Family doctor management system and method for diabetic patients based on rehabilitation model

Abstract

The present invention relates to the field of medical data management, in particular to a family doctor management system and method for diabetic patients based on a rehabilitation model; the family doctor management system for diabetic patients based on a rehabilitation model includes: a data acquisition module for acquiring first health data of a target patient; a data encryption module connected to the data acquisition module by signal, for encrypting the first health data to generate second health data; a family doctor module connected to the data encryption module by signal, for generating a rehabilitation plan for the target patient based on the second health data using a rehabilitation plan generation model; a visualization module connected to the family doctor module by signal, for visualizing the rehabilitation plan. The present invention solves the problems of data islands, privacy leakage and insufficient data sharing by acquiring the patient's health data and by data encryption and model optimization, which helps to ensure the security and privacy of patient data and improve the effect and efficiency of rehabilitation management.

Description

Technical Field

[0001] The present invention relates to the field of medical data management, and in particular to a family doctor management system and method for diabetic patients based on a rehabilitation model. Background Art

[0002] As the number of diabetes patients continues to grow worldwide, the demand for diabetes rehabilitation management is increasing. The core of the rehabilitation model is to help patients stabilize blood sugar levels, reduce complications and improve quality of life through personalized management and continuous monitoring. However, the diabetes rehabilitation process is complex and involves multi-dimensional health data, such as blood sugar fluctuations, eating habits, exercise volume and mental health. In order to achieve accurate personalized rehabilitation guidance, a large amount of health data must be obtained and analyzed, which makes patient privacy protection and data security a core issue.

[0003] The current diabetes management system has the risk of data silos and privacy leakage. Patients' health data are often scattered in different medical institutions, monitoring equipment and applications, lacking unified management and analysis. Even if there is a need for data sharing, the traditional centralized data transmission mode is prone to security issues such as leakage and tampering of medical data during transmission and storage. At the same time, patients' concerns about the privacy of their personal health data make them unwilling to fully open their data, further reducing the effectiveness and accuracy of personalized rehabilitation guidance. Therefore, the diabetes management system under the rehabilitation mode urgently needs a technical mechanism that can both protect data privacy and effectively process and share health data, so as to achieve efficient collaboration between patients and medical teams. Summary of the invention

[0004] In view of the deficiencies of existing technologies and actual application needs, the present invention provides a family doctor management system and method for diabetic patients based on a rehabilitation model, which aims to collect patients' health data and use data encryption and model optimization technology to solve the problems of data silos, privacy leakage and insufficient data sharing, ensure the security and privacy of patient data, and improve the effectiveness and efficiency of rehabilitation management.

[0005] In the first aspect, the family doctor management system for diabetic patients based on the rehabilitation model provided by the present invention comprises:

[0006] A data collection module, used for collecting first health data of a target patient;

[0007] A data encryption module connected to the data acquisition module by signal, used for encrypting the first health data to generate second health data;

[0008] A family doctor module connected to the data encryption module signal, used to generate a rehabilitation plan for the target patient based on the second health data and using a rehabilitation plan generation model;

[0009] A visualization module connected to the family doctor module signal is used to visualize the rehabilitation plan.

[0010] Furthermore, in the family doctor management system for diabetic patients based on the rehabilitation model provided by the first aspect of the present invention,

[0011] The data acquisition module includes a plurality of data acquisition submodules, each of which corresponds to a plurality of target patients, and any data acquisition submodule is used to acquire the first health data of the target patient corresponding thereto;

[0012] The family doctor module includes multiple family doctor sub-modules, and the multiple family doctor sub-modules correspond one-to-one to the multiple data collection sub-modules. Any family doctor sub-module is used to generate a rehabilitation plan for the corresponding target patient based on the second health data encrypted from the first health data collected by the corresponding data collection sub-module, using a rehabilitation plan generation model.

