Nursing information management method and system
By adopting encryption policy selection model and reinforcement learning technology in the nursing information management system, and dynamically selecting encryption policies based on the network environment, the problem that encryption policies in the existing technology cannot adapt to different network environments is solved, and more efficient and secure nursing information transmission is achieved.
Patent Information
- Application Number
- CN202510177189.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-18
- Publication Date
- 2025-05-30
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art cannot adapt to the dynamic needs of different network environments and data characteristics in nursing information management, resulting in excessive or insufficient encryption strength, affecting transmission efficiency and security.
The encryption policy selection model is adopted, through reinforcement learning real-time training, the appropriate encryption policy is selected based on the current network environment data, and the nursing information is encrypted and transmitted. The model is built based on the decision tree model and the training samples are generated through the swarm optimization algorithm.
On the basis of ensuring transmission security, improve network transmission performance and resource utilization, and ensure the adaptability and accuracy of encryption policies.
Smart Images
Figure CN120074899A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of information management, and particularly to a nursing information management method and system. Background Art
[0002] In nursing information management, the secure transmission of information is of great importance. Such information includes health data, diagnosis and treatment records, nursing plans, drug usage records, etc. during the nursing process, involving personal privacy and the confidentiality of medical data.
[0003] The prior art generally uses fixed encryption algorithms and parameter settings. Regardless of how the network environment and data sensitivity change, the encryption strategy remains unchanged. This method cannot adapt to the dynamic requirements of different network environments and data characteristics, resulting in too high encryption intensity in some cases, reduced transmission efficiency, and waste of system resources; while in other cases, the encryption intensity is insufficient, posing risks of data leakage and tampering. Summary of the Invention
[0004] The present invention uses an encryption strategy selection model to quickly select a corresponding encryption strategy for the current network environment, and then encrypts and transmits the nursing information through the selected encryption strategy, which can improve the network transmission performance and the utilization rate of resources during the network transmission process on the basis of ensuring transmission security; and the encryption strategy selection model is trained in real time through reinforcement learning, which can make the encryption strategy selection model more suitable for the actual network environment.
[0005] A nursing information management method includes: Obtain nursing information; Obtain the current network environment data; Send the current network environment data to an encryption strategy selection model for processing, and output an encryption strategy; Configure a key generator through the encryption strategy, generate a key pair through the configured key generator, then encrypt the nursing information with the key pair to obtain encrypted nursing information, and finally transmit the encrypted nursing information to a cloud server; The encryption strategy selection model is established based on a decision tree model, and the encryption strategy selection positive training samples used for training the encryption strategy selection model are simulated and generated through a swarm optimization algorithm; During the management process of nursing information, it also includes training the encryption strategy selection model in real time through reinforcement learning.
[0006] As a preferred aspect of the present invention, training the encryption strategy selection model specifically includes the following steps: Obtain a number of positive training samples for encryption policy selection. The positive training samples for encryption policy selection include feature network environment data and corresponding optimal encryption policies. Combine all the positive training samples for encryption policy selection to form a positive training set for encryption policy selection, and send the positive training set for encryption policy selection into an encryption policy selection model with initialized parameters for training. Obtain the accuracy of the encryption policy selection model, and determine whether the accuracy of the encryption policy selection model is higher than a preset threshold. If the accuracy of the encryption policy selection model is higher than the preset threshold, output the trained encryption policy selection model; otherwise, continue to train the encryption policy selection model using the positive training set for encryption policy selection.
[0007] As a preferred aspect of the present invention, positive training samples for encryption policy selection are generated by simulating with a swarm optimization algorithm, which specifically includes the following steps: Sequentially select feature network environment data from the network environment database. Determine the optimal encryption policy corresponding to the selected feature network environment data through the swarm optimization algorithm, and then combine the selected feature network environment data with the corresponding optimal encryption policy to form a positive training sample for encryption policy selection; The network environment database stores a number of feature network environment data. The construction of the network environment database includes the following content: obtain a number of network environment data, and perform clustering analysis on all the obtained network environment data to output several network environment data clustering clusters F i , i = 1, 2, 3,..., I, where I is the total number of network environment data clustering clusters. Calculate the weight ratio δ i corresponding to each network environment data clustering cluster F i . The weight ratio δ i of the network environment data clustering cluster F i is the ratio of the number of network environment data in the network environment data clustering cluster F i to the total number of all network environment data. For each network environment data clustering cluster F i , perform the following operations: randomly select floor(δ i ·Q i ) network environment data from the network environment data clustering cluster F i as feature network environment data and store them in the network environment database, where floor is the rounding function and Q i is the total number of network environment data in the network environment data clustering cluster F i .
