Health data analysis method and system based on Internet of Things sharing

Abstract

The invention discloses a health data analysis method and system based on Internet of Things sharing, and particularly relates to the field of Internet of Things technology and medical care information.The method comprises the steps that Internet of Things equipment is deployed, and biological health data are collected in real time; analyzing the health data to obtain living body in-vivo health data and living body in-vitro motion data, and performing dynamic analysis and clustering analysis on the data to obtain living body in-vivo health indexes and living body in-vitro health indexes; group analysis is carried out on the two types of health indexes to obtain the correlation degree of the two types of indexes, a prediction model is established according to the correlation degree and optimized, and alliance chain nodes are constructed to establish a health data sharing network; according to the method, data sharing is realized by dynamically analyzing the data, establishing the prediction model and optimizing the model, so that the problem of lack of dynamic analysis, accurate prediction and data sharing of the data in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the fields of Internet of Things technology and healthcare informatics, and more particularly, to a health data analysis method and system based on Internet of Things sharing. Background Art

[0002] With the rapid development of Internet of Things technology, it has been widely used in production and life, including the application of biological health analysis. Through ear tag sensors, collar motion sensors, milk composition analyzers, blood testing equipment, environmental sensors and other terminal devices, the heart rate data, blood data, milk data, exercise steps, temperature and humidity data of dairy cows are collected in real time.

[0003] With the development of large-scale dairy farming, the impact of cow health management on farming efficiency is becoming increasingly significant. However, existing cow health analysis technologies still have some shortcomings. First, the analysis of cow health data is mostly static, single-shot, lacking dynamic tracking and group-related data mining. Second, there is a lack of analysis and matching of the correlation between health indicators. Third, there are significant data barriers between farms, medical institutions, data centers, and other entities, resulting in a conflict between data privacy protection and sharing needs, making collaborative diagnosis and management difficult. Therefore, an integrated method is needed that can achieve dynamic group analysis, high-precision prediction, and secure sharing of dairy cow health data. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a health data analysis method and system based on Internet of Things sharing, and solves the problems raised in the above-mentioned background technology through the following solutions.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a health data analysis method based on Internet of Things sharing, comprising the following steps: S1: Set the predetermined biological group as the target biological group, and collect the health data of the target biological group in real time by deploying IoT devices. The health data includes the in vivo health data and the in vitro movement data. S2: Dynamically analyze the in vivo health data of the organism to obtain in vivo health indicators; divide the in vitro motion data of the organism to obtain biological motion data and in vitro environmental data, and perform cluster analysis on the biological motion data and in vitro environmental data to obtain in vitro health indicators; S3: By performing group analysis on internal and external health indicators, the correlation between internal health data and external movement data is obtained; a health data prediction model is established based on the correlation, and the prediction results of the health data prediction model are encoded and transmitted to the IoT terminal; S4: Optimize the health data prediction model and build a consortium chain architecture network node based on the optimized model. The network nodes include biological type nodes, medical institution nodes, and cow health data nodes. The Internet of Things health data sharing is achieved through data transmission and access between nodes.

[0006] Preferably, the S2 is performed by analyzing the health data in the organism Perform dynamic analysis and dynamic feature extraction to obtain the biological body health data feature vector, and obtain the body health index by calculating the feature vector , specific indicators of health in the body The calculation formula is ,in Represents biological individuals, Indicates time, Indicates the type of health data, Represents an individual organism exist Real-time health data For the Time change rate of in vivo health data, For biological individuals exist time Standard deviation of in vivo health data, is a dynamic weighted summation function.

[0007] The K-means clustering algorithm is used for biological motion data and in vitro environmental data to calculate the Euclidean distance between biological motion data and in vitro environmental data and the cluster center to obtain in vitro health indicators. Specific in vitro health indicators The calculation formula is in Represents an individual organism exist Momentary sports data, Represents an individual organism exist Real-time environmental data, For the Cluster centers, represents the number of cluster categories, represents the Euclidean distance of cluster centers.

[0008] Preferably, the S3 has an effect on the health indicators in the body and in vitro health indicators Perform group analysis, use association rule mining algorithm to extract item sets, and obtain the association degree by calculating the confidence and lift of the item sets. The confidence formula is , Lift Formula , where A represents the motion event outside the organism and B represents the physiological state event inside the organism.