[0013] Furthermore, in the family doctor management system for diabetic patients based on the rehabilitation mode provided in the first aspect of the present invention, the data encryption module includes:

[0014] A key generation submodule, wherein the key generation submodule is respectively connected to the plurality of data acquisition submodules by signal, and is respectively connected to the plurality of family doctor submodules by signal, and the key generation submodule is used to generate a key pair based on the corresponding relationship between the data acquisition submodule and the family doctor submodule, and any key pair includes a public key and a private key;

[0015] Multiple data encryption submodules, the multiple data encryption submodules are signal-connected with the multiple data acquisition submodules in a one-to-one correspondence, the multiple data encryption submodules are signal-connected with the multiple family doctor submodules in a one-to-one correspondence based on the correspondence between the data acquisition submodules and the family doctor submodules, the multiple data encryption submodules are also signal-connected with the key generation submodule, any data encryption submodule is used to encrypt the first health data collected by the corresponding data acquisition submodule based on the corresponding public key to generate the second health data.

[0016] Furthermore, in the family doctor management system for diabetic patients based on the rehabilitation mode provided by the first aspect of the present invention, any key pair is generated by the following steps:

[0017] The key generation submodule obtains the data collection objects of the multiple data collection submodules respectively, and obtains the rehabilitation management objects of the multiple family doctor submodules respectively;

[0018] The key generation submodule obtains the correspondence between the data acquisition submodule and the family doctor submodule based on the data acquisition object of the data acquisition submodule and the rehabilitation management object of the family doctor submodule, and generates a set of key pairs based on each set of correspondence using the Paillier encryption algorithm. Any set of key pairs includes a public key and a private key.

[0019] Furthermore, the family doctor management system for diabetic patients based on the rehabilitation mode provided in the first aspect of the present invention further includes:

[0020] A plurality of model training modules, each of which is used to train and verify the rehabilitation program generation model based on a number of historical samples from a corresponding diabetic patient, and obtain model parameters of the verified rehabilitation program generation model;

[0021] A parameter optimization module, wherein the parameter optimization module is respectively connected to a plurality of model training modules and respectively connected to each family doctor sub-module signal, and is used for optimizing the model parameters of the rehabilitation program generation model in each family doctor sub-module connected thereto based on the model parameters and the number of historical samples of the plurality of model training modules.

[0022] Furthermore, in the family doctor management system for diabetic patients based on the rehabilitation model provided in the first aspect of the present invention, any model training module includes:

[0023] A historical sample submodule, wherein the historical sample submodule stores a plurality of historical samples from the same diabetic patient, any of the historical samples being a rehabilitation program recommended for the diabetic patient under a set of health data;

[0024] A historical sample encryption submodule, which is signal-connected to the historical sample submodule and to the model training submodule, and is used to generate a key pair based on a homomorphic encryption algorithm, encrypt the historical sample using the public key in the key pair, and distribute the private key in the key pair to the corresponding model training submodule;

[0025] A model training submodule is signal-connected to the parameter optimization module, and is used to decrypt the encrypted historical samples based on the private key from the corresponding historical sample encryption submodule, and use the decrypted historical samples to train and verify the rehabilitation program generation model, and obtain the model parameters of the verified rehabilitation program generation model.

[0026] Furthermore, in the family doctor management system for diabetic patients based on the rehabilitation model provided by the first aspect of the present invention, the model parameters are optimized by the following steps:

[0027] The parameter optimization module obtains the number of historical samples used to train the rehabilitation program generation model in each model training submodule, and obtains the weighted weight of the corresponding model training submodule based on the number of historical samples of each model training submodule;

[0028] The model parameters used to train the rehabilitation program generation model in each model training submodule are obtained respectively, and the optimized model parameters are obtained through the weighted average calculation model in combination with the weighted weights of each model training submodule.

[0029] Furthermore, in the family doctor management system for diabetic patients based on the rehabilitation model provided in the first aspect of the present invention, the rehabilitation program generation models in any two family doctor sub-modules are the same.

[0030] In a second aspect, based on the family doctor management system for diabetic patients based on the rehabilitation model provided in the first aspect, the family doctor management method for diabetic patients based on the rehabilitation model provided by the present invention comprises the following steps:

[0031] The data acquisition module acquires first health data of the target patient;

[0032] The data encryption module encrypts the first health data to generate second health data;

[0033] The family doctor module generates a rehabilitation plan for the target patient based on the second health data and using a rehabilitation plan generation model;

[0034] The visualization module visualizes the rehabilitation program.

