A user information security processing control system for smart medical care

By combining two-factor authentication, quantum communication, blockchain and NMM models and attention mechanism log analysis in smart medical systems, the problems of false alarms and underreporting behavior models are solved, and more accurate abnormal behavior recognition and Data security.

CN118349979BActive Publication Date: 2024-11-19LANZHOU WANQIAO INTELLIGENCE TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410780904.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-18
Publication Date
2024-11-19
Estimated Expiration
2044-06-18

AI Technical Summary

Technical Problem

Existing behavioral models have problems with false alarms and missed reports when distinguishing between normal operations and abnormal behaviors, resulting in overly sensitive models that may lead to a large number of false alarms, while being too loose may miss the real red flag.

Method used

Using identity authentication authorization unit, data transmission encryption unit, blockchain trust network unit and log analysis management unit, combining two-factor authentication, quantum communication technology, blockchain and NMM models and attention mechanisms to realize user identity authentication and data Encryption and abnormal behavior recognition.

Benefits of technology

It improves the accuracy of user behavior recognition, ensures the security of data transmission and the authenticity of data, reduces false alarms, and enhances the security and reliability of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118349979B_ABST
    Figure CN118349979B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of smart medical technology, and specifically, to a user information security processing and control system applied to smart medical care. It includes an identity authentication authorization unit, a data transmission encryption unit, a blockchain trust network unit, and a log analysis management unit; wherein the log analysis management unit includes a log collection and processing module, a storage index module, a behavior learning module, and a monitoring and early warning module. This user information security processing and control system applied to smart medical care can automatically focus on the part of the log data that is most relevant to the current analysis task, ignoring irrelevant information, thereby improving the recognition accuracy of abnormal access or operation behavior, and at the same time can better adapt to changes in user behavior and system development, and at the same time, can continuously optimize model parameters through learning, without the need to manually adjust rules frequently.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of smart medical technology, and in particular to a user information security processing and control system applied to smart medical care. Background Art

[0002] With the rapid development of smart healthcare, a huge amount of medical data is generated and collected, including highly sensitive data such as personal health information, genetic information, and treatment records. The storage, processing, and transmission of these data require extremely high security to prevent data leakage or illegal use. The application of advanced technologies such as cloud computing, big data, and artificial intelligence in the field of smart healthcare has greatly improved the efficiency and quality of medical services, but it has also introduced new security threats. For example, centralized storage of data in a cloud computing environment increases the risk of data leakage, while big data analysis may expose personal privacy. The means of network attacks are constantly evolving, and targeted malware, ransomware attacks, and improper behavior of internal personnel all pose a serious threat to the security of smart medical systems. Once patient data is stolen, it will not only damage personal privacy, but may also affect the quality of medical services and even public safety. Building an efficient and reliable user information security processing and control system, using technical means such as data encryption, anonymization, multi-layer access control, security auditing, and timely response mechanisms, has become an indispensable part of the field of smart healthcare to ensure the safe circulation and legal use of data and maintain the trust between doctors and patients.

[0003] Existing behavior models may have problems with false positives and false negatives when distinguishing normal operations from abnormal behaviors. The legal operation range of medical staff is wide and flexible, and the normal working modes are diverse, making it difficult to define abnormal behaviors. An overly sensitive model may lead to a large number of false positive warnings, while an overly loose model may miss real danger signals. Therefore, a user information security processing and control system for smart medical care is designed. Summary of the invention

[0004] The purpose of the present invention is to provide a user information security processing and control system applied to smart medical care to solve the problems of false positives and false negatives in distinguishing normal operations from abnormal behaviors raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides a user information security processing and control system applied to smart medical care, comprising:

[0006] Authentication and authorization unit: used to verify the identity of users using a two-factor authentication security mechanism and assign user access rights based on access control lists;

[0007] Data transmission encryption unit: used to encrypt the medical data transmitted by users using the transport layer security protocol, and based on quantum communication technology, to achieve quantum key distribution, providing a theoretically unconditionally secure key exchange method for data encryption;

[0008] Blockchain Trust Network Unit: used to encrypt and add each piece of medical data as a transaction record to the blockchain to ensure the authenticity and integrity of the data;

[0009] Log analysis management unit: used to record the operation logs of all users, centrally manage and analyze log data, and use the NMM model combined with the attention mechanism to analyze user behavior patterns, automatically identify abnormal access or operation behaviors, and promptly warn of potential security threats;

[0010] Among them, the log analysis management unit includes a log collection and processing module, a storage index module, a behavior learning module and a monitoring and early warning module.

