A ccrc-based health monitoring system and applications thereof

Abstract

The application discloses a health monitoring platform based on CCRC and application thereof and belongs to the technical field of health monitoring, comprising a data acquisition module, which is used for collecting the basic information of the users of the monitoring platform at regular intervals according to the user information recorded by the platform; and a monitoring item recommendation module, which is used for establishing the detection items of the health monitoring data based on the personal preferences of the users of the platform to update the health information. In the application, after the basic information of the users of the platform is collected in real time by the data acquisition module, the health of the users can be monitored based on the user information, the information to be monitored by the users can be pushed based on the personal information and the collected user preference data, the health monitoring plan can be established according to the conditions and preferences of the users, the quality of the pension service is improved, the accurate monitoring of the health of the old people can be realized, and the cost of manpower and material resources is reduced.

Description

Technical Field

[0001] This invention belongs to the field of health monitoring technology, and in particular relates to a health monitoring system based on CCRC and its application. Background Technology

[0002] CCRC (Continuing Care Retirement Community) is a comprehensive elderly care model. It combines the advantages of home-based, community-based, and institutional care, aiming to provide seniors with all-round elderly care services and care. Originating in the United States, CCRC has gained increasing attention in China's elderly care industry due to its aging population. By 2020, China's population aged 60 and above reached 245 million, and is projected to exceed 400 million by 2050. Faced with the contradiction between the ever-increasing demand for elderly care services and the limited supply of resources, the CCRC model offers a more rational and effective way to allocate elderly care resources.

[0003] Patent application CN114708129A discloses a novel CCRC (Continuing Care Retirement Community) integrated health and elderly care complex. This complex comprises two types of land within a community: land for elderly care services and land for construction. The elderly care service land is used to construct a smart community system, a smart elderly care system, a smart medical system, a health management system, a smart financial system, a community living support system, and a service operation and management system. The construction land is used to construct residential-ownership-based home care and wellness apartments for sale or lease, each equipped with a smart home system. This invention combines a socialized elderly care center integrating medical care and rehabilitation with a property-ownership-based elderly care community, allowing seniors to live in a familiar environment and receive care services. However, in practical use, the platform lacks real-time monitoring capabilities for the seniors' health status. This means that the seniors' health in the elderly care environment depends on their own judgment or on health data testing under the guidance of a physician, which consumes significant human and material resources and warrants improvement. Summary of the Invention

[0004] The purpose of this invention is to address the problem that the health status of the elderly in the elderly care environment depends on the elderly’s own judgment or the monitoring of health data under the guidance of health physicians, which consumes a lot of human and material resources. Therefore, this invention proposes a health monitoring system based on CCRC and its application.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A health monitoring system based on CCRC includes a data acquisition module for periodically collecting basic information of users of the maintenance monitoring system based on user information recorded in the system.

[0007] The health monitoring module is used to monitor the health information of system users based on their personal basic information.

[0008] The personal information storage module is used to store the system users' basic personal information and health information;

[0009] The monitoring item recommendation module is used to recommend monitoring items to system users based on their personal preferences, so as to update their health information.

[0010] As a further description of the above technical solution:

[0011] The system includes a data preprocessing module, which is communicatively connected to the data acquisition module. The data preprocessing module processes the acquired basic information data of system users and extracts the data feature information corresponding to the system users.

[0012] As a further description of the above technical solution:

[0013] The data preprocessing module extracts data feature information corresponding to system users, specifically including:

[0014] Data cleaning involves removing missing and outlier values ​​and performing data standardization.

[0015] Data transformation converts text data into numerical features and health monitoring items into feature vectors.

[0016] As a further description of the above technical solution:

[0017] The project recommendation process of the monitoring project recommendation module is as follows: After setting the types of health monitoring projects for system users as a given set of projects, a ranking function is obtained through training. The ranking function generates a ranking list of projects for the target user and recommends the list content to the user. The recommendation of monitoring projects is completed through the ranking learning model.