[0035] In a third aspect, based on the family doctor management system for diabetic patients based on the rehabilitation model provided in the first aspect, the family doctor management method for diabetic patients based on the rehabilitation model provided by the present invention further includes the following steps:

[0036] The model training module trains and verifies the rehabilitation program generation model based on a number of historical samples corresponding to the diabetic patients, and obtains model parameters of the verified rehabilitation program generation model;

[0037] The parameter optimization module optimizes the model parameters of the rehabilitation program generation model in each family doctor sub-module connected to it based on the model parameters and historical sample quantities of multiple model training modules.

[0038] The benefits of the family doctor management system and method for diabetic patients based on the rehabilitation model provided by the present invention are:

[0039] The patient's real-time health data is obtained through the data collection module, and the data is encrypted using the homomorphic encryption algorithm, which effectively protects the privacy and security of the data and prevents data leakage and tampering.

[0040] Through the one-to-one correspondence between each data collection submodule and the family doctor submodule, a personalized rehabilitation plan generation model is generated, combined with encrypted health data, to develop a unique rehabilitation plan for each patient, achieving personalized and precise rehabilitation intervention;

[0041] The rehabilitation program generation model is continuously trained and verified through historical samples of diabetic patients. After the optimized model parameters are generated by weighted average through the parameter optimization module, the rehabilitation program generation model in each family doctor sub-module is optimized respectively, which helps to generate the rehabilitation program with accuracy and personalization. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A first schematic diagram of a family doctor management system for diabetic patients based on a rehabilitation model provided by an embodiment of the present invention;

[0043] Figure 2 A second schematic diagram of a family doctor management system for diabetic patients based on a rehabilitation mode provided by an embodiment of the present invention;

[0044] Figure 3 A third schematic diagram of a family doctor management system for diabetic patients based on a rehabilitation mode provided by an embodiment of the present invention;

[0045] Figure 4 A fourth schematic diagram of a family doctor management system for diabetic patients based on a rehabilitation mode provided by an embodiment of the present invention;

[0046] Figure 5 A fifth schematic diagram of a family doctor management system for diabetic patients based on a rehabilitation mode provided by an embodiment of the present invention;

[0047] Figure 6 A schematic flow chart of a method for family doctor management of diabetic patients based on a rehabilitation model provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described here are only for illustration and are not intended to limit the present invention. In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention.

[0049] However, it will be apparent to one of ordinary skill in the art that these specific details need not be employed to practice the present invention. In other instances, well-known circuits, software, or methods are not described in detail to avoid obscuring the present invention.

[0050] In one embodiment, see Figure 1 , Figure 1This is a first schematic diagram of a family doctor management system for diabetic patients based on a rehabilitation model provided by an embodiment of the present invention.

[0051] All user-related data involved in the technical solution of the present invention is implemented with the user's consent and authorization.

[0052] like Figure 1 As shown, the family doctor management system for diabetic patients based on the rehabilitation mode provided in an embodiment of the present invention includes a data acquisition module, a data encryption module, a family doctor module and a visualization module; further, the data acquisition module is signal-connected to the data encryption module, and the family doctor module is connected to the visualization module.

[0053] In this embodiment, the data acquisition module is used to collect the first health data of the target patient; the data encryption module is used to encrypt the first health data to generate second health data; the family doctor module is used to generate a rehabilitation plan for the target patient based on the second health data and using a rehabilitation plan generation model; the visualization module is used to visualize the rehabilitation plan.

[0054] It should be noted that the target patients described in the embodiment of the present invention refer to diabetic patients who use the system for personalized rehabilitation management, such as type 1 diabetes patients, type 2 diabetes patients, and prediabetes patients.

[0055] It can be understood that the first health data described in the present invention refers to the original health data collected from the target patient by the data acquisition module, such as physiological indicator data, exercise data, diet data, sleep data, drug usage data, etc.; the second health data described in the present invention refers to the first health data processed by the encryption algorithm in the data encryption module (such as homomorphic, RSA, AES and other encryption algorithms).

[0056] Specifically, the data acquisition module is the core part for acquiring the first health data of the target patient, and the acquisition of the first health data can be achieved by a variety of technical components, such as wearable devices, smart home medical devices, mobile applications, telemedicine platforms, etc.

[0057] Similarly, the visualization module is a key part for presenting the generated personalized rehabilitation plan to patients and doctors in an intuitive manner, which can also be implemented through a variety of visualization methods, such as Web-based or mobile-based dashboards, paper or electronic health reports, virtual health assistants, etc.