[0011] As a further improvement of the present technical solution, the two-factor identity authentication in the identity authentication authorization unit includes password verification and fingerprint recognition verification; the access control list can enable the administrator to customize access rules for user groups and specific users. When the verification is completed, the system allocates user access rights based on the access control list.

[0012] As a further improvement of the technical solution, the specific steps of the data transmission encryption unit are as follows:

[0013] S21: The data sender and the data receiver both prepare two quantum bit sequences for sending quantum states;

[0014] S22: The data sender prepares quantum bits, uses quantum states to represent different bit values, and sends the prepared quantum bits to the data receiver;

[0015] S23: After receiving the quantum bits sent by the data sender, the data receiver performs measurement to determine the state of each quantum bit;

[0016] S24: The data sender and the data receiver retain a portion of the quantum bit information during the communication process, and compare them through a public transmission channel after the communication is completed, and determine whether there is an eavesdropper by comparing the measurement results;

[0017] S25: If no eavesdropper is found during the comparison process, the data sender and the data receiver extract a shared key from the remaining matching quantum bits;

[0018] S26: The data sender and the data receiver use the quantum key to encrypt the communication data;

[0019] S27: The data sender and the data receiver use a portion of the extracted key for verification and perform a security analysis on the entire quantum communication process.

[0020] As a further improvement of this technical solution, the specific steps of the blockchain trust network unit are as follows:

[0021] S31: Encapsulate the encrypted medical data and its metadata into a transaction;

[0022] S32: Verify the legitimacy of transactions through the consensus mechanism of the blockchain network;

[0023] S33: The verified transactions are packaged into a new block, which contains the hash value of the previous block, forming a chain structure;

[0024] S34: The new block is broadcast to the entire network, and each participating node updates its local copy of the blockchain.

[0025] As a further improvement of the present technical solution, in S31, the transaction includes a contract code, specified data access rights and data usage rules.

[0026] As a further improvement of the technical solution, in the log analysis management unit,

[0027] The log collection and processing module is used to automatically collect log data from various system components to ensure comprehensive coverage of user activity records, and to clean, standardize and structure the collected data;

[0028] The storage index module is used to efficiently store the processed log data in a database or big data platform and create indexes for fast query;

[0029] The behavior learning module is used to build user behavior models through historical log data to improve the accuracy of anomaly detection and reduce false positives;

[0030] The monitoring and early warning module is used to continuously monitor the log data stream. Once abnormal behavior or potential security threats are detected, the early warning system is immediately triggered and the administrator is notified.

[0031] As a further improvement of this technical solution, the specific steps of the behavior learning module are as follows:

[0032] S431: extract useful features based on the log content, and sort the user behaviors by time to form time series data;

[0033] S432: using sequence pattern mining technology to identify regular paths and time interval patterns of user behaviors, and dividing users into different behavior groups based on user behavior characteristics;

[0034] S433: Count the frequency distribution of various operations and identify common and rare behaviors;

[0035] S434: Based on HMM and combined with the attention mechanism, further screen and construct time window features and behavior frequency features, and use historical log data to train the model to learn the statistical characteristics and time series laws of normal user behaviors;

[0036] S435: Apply the trained model to the new log data, calculate the anomaly score of the behavior, set the anomaly threshold according to the model output and actual needs, and distinguish between normal and abnormal behaviors.

[0037] As a further improvement of the present technical solution, in S431, useful features include login time, operation type, operation frequency and stay time.

[0038] As a further improvement of the technical solution, in S434, the calculation steps of HMM are:

[0039] Forward Probability:

[0040]

[0041] in, For in time In state And generate observation sequence from arrive probability; is the parameter set of HMM; for observation sequence; For time; A certain status of the user; For time Status;

[0042] Backward Probability:

[0043]

[0044] in, For in time In state And generate observation sequence from arrive The probability of represents the total length of the observation sequence;

[0045] At each time step , introduce the attention mechanism to calculate the weight vector , which reflects the generation of the current observation When , the importance of each past state is taken into account. Therefore, the expressions of forward probability and backward probability are:

[0046]

[0047]

[0048] in, The adjusted time In state And generate observation sequence from arrive probability; The adjusted time In state And generate observation sequence from arrive probability; For in time Time Status The attention weight of For in time Time Status The attention weight of From the state Transfer to state probability; For the status Observed below probability; For the status Observed below probability.