[0008] As a preferred aspect of the present invention, the optimal encryption policy corresponding to the selected feature network environment data is determined through the swarm optimization algorithm, which specifically includes the following steps: Construct a number of encrypted policy simulation individuals, and then form a simulation population set with all the encrypted policy simulation individuals; set the maximum number of iterations; Calculate the fitness corresponding to each encrypted policy simulation individual; Based on the fitness corresponding to the encrypted policy simulation individual, perform simulation iteration on the simulation population set through the genetic algorithm; When the number of iterations reaches the maximum number of iterations, output the encrypted policy simulation individual with the maximum fitness as the optimal encrypted policy corresponding to the selected feature network environment data.
[0009] As a preferred aspect of the present invention, the specific steps for calculating the fitness corresponding to the encrypted policy simulation individual are as follows: Concatenate the current feature network environment data with the encrypted policy simulation individual to construct the data to be evaluated, and then send the data to be evaluated into the encrypted policy evaluation model for processing, output the encrypted policy evaluation value, and the output encrypted policy evaluation value is used as the fitness corresponding to the encrypted policy simulation individual.
[0010] As a preferred aspect of the present invention, the encrypted policy evaluation model is established based on the BP neural network, and the encrypted policy evaluation model is trained, specifically including the following steps: Obtain a number of evaluation training samples. The evaluation training samples include network environment data, encrypted policies, and feedback values, and the corresponding feedback values are the weighted sum of the security evaluation value, performance evaluation value, and cost-benefit evaluation value; form all the evaluation training samples into an evaluation training set, and then send the evaluation training set into the encrypted policy evaluation model with initialized parameters for training. Using the feedback value as the target, calculate the loss value, and determine whether the loss value is within the preset range. If the loss value is within the preset range, output the trained encrypted policy evaluation model. Otherwise, continue to train the encrypted policy evaluation model through the evaluation training set.
[0011] As a preferred aspect of the present invention, the encrypted policy selection model is trained in real time through reinforcement learning, specifically including the following steps: Whenever the network environment data and the corresponding encrypted policy are obtained, obtain the feedback value after executing the corresponding encrypted policy; form the network environment data and the encrypted policy into a real-time training sample, and send the real-time training sample into the encrypted policy evaluation model and the target encrypted policy evaluation model respectively, and output the encrypted policy evaluation value and the target encrypted policy evaluation value; Train the encrypted policy selection model in real time based on the difference between the target encrypted policy evaluation value and the feedback value; train the encrypted policy evaluation model in real time based on the difference between the encrypted policy evaluation value and the feedback value; At intervals of a preset period, replace the original target encryption policy evaluation model with the encryption policy evaluation model as the target encryption policy evaluation model; and in the initial state, the encryption policy evaluation model and the target encryption policy evaluation model are the same.
[0012] The present invention also provides a nursing information management system, including: A nursing information acquisition module for acquiring nursing information; A network environment data acquisition module for acquiring current network environment data; An encryption policy selection module for sending the current network environment data to an encryption policy selection model for processing and outputting an encryption policy. The encryption policy selection model is established based on a decision tree model, and the positive training samples for training the encryption policy selection model are generated by simulating a swarm optimization algorithm; A nursing information encryption transmission module for configuring a key generator through an encryption policy, generating a key pair through the configured key generator, encrypting the nursing information with the key pair to obtain encrypted nursing information, and finally transmitting the encrypted nursing information to a cloud server; A real-time training module for performing real-time training on the encryption policy selection model through reinforcement learning.
[0013] The present invention has the following advantages: 1. The present invention uses an encryption policy selection model to quickly select a corresponding encryption policy for the current network environment, and then encrypts and transmits the nursing information through the selected encryption policy, which can improve the network transmission performance and the utilization rate of resources during the network transmission process while ensuring the transmission security; and the encryption policy selection model is trained in real time through reinforcement learning, which can make the encryption policy selection model more suitable for the actual network environment.