[0058] For further information, see Figure 2 , Figure 2 This is a second schematic diagram of the family doctor management system for diabetic patients based on the rehabilitation mode provided by an embodiment of the present invention.

[0059] like Figure 2 As shown, in order to achieve personalized management and efficient data processing for different target patients, the data acquisition module provided in this embodiment includes multiple data acquisition sub-modules, and the family doctor module provided by the present invention also includes multiple family doctor sub-modules.

[0060] Furthermore, for multiple target patients, the multiple data acquisition sub-modules correspond one-to-one to the multiple target patients, and any data acquisition sub-module is used to collect the first health data of the target patient corresponding to it; the multiple family doctor sub-modules correspond one-to-one to the multiple data acquisition sub-modules, and any family doctor sub-module is used to generate a rehabilitation plan for the corresponding target patient through a rehabilitation plan generation model based on the second health data corresponding to the first health data collected by the corresponding data acquisition sub-module.

[0061] For further information, see Figure 3 , Figure 3 This is a third schematic diagram of the family doctor management system for diabetic patients based on the rehabilitation mode provided by an embodiment of the present invention.

[0062] like Figure 3 As shown, in order to further ensure the privacy of health data during transmission and use, the data encryption module described in this embodiment includes a key generation submodule and multiple data encryption submodules.

[0063] Furthermore, the key generation submodule is respectively connected to the multiple data acquisition submodule signals and respectively connected to the multiple family doctor submodule signals, and the key generation submodule is used to generate a key pair based on the corresponding relationship between the data acquisition submodule and the family doctor submodule.

[0064] Furthermore, the multiple data encryption sub-modules are signal-connected with the multiple data acquisition sub-modules in a one-to-one correspondence, and the multiple data encryption sub-modules are signal-connected with the multiple family doctor sub-modules in a one-to-one correspondence based on the corresponding relationship between the data acquisition sub-modules and the family doctor sub-modules. The multiple data encryption sub-modules are also signal-connected with the key generation sub-module.

[0065] In this embodiment, the data encryption module encrypts the first health data by a homomorphic encryption algorithm. Specifically, the key generation submodule generates a key pair based on the corresponding relationship between the data acquisition submodule and the family doctor submodule, including the following steps:

[0066] S011. Acquire data acquisition objects of multiple data acquisition sub-modules respectively.

[0067] In this embodiment, a data acquisition object of a data acquisition submodule , where i is a positive integer and represents the number of the data acquisition submodule. Represents the identity information of the target patient corresponding to the i-th data acquisition submodule in the data acquisition module.

[0068] In this embodiment, the identity information is the ID number of the target patient. It is understandable that in one or some other embodiments, the identity information can also be used to distinguish the identity information of different target patients in the system, such as medical card number, social security number, patient registration ID, patient email address, patient mobile phone number, biometric identification information, etc.

[0069] S012. Obtain rehabilitation management objects of multiple family doctor sub-modules respectively.

[0070] In this embodiment, a rehabilitation management object of a family doctor submodule , where j is a positive integer and represents the number of the family doctor submodule. Represents the identity information of the target patient corresponding to the j-th family doctor sub-module in the family doctor module.

[0071] It should be noted that under the same management system, the data collection module and the family doctor module use the same type of identity information. For example, if the data collection module uses the ID number to distinguish different target patients, the family doctor module also uses the ID number to distinguish different target patients.

[0072] S013. Based on the data collection object of the data collection submodule and the rehabilitation management object of the family doctor submodule, obtain the corresponding relationship between the data collection submodule and the family doctor submodule:

[0073] If the data collection object of a data collection submodule and the rehabilitation management object of a family doctor submodule are the same target patient, then there is a corresponding relationship between this data collection object and this family doctor submodule; otherwise, there is no corresponding relationship.

[0074] In this embodiment, step S013 is based on the data collection object of each data collection submodule. , one by one with the rehabilitation management objects of the multiple family doctor sub-modules to make a match.

[0075] Furthermore, the corresponding relationship between the i-th data collection submodule and the j-th family doctor submodule satisfies the following matching model: ,in, Indicates that there is a corresponding relationship between the i-th data collection submodule and the j-th family doctor submodule. It means that there is no corresponding relationship between the i-th data collection submodule and the j-th family doctor submodule.