[0009] Preferably, the health data analysis system based on Internet of Things sharing includes: Data collection module: Set the predetermined biological group as the target biological group, and collect the health data of the target biological group in real time by deploying IoT devices; Data processing module: by analyzing the health data, obtain the biological body's internal health data and the biological body's external movement data; by calculating the biological body's internal and external data, obtain the internal health indicators and the external health indicators; Prediction model establishment and transmission module: establishes a health data prediction model through correlation analysis and transmits the encoded prediction results to the IoT terminal; Model optimization and sharing module: Use the LSTM algorithm to optimize the health data prediction model, build a health data sharing network with a consortium chain architecture by setting up network nodes, manage access rights through the smart contract engine, and realize the sharing of health data in the Internet of Things.

[0010] Technical effects and advantages of the present invention: 1. The present invention sets a predetermined biological group as the target biological group and collects health data of the target biological group in real time by deploying IoT devices; 2. The present invention obtains in vivo health indicators by dynamically analyzing in vivo health data, and obtains in vitro health indicators based on cluster analysis of biological motion data and in vitro environmental data, thus achieving dynamic tracking and classification of health data; 3. The present invention calculates confidence and lift through association rule mining, establishes the correlation between exercise type and clinical indicators; constructs a health data prediction model, outputs the probability distribution of exercise type, and transmits the prediction results after SHA-256 hash encoding, thereby improving the accuracy of health status prediction; 4. The present invention iteratively optimizes the health data prediction model through the LSTM algorithm, establishes an alliance chain and a sharing network, and stores hashed health data through nodes to achieve cross-institutional collaborative management and health data sharing. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 Schematic diagram of the method of the present invention.

[0012] Figure 2 This is a logical diagram of data collection and processing of the present invention.

[0013] Figure 3 Construct a logic diagram for the prediction model of the present invention.

[0014] Figure 4This is a logical diagram of model optimization and data sharing of the present invention.

[0015] Figure 5 Schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] See also Figure 1 - Figure 4 As shown, the embodiment of the present invention provides a health data analysis method based on Internet of Things sharing. By collecting cow health data, dynamically analyzing cow health data, establishing a prediction model based on correlation, optimizing the model and building a shared network to achieve data sharing, the Internet of Things can share and analyze cow health data. The embodiment of the present invention discloses a health data analysis method based on Internet of Things sharing, including the following steps: S1: Set the predetermined biological group as the target biological group, and collect the health data of the target biological group in real time by deploying IoT devices. The health data includes the in vivo health data and the in vitro movement data. S2: Dynamically analyze the in vivo health data of the organism to obtain in vivo health indicators; divide the in vitro motion data of the organism to obtain biological motion data and in vitro environmental data, and perform cluster analysis on the biological motion data and in vitro environmental data to obtain in vitro health indicators; S3: By performing group analysis on internal and external health indicators, the correlation between internal health data and external movement data is obtained; a health data prediction model is established based on the correlation, and the prediction results of the health data prediction model are encoded and transmitted to the IoT terminal; S4: Optimize the health data prediction model and build a consortium chain architecture network node based on the optimized model. The network nodes include biological type nodes, medical institution nodes, and cow health data nodes. The Internet of Things health data sharing is achieved through data transmission and access between nodes.

[0018] In step S1, a predetermined number of dairy cows with the same breed attributes are selected as the target biological group to ensure data consistency and comparability. Real-time dairy cow health data is collected by deploying IoT devices, including ear tag sensors, collar motion sensors, milk composition analyzers, blood testing equipment, and environmental sensors. Ear tag sensors on cows' ears collect real-time heart rate and body temperature data. Collar motion sensors on cows' necks collect real-time external movement data. Milk component analyzers and blood testing equipment collect real-time milk and blood data. Environmental sensors installed on the cowshed ceiling collect real-time external temperature and humidity data. In vitro motion data of organisms are usually divided by sensors or monitoring equipment. The data collection location and sensor type are the basic basis for division. Sensors directly attached to the organism can collect biological motion data. Sensors fixed in the external environment of the organism can collect in vitro environmental data. In vitro motion data of organisms are divided by different sensor types.