[0014] 2. The present invention performs clustering analysis on a number of network environment data, selects network environment data as characteristic network environment data from different network environment data clustering clusters according to a proportion, and further constructs a network environment database. The characteristic network environment data in the network environment database can generally reflect different network conditions. Training the encryption policy selection model with the positive training samples of the encryption policy selection composed of these characteristic network environment data and their corresponding optimal encryption policies can improve the accuracy of the encryption policy output by the encryption policy selection model.
[0015] 3. Obtain the feedback value after executing the output encryption policy in different network environments, and perform real-time training on the encryption policy selection model based on the difference between the feedback value and the target encryption policy evaluation value output by the target encryption policy evaluation model, so that the encryption policy output by the encryption policy selection model can better conform to the actual network situation; at the same time, setting the target encryption policy evaluation value can ensure the stability of training within a certain period of time. Brief Description of the Drawings
[0016] Figure 1 The figure is a schematic structural diagram of the nursing information management system adopted in the embodiment of the present invention. Detailed Embodiment
[0017] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
[0018] Embodiment 1, a nursing information management method, including: Obtain nursing information, which includes patient vital sign monitoring data and nursing guidance opinions, etc. The patient vital sign monitoring data can be uploaded and obtained through various vital sign monitoring devices, and the nursing guidance opinions can be obtained by combining conversation information, record information, etc. And these nursing information, as the privacy data of patients, need to be kept confidential to ensure the information security of patients and realize the management of nursing information; Obtain the current network environment data, which includes data such as the current network bandwidth, network latency, and device load. Through these network environment data, the current network situation can be analyzed. For different network situations, different encryption policies need to be adopted to encrypt the nursing information; Send the current network environment data to the encryption policy selection model for processing, and output the encryption policy. The encryption policy includes different encryption algorithms, key lengths, and encryption modes, etc. By selecting the corresponding encryption policy for different network environments, the performance of network transmission and the utilization rate of resources during network transmission can be improved on the basis of ensuring transmission security; Configure the key generator through the encryption policy, generate a key pair through the configured key generator, then encrypt the nursing information with the key pair to obtain the encrypted nursing information, and finally transmit the encrypted nursing information to the cloud server; The encryption policy selection model is established based on the decision tree model. The decision tree model is an algorithm based on a tree structure, including a root node, internal nodes, and leaf nodes, etc. And the positive training samples for the encryption policy selection model used for training are generated by simulating through a swarm optimization algorithm. It should be noted that the encryption policy selection model established based on the decision tree model can quickly obtain encryption policies based on regular partitioning, which is extremely important in the information management process that requires efficiency; In the process of managing nursing information, it also includes real-time training of the encryption policy selection model through reinforcement learning. Due to the uncertainty of the network environment, the initially trained encryption policy selection model may not be sufficient to handle all network situations. Therefore, reinforcement learning is used to achieve interaction with the real-time network environment, and real-time feedback is collected to optimize the encryption policy selection model, making the encryption policy selection model more suitable for the actual network environment and obtaining the encryption policy with the highest adaptability.
[0019] This application uses the encryption policy selection model to quickly select the corresponding encryption policy for the current network environment, and then encrypts and transmits the nursing information through the selected encryption policy, which can improve the performance of network transmission and the utilization rate of resources during network transmission on the basis of ensuring transmission security; and the encryption policy selection model is trained in real time through reinforcement learning, which can make the encryption policy selection model more suitable for the actual network environment.
[0020] Training the encryption policy selection model specifically includes the following steps: Obtain a number of positive training samples for encryption policy selection. The positive training samples for encryption policy selection include characteristic network environment data and the corresponding optimal encryption policy. It should be noted that the positive training samples for encryption policy selection here are obtained based on the optimal encryption policies simulated through a swarm optimization algorithm under different standard network environments. All the positive training samples for encryption policy selection are combined into a positive training set for encryption policy selection, and the positive training set for encryption policy selection is sent into the encryption policy selection model with parameter initialization for training. During the training, some features are selected from the network environment data as the root node, and then nodes are selected layer by layer based on the gain of information entropy, and pruning is used to prevent overfitting until the target condition is reached, such as the conditional entropy is less than the set threshold, the accuracy of the encryption policy selection model is obtained, and it is judged whether the accuracy of the encryption policy selection model is higher than the preset threshold. The preset threshold is set by developers according to experience. If the accuracy of the encryption policy selection model is higher than the preset threshold, the trained encryption policy selection model is output; otherwise, the encryption policy selection model is continuously trained through the positive training set for encryption policy selection.