[0076] S014. Based on each set of corresponding relationships, generate a set of key pairs using the Paillier encryption algorithm. Each set of key pairs includes a public key and a private key.

[0077] In this embodiment, based on a set of corresponding relationships , whose public key satisfies the following generation model: , ,in, is the public key modulus, g is the public key base, Representation based on correspondence The generated public key, Representation based on correspondence The first randomly generated prime number , Representation based on correspondence Randomly generated second prime number .

[0078] Furthermore, based on a set of corresponding relationships ( ), whose private key satisfies the following generation model: ,in, is the Carmichael function value used for decryption, is the inverse function value used for decryption, Representation based on correspondence The generated private key, express and The least common multiple of Representation based on correspondence The generated public key modulus, Representation based on correspondence The generated public key base. mod represents the remainder operation.

[0079] S015. Distribute the public key in the key pair to the corresponding data encryption sub-modules, and distribute the private key in the key pair to the corresponding family doctor sub-modules.

[0080] Furthermore, based on a set of corresponding relationships Key pair , the public key Distribute to the i-th data encryption submodule in the corresponding data encryption module, and pass the private key Distributed to the j-th family doctor sub-module in the corresponding family doctor module.

[0081] It can be understood that each data encryption sub-module encrypts the first health data collected by the corresponding data acquisition sub-module based on the corresponding public key to obtain the corresponding second health data; each family doctor sub-module decrypts the second health data based on the corresponding private key, and then generates a rehabilitation plan for the target patient through the rehabilitation plan generation model.

[0082] Furthermore, the rehabilitation program generation model described in the present invention is an intelligent algorithm model for generating personalized rehabilitation programs for target patients; it can be built based on one or more existing model frameworks such as a machine learning framework, a deep learning framework, a reinforcement learning framework, a natural language processing framework, and an intelligent time series analysis framework, and focuses on providing personalized diet, exercise, drug adjustment and lifestyle recommendations for target patients based on their real-time health data.

[0083] In this embodiment, the rehabilitation program generation model provided by the present invention is a multi-task learning model, which includes multiple learning sub-models, and any one of the learning sub-models is used to generate a rehabilitation suggestion under an execution dimension based on the decrypted second health data.

[0084] It can be understood that the execution dimension refers to the classification standard of rehabilitation programs, which is used to define different types of rehabilitation interventions and guidance content; each execution dimension represents a direction or category of rehabilitation activities.

[0085] Furthermore, any learning sub-model satisfies the following expression: , where k is a positive integer and represents the number of the learning sub-model. represents the output dataset of the kth learning sub-model, represents the input dataset of the kth learning sub-model, , represents the health data set after the second health data is decrypted, represents the model parameter set of the kth learning sub-model, Represents model parameters based on The k-th learning sub-model of .

[0086] It can be understood that each learning sub-model outputs personalized rehabilitation suggestions based on its specific input data set and corresponding model parameters, guiding patients to make optimal health decisions in terms of exercise, diet, medication use, and daily routine adjustment.

[0087] It should be noted that the rehabilitation program generation models in any two family doctor sub-modules in the embodiment of the present invention are the same, that is, the model architecture and model parameters are the same.

[0088] For further information, see Figure 4 , Figure 4This is a fourth schematic diagram of the family doctor management system for diabetic patients based on the rehabilitation mode provided by an embodiment of the present invention.

[0089] In order to share the historical data of different patients, the optimization of the rehabilitation program generation model in different family doctor sub-modules is realized synchronously, such as Figure 4 As shown, the family doctor management system for diabetic patients based on the rehabilitation model provided by the embodiment of the present invention also includes multiple model training modules and a parameter optimization module.

[0090] In this embodiment, any model training module is used to train and verify the rehabilitation program generation model based on a number of historical samples from a diabetic patient, and obtain model parameters of the verified rehabilitation program generation model.

[0091] Furthermore, the parameter optimization module is respectively connected to multiple model training modules and respectively connected to each family doctor sub-module signal, and is used to optimize the model parameters of the rehabilitation program generation model in each family doctor sub-module connected thereto based on the model parameters and historical sample numbers of the multiple model training modules.