[0018] As a further description of the above technical solution:

[0019] By employing an adversarial foraging method, different basic recommendation models are established to correspond to individual birds in the flock. The foraging behavior of individual birds is used to obtain the prediction value of each item for a single user. The prediction values ​​of the basic recommendation models are then used to construct a feature matrix, i.e., a dataset of auxiliary information, through feature fusion. Based on the characteristics of adversarial foraging, at least two feeders are established to simulate the foraging behavior of the flock. The foraging of the flock is simulated under the interaction of the two feeders. By monitoring the foraging behavior of individual birds, the optimized ranking recommendation of monitored items is obtained.

[0020] As a further description of the above technical solution:

[0021] The monitoring project recommendation module also includes a discriminator and a parameter update module. The discriminator is used to evaluate the quality of the recommended sequence obtained from the foraging behavior of individual birds. After receiving the evaluation result from the discriminator, the feeder updates the current recommended sequence of the feeder to optimize the recommendation result. The parameter update module is used to update the parameters of the discriminator according to the evaluation score of the discriminator.

[0022] As a further description of the above technical solution:

[0023] The monitoring project recommendation module also includes an alternating training module, which completes the project after alternating training of the feeder and discriminator and meeting the stopping conditions.

[0024] As a further description of the above technical solution:

[0025] The basic information of the system users includes age, gender, health status, hobbies, and records of past participation in health monitoring and care activities.

[0026] As a further description of the above technical solution:

[0027] After the health monitoring module collects health information that exceeds the set threshold through the data acquisition module, it transmits the information to the monitoring item recommendation module for data weighting.

[0028] As a further description of the above technical solution:

[0029] An application of a health monitoring system based on CCRC is disclosed. The health monitoring system is set on a mobile terminal, which includes an interaction module. The mobile terminal can modify, view, or use the data acquisition module and personal information storage module of the health monitoring system through the interaction module, and can also provide feedback on its own health status through the interaction module for health monitoring in the context of continuous care for elderly people in retirement communities.

[0030] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0031] 1. In this invention, after collecting basic information of the user platform in real time through the data acquisition module, health monitoring can be carried out based on user information. Furthermore, based on personal information and collected user preference data, the information to be monitored for the user can be pushed according to their preferences. This is conducive to establishing a health monitoring plan according to the user's own situation and preferences, improving the quality of elderly care services, and is expected to achieve accurate monitoring of the health of the elderly and reduce human and material costs.

[0032] 2. In this invention, the designed monitoring item recommendation module uses simulated bird foraging behavior detection to recommend items to different individuals. This addresses the issue of insufficient preference characteristics among the elderly by improving recommendation accuracy through adversarial interaction between a discriminator and a feeder. The discriminator assesses the matching degree between the currently recommended health monitoring items and the actual needs of the elderly. It continuously learns and optimizes the recommendation strategy based on feedback from the elderly and historical data. The feeder generates new monitoring item recommendations and engages in adversarial interaction with the discriminator, attempting to generate recommendations that increasingly closely match the elderly's actual needs, while the discriminator strives to identify and filter out unsuitable recommendations. This adversarial interaction helps improve recommendation accuracy and facilitates simultaneous optimization using different algorithms to obtain the most suitable order of monitoring item recommendations based on individual health data. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the overall structure of a health monitoring system based on CCRC proposed in this invention;

[0034] Figure 2 This is a schematic diagram of the data preprocessing module structure of a health monitoring system based on CCRC proposed in this invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] Please see Figures 1-2 The present invention provides a technical solution: a health monitoring system based on CCRC, including a data acquisition module, used to periodically collect basic information of users of the maintenance monitoring system according to the user information recorded in the system;

[0037] The health monitoring module is used to monitor the health information of system users based on their personal basic information.

[0038] The basic information of the system users includes age, gender, health status, hobbies and past records of participation in health monitoring and care activities. After the health monitoring module collects health information that exceeds the set threshold through the data acquisition module, it transmits the information to the monitoring item recommendation module for data weighting.