[0049] As a further improvement of the technical solution, the probability of state and update is calculated according to the forward probability and the backward probability, and the calculation steps are as follows:

[0050] Status probability:

[0051]

[0052] in, For in time When in state probability; is the number of states; , It's in time In state And generate observation sequence from 1 to The probability of It's in time In state And generate observation sequence from arrive probability;

[0053] State transition probability:

[0054]

[0055] in, For at the moment arrive Between, from the state Transfer to state probability; From the state Transfer to state probability; For the status Observed below probability; For in time In state And generate observation sequence from arrive probability; For in time In state And generate observation sequence from arrive probability; From the state Transfer to state probability; For the status Observed below probability; For in time In state And generate observation sequence from arrive probability; ; ;

[0056] Update the initial state probability:

[0057]

[0058] in, The initial state is probability; for When in state probability;

[0059] Update state transition probability:

[0060]

[0061] Update observation probabilities:

[0062]

[0063] in, For the status Observed below probability; For in time When in state The probability of Represents a time instant before the last time instant in the observation sequence.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] 1. This user information security processing and control system applied to smart medical care uses the NMM model and combines it with the attention mechanism to analyze user behavior. The attention mechanism can automatically focus on the part of the log data that is most relevant to the current analysis task and ignore irrelevant information, thereby improving the recognition accuracy of abnormal access or operation behavior. The log analysis system combining the NMM model and the attention mechanism can better adapt to changes in user behavior and system development. At the same time, it can continuously optimize model parameters through learning without the need to manually adjust the rules frequently.

[0066] 2. In this user information security processing and control system applied to smart medical care, quantum key distribution is used in the process of data transmission. In this way, any attempt to eavesdrop on the key exchange process will be discovered immediately, because any measurement of the quantum state will change its state, thereby ensuring the absolute security of the key. Combined with the transport layer security protocol, end-to-end encryption can be achieved from the data source to the destination, greatly enhancing the security of the data transmission process. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 It is a flowchart of the whole process of the present invention;

[0068] Figure 2 This is a system block diagram of the log analysis management unit in the present invention.

[0069] The meaning of each number in the figure is:

[0070] 1. Identity authentication and authorization unit; 2. Data transmission encryption unit; 3. Blockchain trust network unit; 4. Log analysis and management unit; 41. Log collection and processing module; 42. Storage index module; 43. Behavior learning module; 44. Monitoring and early warning module. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Example

[0072] See also Figure 1-2 As shown, a user information security processing and control system applied to smart medical care is provided, comprising:

[0073] Identity authentication and authorization unit 1: used to verify the identity of the user using a two-factor authentication security mechanism and assign the user's access rights based on the access control list;

[0074] Two-factor authentication includes password verification and fingerprint recognition verification. The system greatly improves the reliability and security of user identity verification. Password verification provides a layer of cognitive authentication, while fingerprint recognition, as a biometric verification, adds physical protection to identity authentication, effectively preventing the risk of password theft or simple guessing.

[0075] Access control lists allow administrators to customize access rules for user groups and specific users. Once verification is completed, the system assigns user access rights based on the access control list. Access control lists allow administrators to flexibly configure access rights based on user roles, responsibilities, or specific needs. This not only sets common rules for different user groups, but also provides customized permissions for individual users, ensuring that each user can only access necessary information related to their work, reducing the risk of data leakage caused by excessive authorization of permissions.

[0076] Through strict identity authentication and access control, the system can better comply with the relevant laws and regulations of the medical industry regarding patient privacy protection. At the same time, detailed access log records help with post-audits, track any improper access behavior, and enhance the transparency and accountability traceability of the system.

[0077] Data transmission encryption unit 2: used to encrypt the medical data transmitted by the user using the transport layer security protocol, and based on quantum communication technology, realize quantum key distribution, and provide a theoretically unconditionally secure key exchange method for data encryption; the specific steps are as follows:

[0078] S21: The data sender and the data receiver both prepare two quantum bit sequences for sending quantum states;

[0079] S22: The data sender prepares quantum bits and uses quantum states to represent different bit values. For example, the horizontal and vertical polarizations of single photons can be used to represent different bit values ​​and the prepared quantum bits are sent to the data receiver. A secure channel is required to ensure that the quantum bits are not stolen or tampered with during the transmission process.

