior data of user i on operation j, is the average behavior of users on operation j, D i is the behavioral deviation of user i; Risk score calculation: Among them, R i is the risk score of user i, max(D) is the maximum value of the user behavior deviation; Dynamic permission adjustment: Among them, P i is the permission of user i, T1 and T2 are the risk thresholds for adjusting permissions.

10. A clinical data collection and management system according to claim 3, characterized in that: In the training unit, a personalized learning path recommendation algorithm is used to dynamically recommend the most suitable learning content and path based on the user's learning behavior and history. The personalized learning path recommendation formula is as follows: User similarity calculation: Among them, Sim(u i ,u j ) is user u i and u j The similarity between i,k Is user u i The score of learning content k, Is user u i The average rating of user u i and u j A collection of commonly rated content; Recommended content: Among them, P(C i |U i ) is a given user u i Time Content C i The recommendation probability, P(U i |C i ) is user u i Learning Content C i The conditional probability, P(C i ) is the content C i The prior probability, P(U i ) is user U i The probability distribution of .

Images