[0039] The data acquisition module collects user information, including cloud-based acquisition of user information for detection data, and wearable monitoring units. By wearing the monitoring unit on a designated location on the patient's body using a pre-made wearable device, the monitoring unit monitors the patient's heart rate, body temperature, and blood pressure in real time. The monitoring unit then transmits the collected user health information to the platform server and stores the information through the personal information storage module.

[0040] Meanwhile, user personal information can be stored using a blockchain-based information dissemination system. This system protects user privacy data through the blockchain's public and private keys. The health monitoring module monitors user health information based on their basic personal information. Specifically, the platform server compares the patient's health information stored internally with its own data. The difference is then transmitted wirelessly to the platform server, which in turn sends the information to the user's mobile terminal and related mobile devices of their family members. When the monitoring unit detects a low health index or a fall based on position and posture sensor data, the wearable terminal transmits the patient's current status to the platform via wireless communication. The platform then calls upon CCRC service community personnel for timely treatment, improving patient safety.

[0041] The personal information storage module is used to store the system users' basic personal information and health information;

[0042] The monitoring item recommendation module is used to recommend monitoring items to system users based on their personal preferences, so as to update their health information.

[0043] It also includes a data preprocessing module, which is communicatively connected to the data acquisition module. The data preprocessing module processes the acquired basic information data of system users and extracts the data feature information corresponding to system users.

[0044] The data preprocessing module extracts data feature information corresponding to system users, specifically including:

[0045] Data cleaning involves removing missing and outlier values ​​and performing data standardization.

[0046] Data transformation converts text data into numerical features and health monitoring items into feature vectors;

[0047] The monitoring item recommendation module is trained using data from the data preprocessing module to construct a monitoring item recommendation model, which specifically includes:

[0048] When recommending monitoring items to system users, the system users' health monitoring items are set as a given set of items. After training, a ranking function is obtained. The ranking function generates a ranking list of items for the target users and recommends the list to the users. The recommendation of monitoring items is completed through the ranking learning model.

[0049] By adapting the adversarial foraging method of bird flocks, different recommendation models are used to correspond to individual birds in the flock. The foraging behavior of individual birds is used to obtain the prediction value of each user for each item. The prediction values ​​of the basic recommendation model are used to construct a feature matrix through feature fusion, which is the dataset of auxiliary information. Based on the characteristics of adversarial foraging, the optimal foraging behavior of individual birds is determined to obtain the optimal recommendation of monitoring items.

[0050] Initialization: Initialize the parameters of the bird flock foraging algorithm, such as flock size, maximum number of iterations, step size, etc.

[0051] Building the base model: Construct multiple base recommendation models as individual birds, and fuse their predicted rating vectors to construct a feature matrix.

[0052] Construct a ranking learning recommendation dataset: Combine user, activity, and predicted rating vectors into quadruples to form a ranking learning recommendation dataset, and divide the dataset into training, validation, and test sets.

[0053] Simulate bird flock foraging behavior: Establish at least two feeders that simulate bird flock foraging behavior, and find the optimal recommended sequence by iteratively updating the recommended sequence and identifying the foraging behavior of the individual bird with the best foraging ability in the flock.

[0054] After obtaining the optimal recommended sequence through simulation, the feeder uses a discriminator to determine whether the current foraging behavior is optimal.

[0055] Evaluate the quality of the recommended sequences: Input the recommended sequences generated by the feeder into the discriminator to evaluate their quality.

[0056] Update the recommended sequence: Based on the discriminator's evaluation results, update the recommended sequence to make it closer to the optimal sequence.

[0057] Training the discriminator: Evaluating the quality of recommended sequences. The discriminator gives an evaluation score based on the quality of the recommended sequences. During training, the discriminator learns the quality of the current sequence by ranking the recommended dataset.

[0058] Both the feeder and the discriminator are neural networks. The neural network can fit the ranking results and calculate the loss function and probability distribution list. At the same time, it retains the feedback information and can provide auxiliary feedback to the two feeders.