[0092] It should be noted that the model architecture of the rehabilitation program generation model in any two model training modules is the same, and the model parameters may be the same or different.

[0093] For further information, see Figure 5 , Figure 5 This is a fifth schematic diagram of a family doctor management system for diabetic patients based on a rehabilitation model provided by an embodiment of the present invention. Figure 5 As shown, any model training module includes a historical sample sub-module, a historical sample encryption sub-module and a model training sub-module; the historical sample sub-module is signal-connected to the historical sample encryption sub-module, the historical sample encryption sub-module is signal-connected to the model training sub-module, and the model training sub-module is signal-connected to the parameter optimization module.

[0094] The historical sample submodule stores a number of historical samples from the same diabetic patient, and any of the historical samples is a rehabilitation program recommended for the diabetic patient under a set of health data.

[0095] Furthermore, the historical sample encryption submodule is used to generate a key pair based on a homomorphic encryption algorithm, and use the public key in the key pair to encrypt the historical sample, and distribute the private key in the key pair to the corresponding model training submodule. The generation method of the key pair can be referred to the above embodiment, which will not be repeated here.

[0096] Furthermore, the model training submodule is used to decrypt the encrypted historical samples based on the private key from the corresponding historical sample encryption submodule, and use the decrypted historical samples to train and verify the rehabilitation program generation model, and obtain the model parameters of the verified rehabilitation program generation model.

[0097] In this embodiment, the parameter optimization module optimizes the model parameters of the rehabilitation program generation model in each family doctor module based on a weighted average method. Specifically, the parameter optimization module optimizes the model parameters of the rehabilitation program generation model in each family doctor submodule connected thereto based on the model parameters and the number of historical samples of multiple model training modules, and includes the following steps:

[0098] S021. Obtain the number of historical samples used to train the rehabilitation program generation model in each model training submodule respectively, and obtain the weighted weight of the corresponding model training submodule based on the number of historical samples of each model training submodule.

[0099] Specifically, the weighted weight of the rth model training submodule in the model training submodule satisfies the following calculation model: , where r is a positive integer and represents the number of the model training submodule. is the weighted weight of the rth model training submodule in the model training submodule, represents the number of historical samples used for the rehabilitation scheme generation model in the rth model training submodule in the model training submodule, The total number of submodules trained for the model.

[0100] S022. Obtain the model parameters used for training the rehabilitation program generation model in each model training submodule respectively, and obtain the optimized model parameters through the weighted average calculation model in combination with the weighted weights of each model training submodule.

[0101] Specifically, the model parameter set Medium model parameters , satisfying the following optimization model: , where s is a positive integer and represents the number of the model parameter. represents the sth model parameter in the model parameter set, Represents the sth model parameter in the model parameter set of the rth model training submodule.

[0102] S023. Transmit the optimized model parameters to each family doctor sub-module connected to the parameter optimization module.

[0103] Furthermore, each family doctor submodule updates the corresponding rehabilitation program generation model based on the optimized model parameters.

[0104] It should be noted that after the update, the rehabilitation program generation models in any two family doctor sub-modules are also the same, that is, the model architecture and model parameters are the same.

[0105] In one embodiment, see Figure 6 , Figure 6 This is a flow chart of the family doctor management method for diabetic patients based on the rehabilitation model provided by an embodiment of the present invention. The family doctor management method for diabetic patients based on the rehabilitation model provided by this embodiment is applicable to the family doctor management system for diabetic patients based on the rehabilitation model provided by the above embodiments.

[0106] like Figure 6 As shown, based on the data acquisition module, data encryption module, family doctor module and visualization module in the family doctor management system for diabetic patients based on the rehabilitation mode provided in the above embodiment, in order to realize the diabetes rehabilitation monitoring and guidance of a target patient, the method includes the following steps:

[0107] S01. The data acquisition module acquires first health data of a target patient.

[0108] S02. The data encryption module encrypts the first health data to generate second health data.

[0109] S03. The family doctor module generates a rehabilitation plan for the target patient based on the second health data and using a rehabilitation plan generation model.

[0110] S04. The visualization module visualizes the rehabilitation plan.