[0021] The pruning method during the training process specifically includes the following content: For any decision node in the encryption policy selection model, perform the following operations. Obtain the corresponding two subtrees according to the decision node. It should be noted that at the decision node, there will be a branch and it will split into two child nodes. These two child nodes and their respective descendants constitute two subtrees. Determine the embedding vectors corresponding to the subtrees according to the node features in the subtrees and the features corresponding to the network environment data. Calculate the distance between the embedding vectors corresponding to the two subtrees, and denote it as the embedding distance. Determine whether the embedding distance is less than the set threshold. If the embedding distance is less than the set threshold, it means that the functions of the two subtrees are similar, and delete the subtree with fewer nodes to complete the pruning operation; otherwise, do not perform the pruning operation. By judging the differences between the subtrees, an adaptive pruning operation can be realized. It should be added that generally, there are three types of nodes in a decision tree, namely the root node, the decision node, and the leaf node. Among them, the root node represents the entire data set and serves as the starting point of the decision tree; the decision node is located between the root node and the leaf node, and each decision node represents a feature used to divide the data; the leaf node is the end of the decision tree and no longer conducts division, and is used to represent the decision result. Determine the embedding vectors corresponding to the subtrees according to the node features in the subtrees and the features corresponding to the network environment data, which specifically includes the following content: Traverse the nodes in the subtree. The traversal order is generally from left to right and from top to bottom. Store the node features corresponding to the traversed nodes in the feature analysis set in the traversal order. It should be noted that the node features here refer to the features for dividing data at the node; send the network environment data into a multi-layer perceptron for feature extraction, construct the features corresponding to the network environment data, and denote them as network environment features. The parameters of the multi-layer perceptron here are adjusted following the training of the encryption policy selection model; add the network environment features at the end of each node feature in the feature analysis set to reconstruct the node features. Through the fusion of the network environment features, the pruning operation can pay more attention to the adaptability of the network environment; then send the reconstructed feature analysis set into the RNN network for processing to output the embedding vectors corresponding to the subtrees. Simulate and generate positive training samples for encryption policy selection through a swarm optimization algorithm, which specifically includes the following steps: Sequentially select characteristic network environment data from the network environment database, determine the optimal encryption policy corresponding to the selected characteristic network environment data through the swarm optimization algorithm, and then form positive training samples for encryption policy selection by combining the selected characteristic network environment data with the corresponding optimal encryption policy. Several pieces of characteristic network environment data are stored in the network environment database. The construction of the network environment database includes the following content. Obtain several pieces of network environment data, and perform clustering analysis on all the obtained network environment data. The clustering analysis method can adopt the k-means algorithm, and output several network environment data clustering clusters F i, where \(i = 1, 2, 3, \ldots, I\) and \(I\) is the total number of network environment data clustering clusters, calculate each network environment data clustering cluster \(F\) i The corresponding weight ratio \(\delta\) i , the network environment data clustering cluster \(F\) i The weight ratio \(\delta\) i is the ratio of the number of network environment data in the network environment data clustering cluster \(F\) to the total number of all network environment data. For each network environment data clustering cluster \(F\) i , perform the following operations. Randomly select \(\lfloor\delta\cdot Q\rfloor\) network environment data from the network environment data clustering cluster \(F\) i as characteristic network environment data and store them in the network environment database, where \(\lfloor\ \rfloor\) is the floor function and \(Q\) i is the total number of network environment data in the network environment data clustering cluster \(F\). i ·Q i ), where \(\lfloor\ \rfloor\) is the floor function, and \(Q\) i is the total number of network environment data in the network environment data clustering cluster \(F\). i
[0022] In this application, by performing clustering analysis on a number of network environment data, selecting network environment data from different network environment data clustering clusters as characteristic network environment data according to a proportion, and further constructing a network environment database, the characteristic network environment data in the network environment database can generally reflect different network situations. Using the encryption policy selection positive training samples composed of these characteristic network environment data and their corresponding optimal encryption policies to train the encryption policy selection model can improve the accuracy of the encryption policy output by the encryption policy selection model.