[0080] S23: After receiving the quantum bits sent by the data sender, the data receiver performs measurement to determine the state of each quantum bit;

[0081] S24: The data sender and the data receiver retain a portion of the quantum bit information during the communication process, and compare them through a public transmission channel after the communication is completed, and determine whether there is an eavesdropper by comparing the measurement results;

[0082] S25: If no eavesdropper is found during the comparison process, the data sender and the data receiver extract the shared key from the remaining matching quantum bits and use the properties of the quantum state to ensure the security of the key;

[0083] S26: The data sender and the data receiver use quantum keys to encrypt communication data; commonly used quantum encryption algorithms include BB84 algorithm and E91 algorithm. These algorithms use the characteristics of quantum states to provide stronger security than traditional encryption algorithms and can resist possible quantum computer attacks in the future;

[0084] In the BB84 protocol, the data sender and the data receiver extract the key by comparing a portion of matching quantum bits. If the states of the two bits are the same, then the corresponding bit value can be used as part of the key. For example, if the data sender sends a vertical polarization state and the data receiver also measures a vertical polarization state, then the corresponding bit value can be used as part of the key;

[0085] In the E91 protocol, the data sender and data receiver use quantum entanglement to distribute keys.

[0086] S27: The data sender and the data receiver use a portion of the extracted key for verification to ensure the correctness and integrity of the key, and perform a security analysis on the entire quantum communication process to ensure that the system can resist various potential attacks.

[0087] Quantum key distribution uses the principles of quantum mechanics to ensure that even in theory, the key cannot be eavesdropped or copied by a third party without being detected. Through S21 to S25, the data sender and receiver can generate a shared key that is only known to both parties and is completely secure, providing long-term security against the potential threat of quantum computers compared to traditional encryption methods.

[0088] In step S26, the communication data is encrypted using the security key obtained from quantum key distribution, and the medical data is encrypted in transmission in combination with the transport layer security protocol. This two-pronged approach ensures the confidentiality and integrity of the data during transmission, and effectively prevents the data from being intercepted or tampered with in the middle.

[0089] In step S27, by using a part of the key for verification, the system can self-verify the correctness and integrity of the key, ensuring the reliability of the key distribution and data encryption process. In addition, security analysis of the entire quantum communication process helps to identify and prevent potential security vulnerabilities, further strengthening the security of the system.

[0090] Blockchain trust network unit 3: used to encrypt each piece of medical data as a transaction record and add it to the blockchain to ensure the authenticity and integrity of the data; the specific steps are as follows:

[0091] S31: Encapsulate the encrypted medical data and its metadata into a transaction and sign the transaction. The signature ensures the integrity and immutability of the data, and only the data owner can sign the data; the transaction includes the contract code, data access rights, and data usage rules.

[0092] S32: Verify the legitimacy of the transaction through the consensus mechanism of the blockchain network; send the transaction to the blockchain network and wait for blockchain node verification. The blockchain node will verify the legitimacy of the transaction, including whether the signature is valid and whether the transaction meets the rules of the blockchain network.

[0093] S33: Verified transactions are packaged into a new block, which contains the hash value of the previous block to form a chain structure; ensuring distributed storage and consensus consistency of data.

[0094] S34: The new block is broadcast to the entire network, and each participating node updates its local copy of the blockchain. This ensures that all nodes reach a consensus on the state of the blockchain. Only blocks that have passed consensus can be added to the blockchain, thus ensuring the authenticity and immutability of the data;

[0095] Data owners or authorized users can access and verify the authenticity and integrity of data through the blockchain network, and ensure that only legitimate users can access the data by providing corresponding private keys or authorization mechanisms.

[0096] Log analysis management unit 4: used to record the operation logs of all users, centrally manage and analyze the log data, and adopt the NMM model in combination with the attention mechanism to analyze user behavior patterns, automatically identify abnormal access or operation behaviors, and timely warn of potential security threats; the log analysis management unit 4 includes a log collection and processing module 41, a storage index module 42, a behavior learning module 43 and a monitoring and warning module 44.