[0111] Furthermore, in order to further optimize the model parameters of the rehabilitation program generation model in the family doctor module to obtain more accurate and personalized rehabilitation program recommendations, based on the multiple model training modules and parameter optimization modules in the family doctor management system for diabetic patients based on the rehabilitation mode provided in the above embodiment, the method also includes:

[0112] S001. The model training module trains and verifies a rehabilitation program generation model based on a number of historical samples corresponding to diabetic patients, and obtains model parameters of the verified rehabilitation program generation model.

[0113] It is understandable that the sources of historical samples for different model training modules are different.

[0114] S002. The parameter optimization module optimizes the model parameters of the rehabilitation program generation model in each family doctor sub-module connected to it based on the model parameters and historical sample quantities of multiple model training modules.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents.

Claims

1. A family doctor management system for diabetic patients based on rehabilitation mode, characterized by: include: A data collection module, used for collecting first health data of a target patient; A data encryption module connected to the data acquisition module by signal, used for encrypting the first health data to generate second health data; A family doctor module connected to the data encryption module signal, used to generate a rehabilitation plan for the target patient based on the second health data and using a rehabilitation plan generation model; A visualization module connected to the family doctor module signal, used to visualize the rehabilitation program; The data acquisition module includes a plurality of data acquisition submodules, each of which corresponds to a plurality of target patients, and any data acquisition submodule is used to acquire the first health data of the target patient corresponding thereto; The family doctor module includes a plurality of family doctor submodules, and the plurality of family doctor submodules correspond to the plurality of data collection submodules one by one. Any family doctor submodule is used to generate a rehabilitation plan for a target patient corresponding to the family doctor based on the second health data encrypted after the first health data collected by the corresponding data collection submodule and using the rehabilitation plan generation model; The data encryption module includes: a key generation submodule, the key generation submodule is respectively connected to the plurality of data acquisition submodules by signal, and is respectively connected to the plurality of family doctor submodules by signal, the key generation submodule is used to generate a key pair based on the corresponding relationship between the data acquisition submodule and the family doctor submodule, and any key pair includes a public key and a private key; The key generation submodule generates a key pair based on the corresponding relationship between the data collection submodule and the family doctor submodule, including the following steps: S011, respectively obtaining data acquisition objects of multiple data acquisition submodules; A data acquisition object of a data acquisition submodule , where i is a positive integer and represents the number of the data acquisition submodule. represents the identity information of the target patient corresponding to the i-th data acquisition submodule in the data acquisition module, and the identity information is the ID card number of the target patient; S012. Obtain rehabilitation management objects of multiple family doctor sub-modules respectively; A rehabilitation management object of the family doctor submodule , where j is a positive integer and represents the number of the family doctor submodule. Represents the identity information of the target patient corresponding to the j-th family doctor submodule in the family doctor module; S013. Based on the data collection object of the data collection submodule and the rehabilitation management object of the family doctor submodule, obtain the corresponding relationship between the data collection submodule and the family doctor submodule: If the data collection object of a data collection submodule and the rehabilitation management object of a family doctor submodule are the same target patient, then there is a corresponding relationship between this data collection object and this family doctor submodule; otherwise, there is no corresponding relationship; Step S013 is based on the data collection object of each data collection submodule. , one by one with the rehabilitation management objects of the multiple family doctor sub-modules Make a match; Among them, the corresponding relationship between the i-th data collection submodule and the j-th family doctor submodule satisfies the following matching model: ,in, Indicates that there is a corresponding relationship between the i-th data collection submodule and the j-th family doctor submodule. It means that there is no corresponding relationship between the i-th data collection submodule and the j-th family doctor submodule; S014. Based on each set of corresponding relationships, generate a set of key pairs using the Paillier encryption algorithm, where each set of key pairs includes a public key and a private key; Based on a set of corresponding relationships , whose public key satisfies the following generation model: , ,in, is the public key modulus, g is the public key base, Representation based on correspondence The generated public key, Representation based on correspondence The first randomly generated prime number , Representation based on correspondence Randomly generated second prime number ; Among them, based on a set of corresponding relations ( ), whose private key satisfies the following generation model: ,in, is the Carmichael function value used for decryption, is the inverse function value used for decryption, Representation based on correspondence The generated private key, express and The least common multiple of Representation based on correspondence The generated public key modulus, Representation based on correspondence The generated public key base, mod represents the remainder operation; S015. Distribute the public key in the key pair to the corresponding data encryption submodule, and distribute the private key in the key pair to the corresponding family doctor submodule; Among them, based on a set of corresponding relations Key pair , the public key Distribute to the i-th data encryption submodule in the corresponding data encryption module, and pass the private key Distribute to the jth family doctor submodule in the corresponding family doctor module; The data encryption module also includes multiple data encryption sub-modules, and the multiple data encryption sub-modules are respectively connected to the multiple data acquisition sub-modules in a one-to-one signal manner. The multiple data encryption sub-modules are respectively connected to the multiple family doctor sub-modules in a one-to-one signal manner based on the correspondence between the data acquisition sub-modules and the family doctor sub-modules. The multiple data encryption sub-modules are also signal-connected to the key generation sub-module. Any data encryption sub-module is used to encrypt the first health data collected by the corresponding data acquisition sub-module based on the corresponding public key to generate second health data; each family doctor sub-module decrypts the second health data based on the corresponding private key, and then generates a rehabilitation plan for the target patient through a rehabilitation plan generation model.