[0023] Determine the optimal encryption policy corresponding to the selected characteristic network environment data through a swarm optimization algorithm, which specifically includes the following steps: Construct a number of encryption policy simulation individuals, and then form a simulation population set with all encryption policy simulation individuals. It should be noted that in the process of constructing encryption policy simulation individuals, discrete data such as encryption algorithms and encryption modes can be assigned values by encoding, while continuous data such as key lengths can be assigned values; set the maximum number of iterations; Calculate the fitness corresponding to each encryption policy simulation individual. The fitness of the encryption policy simulation individual can reflect whether the encryption policy corresponding to the encryption policy simulation individual is suitable for the current characteristic network environment data; The specific steps for calculating the fitness corresponding to the encryption policy simulation individual are as follows: Concatenate the current characteristic network environment data with the encryption policy simulation individual to construct the data to be evaluated, and then send the data to be evaluated into the encryption policy evaluation model for processing to output the encryption policy evaluation value. The output encryption policy evaluation value is used as the fitness corresponding to the encryption policy simulation individual; Based on the fitness of the encryption policy simulation individuals, the simulation population set is simulated and iterated through the genetic algorithm. During this process, selection, recombination, and mutation operations are performed. It should be noted that the mutation of discrete data is transformed into other encodings, while the mutation of continuous data is transformed into a random number within the range. When the number of iterations reaches the maximum number of iterations, the encryption policy simulation individual with the maximum fitness is output as the optimal encryption policy corresponding to the selected feature network environment data.
[0024] The encryption policy evaluation model is established based on the BP neural network. Training the encryption policy evaluation model specifically includes the following steps: Obtain a number of evaluation training samples. The evaluation training samples include network environment data, encryption policies, and feedback values. It should be noted that the evaluation training samples are obtained during the actual data transmission process, and the corresponding feedback value is the weighted sum of the security evaluation value, performance evaluation value, and cost-benefit evaluation value. Among them, the security evaluation value is the evaluation of the encryption policy's resistance to various attacks, which can be analyzed by methods such as the attack complexity analysis of cryptography. The performance evaluation value is the performance performance after adopting the encryption policy, such as the comprehensive analysis value of decryption speed, data transmission rate, and network latency. The cost-benefit evaluation value is the evaluation of the resource consumption for implementing the encryption policy. All evaluation training samples are combined into an evaluation training set, and then the evaluation training set is sent into the encryption policy evaluation model with initialized parameters for training. Using the feedback value as the target, calculate the loss value and determine whether the loss value is within the preset range. The preset range is determined by the developer. If the loss value is within the preset range, output the trained encryption policy evaluation model; otherwise, continue to train the encryption policy evaluation model through the evaluation training set.
[0025] The encryption policy selection model is trained in real time through reinforcement learning, specifically including the following steps: Whenever the network environment data and the corresponding encryption policy are obtained, obtain the feedback value after executing the corresponding encryption policy; combine the network environment data and the encryption policy into a real-time training sample, and send the real-time training sample into the encryption policy evaluation model and the target encryption policy evaluation model respectively, and output the encryption policy evaluation value and the target encryption policy evaluation value; Based on the difference between the target encryption policy evaluation value and the feedback value, train the encryption policy selection model in real time, specifically by adjusting the parameters in the encryption policy selection model to make the difference between the target encryption policy evaluation value and the feedback value smaller; based on the difference between the encryption policy evaluation value and the feedback value, train the encryption policy evaluation model in real time; At intervals of a preset period, replace the original target encryption policy evaluation model with the encryption policy evaluation model as the target encryption policy evaluation model; and in the initial state, the encryption policy evaluation model and the target encryption policy evaluation model are the same.
[0026] This application obtains the feedback value after executing the output encryption policy in different network environments, and trains the encryption policy selection model in real time based on the difference between the feedback value and the target encryption policy evaluation value output by the target encryption policy evaluation model, so that the encryption policy output by the encryption policy selection model can better conform to the actual network situation; at the same time, setting the target encryption policy evaluation value can ensure the stability of training within a certain period of time.