[0059] By setting up two feeders, at least one feeder becomes the second feeder after being externally connected to the discriminator for information feedback, and the first feeder is internally connected to the other feeder through the second feeder. The internal feedback connection between the first feeder and the second feeder can optimize the transmission of recommended items, enhance the ranking ability, and allow training information and content to be fed back to each other. This is beneficial for exploring the probability distribution through the feeder combined with the discriminator and improving the performance of the second feeder itself.

[0060] The discriminator takes real data as input and outputs the level of interest of system users in the project. The discriminator continuously judges based on the optimized content provided by the second feeder and updates itself after the judgment feedback.

[0061] The internal objective function of the feeder improves its performance by maximizing the cross-entropy loss, in response to feedback from the discriminator. During training, the first feeder samples the generated samples. The internal objective function... for:

[0062] ;

[0063] Where E represents the expected average value, i represents the sample input to the network, and u n Let n represent the user, r represent the sample category, and r = 1 represent the real sample, and r = 0 represent the generated sample. For the scoring function of the second feeder;

[0064] Let be the probability distribution of the real sample i. To generate the probability distribution of sample i, and The parameters for the first feeder and the second feeder are respectively;

[0065] First, fix the parameters. The parameters of the second feeder are not updated; the parameters of the first feeder are optimized, and only the parameters of the second feeder are updated. When r=0, the first feeder generates sample i, and the scoring function of the second feeder is expected to... A high score reduces the overall expected value; then, the parameters are fixed. Without updating the first feeder, repeat the above steps to optimize the second feeder;

[0066] Update discriminator parameters: Based on the evaluation score, update the discriminator parameters to enable them to more accurately evaluate the quality of the recommended sequences;

[0067] The external objective function of the discriminator is to improve the discriminator's performance by maximizing the cross-entropy loss, aiming to make the discriminator classify real samples as 1 and generated samples as 0.

[0068] Specifically:

[0069]

[0070] Where i represents the input network sample, u n Represents the system user, n is the response code generated for the user, r represents the input sample category, r=1 represents a real sample, r=0 represents a generated fake sample;

[0071] and Let be the probability distribution of the real sample i. Let θ be the probability distribution of the generated samples from the second feeder. These are the parameters for the second feeder and the discriminator, respectively. The rating given to the user is [0, 1], therefore... and The range of values ​​for all of them is [-∞, 0];

[0072] After parameter sampling, the training process is as follows:

[0073] Training the first feeder:

[0074] Using LambdaRank as the base model, the objective function is formed by combining the LambdaRank loss function with the generation probability of the first feeder, denoted as . ;

[0075] make ,but

[0076]

[0077] Through the The objective of finding γ is to find its partial derivative, and the optimal parameter γ is approximately obtained as follows:

[0078] ;

[0079] Where N is the reciprocal of the maximum cumulative gain, and the loss function of the first feeder is Pointloss, used to generate probabilities:

[0080] ;

[0081] Where K is the preference number of the characteristic user.

[0082] Repeat the above steps to train the second feeder, receiving difficult samples from the first feeder and providing fake samples to the discriminator, and fitting the function. Generation probability ;

[0083] By using information feedback, the optimal parameter θ* is found. The second feeder may perform poorly in the early training, so the first feeder is introduced to assist in training.

[0084] Training the discriminator:

[0085] The objective function is minimized by fitting the Dδscore(i, u) function;

[0086]

[0087] in, It is a latent correlation distribution, representing the true probability. The estimated probability of the second feeder is obtained through the interaction between models to achieve the optimal result. The discriminator uses RankNet as the bottom-side model and pairloss as the loss function.

[0088]

[0089] By training the second feeder, it is possible to generate fake items that are of interest to the current user. After the items are generated, a discriminator is used to adversarially determine whether the items actually exist.

[0090] Alternating training: The feeder and discriminator are trained alternately until the stopping condition is met, the number of iterations reaches the maximum value or the model performance no longer improves, that is, the model training is complete.

[0091] Generate a recommendation list: Use the recommendation sequence generated by the trained feeder as the final recommendation result.