2. The family doctor management system for diabetic patients based on rehabilitation mode according to claim 1 is characterized in that: Also includes: A plurality of model training modules, each of which is used to train and verify the rehabilitation program generation model based on a number of historical samples from a corresponding diabetic patient, and obtain model parameters of the verified rehabilitation program generation model; A parameter optimization module, wherein the parameter optimization module is respectively connected to a plurality of model training modules and respectively connected to each family doctor sub-module signal, and is used for optimizing the model parameters of the rehabilitation program generation model in each family doctor sub-module connected thereto based on the model parameters and the number of historical samples of the plurality of model training modules.

3. The family doctor management system for diabetic patients based on rehabilitation mode according to claim 2 is characterized in that: Any model training module includes: historical sample submodule, historical sample encryption submodule, model training submodule; The historical sample submodule stores a number of historical samples from the same diabetic patient, any of which is a rehabilitation program recommended for the diabetic patient under a set of health data; The historical sample encryption submodule is signal-connected to the historical sample submodule and signal-connected to the model training submodule, and is used to generate a key pair based on a homomorphic encryption algorithm, encrypt the historical sample using the public key in the key pair, and distribute the private key in the key pair to the corresponding model training submodule; The model training submodule is signal-connected to the parameter optimization module, and is used to decrypt the encrypted historical samples based on the private key from the corresponding historical sample encryption submodule, and use the decrypted historical samples to train and verify the rehabilitation program generation model, and obtain the model parameters of the verified rehabilitation program generation model.

4. The family doctor management system for diabetic patients based on rehabilitation mode according to claim 3 is characterized in that: The model parameters are optimized by the following steps: The parameter optimization module obtains the number of historical samples used to train the rehabilitation program generation model in each model training submodule, and obtains the weighted weight of the corresponding model training submodule based on the number of historical samples of each model training submodule; The model parameters used to train the rehabilitation program generation model in each model training submodule are obtained respectively, and the optimized model parameters are obtained through the weighted average calculation model in combination with the weighted weights of each model training submodule.

5. The family doctor management system for diabetic patients based on rehabilitation mode according to claim 4 is characterized in that: The rehabilitation program generation model in any two family doctor sub-modules is the same.

6. A family doctor management method for diabetic patients based on a rehabilitation model, wherein the family doctor management method for diabetic patients based on a rehabilitation model is applicable to the family doctor management system for diabetic patients based on a rehabilitation model as claimed in claim 2, characterized in that: The steps include: The data acquisition module acquires first health data of the target patient; The data encryption module encrypts the first health data to generate second health data; The family doctor module generates a rehabilitation plan for the target patient based on the second health data and using a rehabilitation plan generation model; The visualization module visualizes the rehabilitation program.

7. The family doctor management method for diabetic patients based on the rehabilitation model according to claim 6 is characterized in that: The steps include: The model training module trains and verifies the rehabilitation program generation model based on a number of historical samples corresponding to the diabetic patients, and obtains model parameters of the verified rehabilitation program generation model; The parameter optimization module optimizes the model parameters of the rehabilitation program generation model in each family doctor sub-module connected to it based on the model parameters and historical sample quantities of multiple model training modules.

Images