[0097] The log collection and processing module 41 is used to automatically collect log data from various system components to ensure comprehensive coverage of user activity records, and to clean, standardize and structure the collected data; ensure that the log information format from different sources is consistent, which is convenient for centralized management and analysis;

[0098] The storage index module 42 is used to efficiently store the processed log data in a database or a big data platform and create an index for quick query; creating an index speeds up the query speed so that a specific query or pattern matching operation can be completed in a short time, improving the response speed, ensuring the long-term storage of log data, and meeting compliance and audit requirements;

[0099] The behavior learning module 43 is used to build a user behavior model through historical log data to improve the accuracy of anomaly detection and reduce false positives;

[0100] The monitoring and early warning module 44 is used to continuously monitor the log data flow, apply the model output by the behavior learning module, and identify operations that deviate from normal behavior patterns, such as illegal login attempts, abnormal access patterns, etc. Once abnormal behavior or potential security threats are detected, the early warning system is immediately triggered and the administrator is notified. According to preset rules, corresponding security response measures are automatically executed or recommended, such as restricting access, starting backup processes, etc.

[0101] Furthermore, the specific steps of the behavior learning module 43 are as follows:

[0102] S431: Extract useful features based on log content to ensure that analysis focuses on core security-related indicators, and sort user behaviors by time to form time series data to help capture the changing trends of behaviors over time and provide a time dimension for understanding behavior patterns;

[0103] Useful features include login time, action type, action frequency, and duration of stay.

[0104] Among them, analyzing the distribution of user login time can identify the user's active time period. For example, some users may log in frequently during working hours on weekdays, while other users may prefer to access at night or on weekends. This helps to understand user behavior habits and optimize system resource allocation. Abnormal login time, such as login late at night or during non-normal working hours, may indicate the risk of unauthorized access or account theft, which requires further investigation;

[0105] Different users have different operation preferences based on their roles and permissions. For example, doctors are more likely to view medical records and prescribe prescriptions, while administrators may be involved in system configuration changes. By analyzing the operation type, you can identify the user's professional characteristics or permission usage; if a user performs an operation that is inconsistent with his or her role, such as an ordinary user trying to access the administrator interface, this may be a signal of a security violation;

[0106] Frequent operations may indicate an efficient workflow, or they may indicate illegal use of automated scripts. Monitoring the frequency of operations can help identify abnormal activities, such as large amounts of data accessed in a short period of time, which may indicate data theft. High-frequency operations may consume more system resources, and understanding the frequency of operations can help rationally plan system load and performance optimization.

[0107] The length of time a user spends on a particular page or feature can reflect their interest or the difficulties they encounter. A long stay may mean that the operation is complex or the user is looking for information, while a short stay may mean that the operation is smooth or the goal is clear; abnormally long stays, such as long periods of inactivity after logging in, may indicate that the account has been forgotten or maliciously occupied, while abnormally short stays and quick logins and exits may be related to scanning or exploratory attacks.

[0108] S432: using sequence pattern mining technology to identify regular paths and time interval patterns of user behaviors, and dividing users into different behavior groups based on user behavior characteristics;

[0109] S433: Count the frequency distribution of various operations, identify common and rare behaviors, and provide statistical basis for abnormal behavior identification;

[0110] S434: Based on HMM, further screen and construct time window features and behavior frequency features, and use historical log data to train the model to learn the statistical characteristics and time series laws of normal user behaviors;

[0111] S435: Apply the trained model to the new log data, calculate the anomaly score of the behavior, set the anomaly threshold according to the model output and actual needs, and distinguish between normal and abnormal behaviors.

[0112] Among them, the calculation steps of HMM are:

[0113] Forward Probability:

[0114]

[0115] in, For in time In state And generate observation sequence from arrive probability; is the parameter set of HMM, including the initial state probability vector, state transfer matrix and observation probability matrix; for observation sequence; For time; A certain status of the user; For time Status;

[0116] Backward Probability:

[0117]

[0118] in, For in time In state And generate observation sequence from arrive The probability of represents the total length of the observation sequence;

[0119] At each time step , introduce the attention mechanism to calculate the weight vector , which reflects the generation of the current observation When , the importance of each past state is enhanced, and the model's ability to focus on key time points or states in the sequence is enhanced. Therefore, the expressions of forward probability and backward probability are:

[0120]

[0121]

[0122] in, The adjusted time In state And generate observation sequence from arrive probability; The adjusted time In state And generate observation sequence from arrive probability; For in time Time Status The attention weight of When, the past state By introducing the attention mechanism, the model can dynamically emphasize the historical states that are more critical to prediction, thereby improving the ability to capture key information in the sequence;

[0123] For in time Time Status The attention weight of The relative importance of , which corresponds to the positive attention mechanism in the forward probability, works together on the entire sequence, helping the model pay more attention to the state transitions and observations in the sequence that are most critical to prediction;

[0124] From the state Transfer to state probability; For the status Observed below probability; For the status Observed below probability.