[0092] Throughout the training process, the first and second feeders alternately generate recommended sequences and engage in adversarial training with the discriminator. Through continuous iterative optimization, the two feeders gradually improve the quality of the generated recommended sequences, while the discriminator gradually improves its ability to distinguish between real and generated sequences. Finally, when the model reaches the predetermined stopping condition, such as when the number of iterations reaches the maximum value or when the model performance no longer improves significantly within a certain period of time, the training process ends, and the model training is complete.

[0093] An application of a health monitoring system based on CCRC is disclosed. The health monitoring system is set on a mobile terminal, which includes an interaction module. The mobile terminal can modify, view, or use the data acquisition module and personal information storage module of the health monitoring system through the interaction module, and can also provide feedback on its own health status through the interaction module for health monitoring in the context of continuous care for elderly people in retirement communities.

[0094] In the above embodiments, the personal information storage module applied to the health monitoring system for executing the above method embodiments is a corresponding data storage system. The data storage system has at least one processor, a control module (chipset) coupled to at least one of the processors, a memory coupled to the control module, a non-volatile memory (NVM) / storage device coupled to the control module, at least one input / output device coupled to the control module, and a network interface coupled to the control module.

[0095] The processor may include at least one single-core or multi-core processor, and may include any combination of general-purpose processors or special-purpose processors (e.g., graphics processors, application processors, baseband processors, etc.). In some alternative implementations, the data storage system used in the health monitoring system can serve as the gateway or other data storage system device used in the health monitoring system described in the embodiments of this application.

[0096] In some alternative implementations, the data storage system applied to the health monitoring system may include at least one computer-readable medium (e.g., a memory or NVM / storage device) having instructions and at least one processor fused with the at least one computer-readable medium and configured to execute the instructions to implement the module thereby performing the actions described in this disclosure.

[0097] In one embodiment, the control module may include any suitable interface controller to provide any suitable interface to at least one of the processors and / or any suitable device or component communicating with the control module.

[0098] The control module may include a memory controller module to provide an interface to the memory. The memory controller module may be a hardware module, a software module, and / or a firmware module.

[0099] The memory can be used, for example, to load and store data and / or instructions for a data storage system applied to a health monitoring system. In one embodiment, the memory may include any suitable volatile memory, such as suitable DRAM.

[0100] In one embodiment, the control module may include at least one input / output controller to provide an interface to the NVM / storage device and (at least one) input / output device.

[0101] For example, an NVM / storage device can be used to store data and / or instructions. An NVM / storage device may include any suitable non-volatile memory (e.g., flash memory) and / or may include any suitable (at least one) non-volatile storage device (e.g., at least one hard disk drive (HDD), at least one optical disc (CD) drive, and / or at least one digital universal optical disc (DVD) drive).

[0102] An NVM / storage device may include storage resources that are physically installed as part of a data storage system used in a health monitoring system, or that can be accessed by the device without being part of it. For example, an NVM / storage device may be accessed via a network through at least one input / output device.

[0103] At least one input / output device provides an interface for the data storage system used in the health monitoring system to communicate with any other suitable device. The input / output device may include communication components, input components, sensor components, etc. A network interface provides an interface for the data storage system used in the health monitoring system to communicate via at least one network. The data storage system used in the health monitoring system can wirelessly communicate with at least one component of a wireless network according to any standard and / or protocol of at least one wireless network standard, such as accessing a wireless network based on a communication standard.