[0027] Embodiment 2, a nursing information management system, as Figure 1 shown, includes: A nursing information acquisition module, used to acquire nursing information, which includes patient vital sign monitoring data and nursing guidance opinions, etc. The patient vital sign monitoring data can be uploaded and acquired through various vital sign monitoring devices, and the nursing guidance opinions can be acquired by combining conversation information, record information, etc. And these nursing information, as the privacy data of patients, need to be kept confidential to ensure the information security of patients and realize the management of nursing information; A network environment data acquisition module, used to acquire the current network environment data, which includes data such as the current network bandwidth, network latency, and device load. Through these network environment data, the current network situation can be analyzed. For different network situations, different encryption policies need to be adopted to encrypt the nursing information; An encryption policy selection module, used to send the current network environment data into the encryption policy selection model for processing and output an encryption policy. The encryption policy includes different encryption algorithms, key lengths, and encryption modes, etc. By selecting the corresponding encryption policy for different network environments, the performance of network transmission and the utilization rate of resources during network transmission can be improved on the basis of ensuring transmission security. The encryption policy selection model is established based on the decision tree model. The decision tree model is an algorithm based on a tree structure, including a root node, internal nodes, and leaf nodes, etc. And the encryption policy selection positive training samples used to train the encryption policy selection model are generated by simulating the swarm optimization algorithm. It should be noted that the encryption policy selection model established based on the decision tree model can quickly obtain the encryption policy based on the regularized division, which is extremely important in the information management process that requires efficiency; A nursing information encrypted transmission module, used to configure the key generator through the encryption policy, generate a key pair through the configured key generator, then encrypt the nursing information with the key pair to obtain the encrypted nursing information, and finally transmit the encrypted nursing information to the cloud server; A real-time training module for real-time training of an encryption policy selection model through reinforcement learning. Due to the uncertainty of the network environment, the initially trained encryption policy selection model may not be sufficient to handle all network situations. Therefore, reinforcement learning is used to interact with the real-time network environment and collect real-time feedback to optimize the encryption policy selection model, making the encryption policy selection model more suitable for the actual network environment and obtaining the encryption policy with the highest adaptability.
[0028] It should be understood that those of ordinary skill in the art can make improvements or transformations based on the above description, and all such improvements and transformations shall fall within the protection scope of the appended claims of the present invention. The parts not described in detail in this specification belong to the prior art well-known to those skilled in the art.
Claims
1. A nursing information management method, characterized in that: include: Get care information; Get the current network environment data; Send the current network environment data to the encryption strategy selection model for processing and output the encryption strategy; The key generator is configured through the encryption strategy, and a key pair is generated through the configured key generator, and then the nursing information is encrypted through the key pair to obtain the encrypted nursing information, and finally the encrypted nursing information is transmitted to the cloud server; The encryption strategy selection model is established based on a decision tree model, and the encryption strategy selection positive training samples used to train the encryption strategy selection model are simulated and generated through a swarm optimization algorithm; The management process of nursing information also includes real-time training of the encryption strategy selection model through reinforcement learning.
2. A nursing information management method according to claim 1, characterized in that: The encryption strategy selection model is trained, which includes the following steps: Obtain several encryption strategy selection forward training samples, the encryption strategy selection forward training samples include characteristic network environment data and the corresponding optimal encryption strategy, all encryption strategy selection forward training samples form an encryption strategy selection forward training set, and send the encryption strategy selection forward training set to the encryption strategy selection model with parameter initialization for training, obtain the accuracy of the encryption strategy selection model, and determine whether the accuracy of the encryption strategy selection model is higher than a preset threshold. If the accuracy of the encryption strategy selection model is higher than the preset threshold, output the trained encryption strategy selection model, otherwise, continue to train the encryption strategy selection model through the encryption strategy selection forward training set.
3. A nursing information management method according to claim 2, characterized in that: The encryption strategy is simulated and generated through the swarm optimization algorithm to select positive training samples, which specifically includes the following steps: Select characteristic network environment data from the network environment database in sequence, determine the optimal encryption strategy corresponding to the selected characteristic network environment data through a group optimization algorithm, and then combine the selected characteristic network environment data and the corresponding optimal encryption strategy to form an encryption strategy selection positive training sample; The network environment database stores several sets of characteristic network environment data. The construction of the network environment database includes the following contents: obtaining several sets of network environment data, performing cluster analysis on all the obtained network environment data, and outputting several network environment data clusters F i , i=1,2,3,…,I, I is the total number of network environment data clusters, calculate each network environment data cluster F i The corresponding weight ratio δ i , network environment data cluster F i Weight ratio δ i Cluster F is the network environment data cluster i The ratio of the number of network environment data in to the number of all network environment data, for each network environment data cluster F i , perform the following operations to cluster cluster F from network environment data i Randomly select floor (δ i Q i ) network environment data are stored in the network environment database as characteristic network environment data, where floor is the rounding function, Q i Cluster F is the network environment data cluster i The total amount of network environment data in .