[0125] The probability of state and update is calculated based on the forward probability and backward probability. The calculation steps are as follows:

[0126] Status probability:

[0127]

[0128] in, For in time When in state The probability of and the backward probability The product of and normalized by the sum of the probabilities of all states provides Status at all times The global importance measure of is the number of states; , It's in time In state And generate observation sequence from 1 to The probability of It's in time In state And generate observation sequence from arrive probability;

[0129] To calculate the time step When in position, in state The probability of the state, called the state probability, combines the information of the forward probability and the backward probability to provide the time point A global importance measure of the sequence of observations in a state, normalized to ensure that the sum of all state probabilities is 1, which helps understand the relative significance of each state in the sequence;

[0130] State transition probability:

[0131]

[0132] in, For at the moment arrive Between, from the state Transfer to state probability; From the state Transfer to state probability; For the status Observed below probability; For in time In state And generate observation sequence from arrive probability; For in time In state And generate observation sequence from arrive probability; From the state Transfer to state probability; For the status Observed below probability; For in time In state And generate observation sequence from arrive probability; ; ;

[0133] To estimate the state Transfer to state The probability of the transition is quantified by using the adjusted forward probability, state transfer matrix and backward probability. The rationality of the transition probability and the principle that the sum of probabilities is 1 are ensured by normalizing the sum of all possible state transitions.

[0134] Update the initial state probability:

[0135]

[0136] in, The initial state is probability; for When in state The probability of the sequence starting at state probability;

[0137] By directly adopting the time The state probability , used to update the initial state probability distribution of the model, reflecting the probability that the sequence is in different states at the beginning;

[0138] Update state transition probability:

[0139]

[0140] By accumulating all time steps arrive From the state Transfer to state Probability , and then divided by the state The state transition matrix is ​​updated by summing up the probabilities of all the states at the start, which ensures the average estimation of the transition probability and better reflects the statistical law of state transitions in the entire sequence.

[0141] Update observation probabilities:

[0142]

[0143] in, For the status Observed below probability; For in time When in state The probability of Represents a time instant before the last time instant in the observation sequence.

[0144] Update observation probability The process is to obtain a more accurate generation probability of the observed symbol in a given hidden state; by re-estimating each element in the observation probability matrix , the model can more accurately reflect the The symbol is observed This means that the model has learned the true mapping relationship from hidden states to observable events, improving the simulation accuracy of the observation sequence generation process.

[0145] As the algorithm iterates, the update of observation probability enables the model to better adapt to the characteristics of the training data. A symbol is observed The frequency of is much higher than other symbols, so after updating, will increase, otherwise it will decrease, so that the model can reflect the patterns and regularities in the data.

[0146] The above steps are repeated in each iteration until the change in model parameters is less than a preset threshold, or the maximum number of iterations is reached, at which point the algorithm converges and it is considered that the optimal model parameters have been obtained.