[0104] In this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance; the term "multiple" refers to two or more unless otherwise explicitly defined. The terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; "linking" can be a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0105] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A CCRC-based health monitoring system, characterized by, The system comprises a data collection module, which is configured to collect basic information of a user of a maintenance monitoring system at a predetermined interval according to user information recorded by the system; a health monitoring module, which is configured to monitor health information of the user of the monitoring system through personal basic information of the user of the system; a personal information storage module, which is configured to store the personal basic information and the health information of the user of the system; a monitoring item recommendation module, which is configured to establish a monitoring item recommendation module to recommend a detection item of health monitoring data based on personal preferences of the user of the system, so as to update the health information; The project recommendation process of the monitoring item recommendation module: after setting the type of the health monitoring item of the user of the system as a given item set, the sorting function is obtained through training, the sorting function generates an item ranking list for the target user, and the list content is recommended to the user, and the recommendation of the monitoring item is completed through the sorting learning model; Through the bird foraging anti-foraging method, different basic recommendation models corresponding to individual birds in the bird group are established, the behavior of individual birds foraging is used to obtain the predicted value of each item of a single user, and the predicted value of the basic recommendation model is used to construct a feature matrix, that is, a data set of auxiliary information, at least two feeding devices simulating the foraging behavior of the bird group are established according to the anti-foraging characteristics, the foraging simulation of the bird group is carried out under the interaction of the two feeding devices, the foraging behavior of the individual bird is monitored to obtain the optimized monitoring item ranking recommendation; The monitoring item recommendation module further comprises a discriminator and a parameter updating module, the discriminator is used to evaluate the quality of the recommended sequence obtained by the foraging behavior of the individual bird, the feeding device updates the current recommended sequence of the feeding device according to the evaluation result of the discriminator to optimize the recommended result, and the parameter updating module is used to update the parameters of the discriminator according to the evaluation score of the discriminator. The basic information of the user of the system includes age, gender, health status, interest, and record of past participation in health monitoring and nursing activities.

2. The CCRC-based health monitoring system of claim 1, wherein, The system further comprises a data preprocessing module, which is in communication connection with the data collection module, and performs data processing on the collected basic information data of the user of the system through the data preprocessing module, and extracts the data feature information corresponding to the user of the system through the data preprocessing module.

3. The CCRC-based health monitoring system of claim 2, wherein, The data preprocessing module extracts the data feature information corresponding to the user of the system, which specifically includes: Data cleaning, removing missing values and outliers, and performing data standardization processing; Data conversion, converting text data into numerical features, and converting health monitoring items into feature vectors.

4. The CCRC-based health monitoring system of claim 1, wherein, The internal objective function of the feeder improves the performance of the feeder with discriminator feedback by maximizing the cross-entropy loss. During the training process, the first feeder samples the generated samples, and the internal objective function is: f(x) = -log p(x) ; where E represents the expected average value, i represents a sample input to the network, N represents the total number of users, u n represents a user, n is the user number, r represents a sample category, and if r is 1, it is real, and if r is 0, it represents a generated sample, is a score function of the second feeding device; is the probability distribution for real sample i, is the probability distribution for generated sample i, and are the trained weight parameters for the first and second feeders, respectively; First, fix the parameters , do not update the second feeder parameters, optimize the first feeder, only update , when r = 0, the first feeder generates sample i, and hopes that the second feeder's score function , the evaluation score is high, the overall expectation value is small, then fix the parameters , do not update the first feeder, repeat the optimization of the second feeder; Objective function outside the discriminator That is, by maximizing the cross-entropy loss to improve the performance of the discriminator, hoping that the discriminator can distinguish the real samples as 1 and the generated samples as 0; and is a probability distribution of real samples i, is a probability distribution of generated samples of the second feeder, is a training weight parameter of the discriminator, represents the score given to the user, the output result is [0, 1], so and The value range of is [-∞, 0].

5. The CCRC-based health monitoring system of claim 1, wherein, The monitoring item recommendation module further comprises an alternating training module, which is configured to complete the project after alternating training of the feeding device and the discriminator and meeting the stop condition.

6. The CCRC-based health monitoring system of claim 1, wherein, The health monitoring module transmits the health information exceeding the set threshold to the monitoring item recommendation module for data weighting after collecting the health information through the data collection module.

7. The CCRC-based health monitoring system of claim 1, wherein, The health monitoring system is arranged in a mobile terminal, the mobile terminal comprises an interaction module, and the data collection module and the personal information storage module in the health monitoring system are modified, viewed or used through the interaction module, and the health state of the mobile terminal can be fed back through the interaction module to be applied to the health monitoring of the retired community old people in continuous care.

Images