4. A nursing information management method according to claim 3, characterized in that: The optimal encryption strategy corresponding to the selected characteristic network environment data is determined by a group optimization algorithm, which specifically includes the following steps: Construct several encryption strategy simulation individuals, and then form a simulation population set from all encryption strategy simulation individuals; set the maximum number of iterations; For each encryption strategy simulation individual, calculate the fitness of the encryption strategy simulation individual; Based on the fitness of the individuals simulated by the encryption strategy, the simulated population set is simulated and iterated through the genetic algorithm; When the number of iterations reaches the maximum number of iterations, the encryption strategy with the largest fitness is output to simulate the optimal encryption strategy corresponding to the selected characteristic network environment data.
5. A nursing information management method according to claim 4, characterized in that: The specific steps for calculating the fitness of the encrypted strategy simulation individuals are as follows: The current characteristic network environment data is spliced with the encryption strategy simulation individual to construct the data to be evaluated, and then the data to be evaluated is sent to the encryption strategy evaluation model for processing, and the encryption strategy evaluation value is output. The output encryption strategy evaluation value is used as the fitness corresponding to the encryption strategy simulation individual.
6. A nursing information management method according to claim 5, characterized in that: The encryption strategy evaluation model is established based on the BP neural network. The encryption strategy evaluation model is trained, which specifically includes the following steps: Obtain a number of evaluation training samples, the evaluation training samples include network environment data, encryption strategy and feedback value, and the corresponding feedback value is the weighted sum of security evaluation value, performance evaluation value and cost-effectiveness evaluation value; all evaluation training samples are combined into an evaluation training set, and then the evaluation training set is sent to the encryption strategy evaluation model with parameter initialization for training, with the feedback value as the target, the loss value is calculated, and it is judged whether the loss value is within the preset range. If the loss value is within the preset range, the trained encryption strategy evaluation model is output, otherwise, the encryption strategy evaluation model is continuously trained through the evaluation training set.
7. A nursing information management method according to claim 6, characterized in that: The encryption strategy selection model is trained in real time through reinforcement learning, which includes the following steps: Whenever network environment data and corresponding encryption strategies are obtained, feedback values after the corresponding encryption strategies are executed are obtained; network environment data and encryption strategies are combined into real-time training samples, and the real-time training samples are respectively sent to the encryption strategy evaluation model and the target encryption strategy evaluation model, and the encryption strategy evaluation value and the target encryption strategy evaluation value are output; Based on the difference between the target encryption strategy evaluation value and the feedback value, the encryption strategy selection model is trained in real time; Based on the difference between the encryption strategy evaluation value and the feedback value, the encryption strategy evaluation model is trained in real time; At preset intervals, the encryption policy evaluation model is used as the target encryption policy evaluation model to replace the original target encryption policy evaluation model; and in the initial state, the encryption policy evaluation model and the target encryption policy evaluation model are consistent.
8. A nursing information management system, characterized in that: The system applies a nursing information management method according to any one of claims 1 to 7, including: A nursing information acquisition module is used to acquire nursing information; A network environment data acquisition module is used to obtain current network environment data; An encryption strategy selection module is used to input the current network environment data into the encryption strategy selection model for processing and output the encryption strategy. The encryption strategy selection model is established based on the decision tree model, and the encryption strategy selection positive training samples used to train the encryption strategy selection model are simulated and generated by a swarm optimization algorithm. The nursing information encryption transmission module is used to configure the key generator through the encryption strategy, generate a key pair through the configured key generator, encrypt the nursing information through the key pair, obtain the encrypted nursing information, and finally transmit the encrypted nursing information to the cloud server; A real-time training module for real-time training of the encryption strategy selection model through reinforcement learning.