[0147] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A user information security processing and control system applied to smart medical care, characterized in that: include: Identity authentication and authorization unit (1): used to verify the identity of the user using a two-factor authentication security mechanism and to assign access rights to the user based on an access control list; Data transmission encryption unit (2): used to encrypt the medical data transmitted by the user using the transport layer security protocol, and based on quantum communication technology, to realize quantum key distribution and provide a key exchange method for data encryption; Blockchain trust network unit (3): used to encrypt and add each piece of medical data as a transaction record to the blockchain; Log analysis management unit (4): used to record the operation logs of all users, centrally manage and analyze the log data, and use the HMM model combined with the attention mechanism to analyze user behavior patterns and automatically identify abnormal access or operation behaviors; The log analysis management unit (4) includes a log collection and processing module (41), a storage index module (42), a behavior learning module (43) and a monitoring and early warning module (44); In the log analysis management unit (4), The log collection and processing module (41) is used to automatically collect log data from various system components to ensure comprehensive coverage of user activity records, and to clean, standardize and structure the collected data; The storage index module (42) is used to efficiently store the processed log data in a database or a big data platform, and to create an index for quick query; The behavior learning module (43) is used to build a user behavior model through historical log data, improve the accuracy of anomaly detection and reduce false positives; the specific steps of the behavior learning module (43) are as follows: S431: extract useful features based on the log content, and sort the user behaviors by time to form time series data; S432: using sequence pattern mining technology to identify regular paths and time interval patterns of user behaviors, and dividing users into different behavior groups based on user behavior characteristics; S433: Count the frequency distribution of various operations and identify common and rare behaviors; S434: Based on HMM and combined with the attention mechanism, further screen and construct time window features and behavior frequency features, and use historical log data to train the model to learn the statistical characteristics and time series laws of normal user behaviors; S435: Apply the trained model to the new log data, calculate the anomaly score of the behavior, set the anomaly threshold according to the model output and actual needs, and distinguish between normal and abnormal behaviors; In S434, the calculation steps of HMM are: Forward Probability: in, For in time In state And generate observation sequence from arrive probability; is the parameter set of HMM; for observation sequence; For time; A certain status of the user; For time Status; Backward Probability: in, For in time In state And generate observation sequence from arrive The probability of represents the total length of the observation sequence; At each time step , introduce the attention mechanism to calculate the weight vector , which reflects the generation of the current observation When , the importance of each past state is taken into account. Therefore, the expressions of forward probability and backward probability are: in, The adjusted time In state And generate observation sequence from arrive probability; The adjusted time In state And generate observation sequence from arrive probability; For in time Time Status The attention weight of For in time Time Status The attention weight of From the state Transfer to state probability; For the status Observed below probability; For the status Observed below probability; The probability of state and update is calculated based on the forward probability and backward probability. The calculation steps are as follows: Status probability: in, For in time When in state probability; is the number of states; ; It's in time In state And generate observation sequence from 1 to The probability of It's in time In state And generate observation sequence from arrive probability; State transition probability: in, For at the moment arrive Between, from the state Transfer to state probability; From the state Transfer to state probability; For the status Observed below probability; For in time In state And generate observation sequence from arrive probability; For in time In state And generate observation sequence from arrive probability; From the state Transfer to state probability; For the status Observed below probability; For in time In state And generate observation sequence from arrive probability; ; ; Update the initial state probability: in, The initial state is probability; for When in state probability; Update state transition probability: Update observation probabilities: in, For the status Observed below probability; For in time When in state The probability of represents a moment before the last moment of the observation sequence; The monitoring and early warning module (44) is used to continuously monitor the log data stream, and once abnormal behavior or potential security threats are detected, the early warning system is immediately triggered and the administrator is notified.

2. According to claim 1, a user information security processing and control system applied to smart medical care is characterized by: The two-factor authentication in the identity authentication and authorization unit (1) includes password authentication and fingerprint recognition authentication; the access control list can be used by the administrator to customize access rules for user groups and specific users. When the authentication is completed, the system allocates user access rights based on the access control list.

3. A user information security processing and control system for smart medical care according to claim 2, characterized in that: The specific steps of the data transmission encryption unit (2) are as follows: S21: The data sender and the data receiver both prepare two quantum bit sequences for sending quantum states; S22: The data sender prepares quantum bits, uses quantum states to represent different bit values, and sends the prepared quantum bits to the data receiver; S23: After receiving the quantum bits sent by the data sender, the data receiver performs measurement to determine the state of each quantum bit; S24: The data sender and the data receiver retain a portion of the quantum bit information during the communication process, and compare them through a public transmission channel after the communication is completed, and determine whether there is an eavesdropper by comparing the measurement results; S25: If no eavesdropper is found during the comparison process, the data sender and the data receiver extract a shared key from the remaining matching quantum bits; S26: The data sender and the data receiver use the quantum key to encrypt the communication data; S27: The data sender and the data receiver use a portion of the extracted key for verification and perform a security analysis on the entire quantum communication process.

4. A user information security processing and control system for smart medical care according to claim 3, characterized in that: The specific steps of the blockchain trust network unit (3) are as follows: S31: Encapsulate the encrypted medical data and its metadata into a transaction; S32: Verify the legitimacy of transactions through the consensus mechanism of the blockchain network; S33: The verified transactions are packaged into a new block, which contains the hash value of the previous block, forming a chain structure; S34: The new block is broadcast to the entire network, and each participating node updates its local copy of the blockchain.

5. A user information security processing and control system for smart medical care according to claim 4, characterized in that: In S31, the transaction includes the contract code, specified data access rights, and data usage rules.

6. The user information security processing and control system applied to smart medical care according to claim 1 is characterized by: In S431, useful features include login time, operation type, operation frequency and stay time.