[0019] In step S2, the health data of the organism is analyzed Perform dynamic analysis and dynamic feature extraction to obtain the biological body health data feature vector, and obtain the body health index by calculating the feature vector , The calculation formula is ,in Represents an individual organism (such as a cow), Indicates time, Indicates the type of health data (such as heart rate, blood, etc.), Represents an individual organism exist Real-time health data For the The time change rate of the health data in the class reflects the data trend, For biological individuals exist time The standard deviation of the in vivo health data reflects the stability of the data. is a dynamic weighted sum function, where the time change rate weight is set to 0.6 and the standard deviation weight is set to 0.4; For example: cows Internal heart rate Time rate of change of health data ,in The time interval is set to 6 hours, and the cows If the heart rate is 70, 72, 75, or 73 (beats / minute), then hour , at this time the heart rate is on the rise, hour , at this time the heart rate is on a downward trend; For biological individuals exist time Standard deviation of in vivo health data, ,in Indicates the mean heart rate, Indicates the number of data, calculate the stability of four sets of data as of Standard deviation of time ; In vitro health indicators are obtained by performing K-means clustering on biological motion data and in vitro environmental data, and calculating the Euclidean distance between biological motion data and in vitro environmental data and the cluster center. The specific calculation formula is: in Represents an individual organism exist Momentary sports data, Represents an individual organism exist Real-time environmental data, For the Cluster centers, represents the number of cluster categories, represents the Euclidean distance of cluster centers; For example: The formula for calculating Euclidean distance is: in Indicates the The motion data and environmental data of the cluster centers are collected. For example, the daily step counts and cowshed temperatures of three cows are selected. The step counts of cow A, cow B, and cow C are set to 7500, 5200, and 3000 respectively. The cowshed temperatures of cow A, cow B, and cow C are set to 25°C, 28°C, and 30°C, respectively. Based on the data distribution, samples with large differences are selected first. A (7500, 25) and C (3000, 30) are selected as the initial cluster centers. A round of clustering is performed first: The Euclidean distance from cow A to the initial cluster center A is 0, and the Euclidean distance from cow A to the initial cluster center C is , then cattle A belongs to category A; The Euclidean distance from cow B to the initial cluster center A is 2300, and the Euclidean distance from cow B to the initial cluster center C is 2200, so cow B belongs to cluster C. The Euclidean distance from cow C to the initial cluster center A is 4500, and the Euclidean distance from cow C to the initial cluster center C is 0, so cow C belongs to class C; Next, determine the final cluster center: Class A center is (7500,25), Class C sports data center , Environmental Data Center , so the center of Class C is (4100,29). Recalculating the Euclidean distance based on the new Class C center shows that the in vitro health index of cow A is 0, indicating that its exercise and environmental status are consistent with those of the same group (high activity, suitable temperature). The in vitro health indexes of cows B and C are approximately 1100, indicating that although they belong to the same group (low activity, high temperature), there are certain differences from the centers of the same group (the number of exercise steps deviates from the center value), which may indicate abnormal activity (for example, the number of steps of cow C is too low, and attention should be paid to whether it is sick).

[0020] In step S3, the correlation is used to measure the correlation strength between external health indicators (such as exercise environment status) and internal health indicators (such as physiological status). The association rule mining algorithm is used to analyze the cow health dynamic behavior type data. First, the item set is extracted from the data, and the minimum support is set to 10%. The item set includes regular exercise, blood standard, heart rate standard, different temperature behavior type labels and health indicator features. The correlation is obtained by calculating the confidence and lift between the item sets. The confidence formula is: The lift formula is , A represents the biological external motion event, B represents the biological internal physiological state event, and finally the correlation degree between the biological internal health data and the biological external motion data is obtained; For example, when a cow's exercise environment (external indicators) is abnormal, its physiological state (internal indicators) is likely to be abnormal as well. By monitoring external indicators (such as a sudden drop in exercise steps or being in a high-temperature environment), early warning of internal health risks (such as diseases and metabolic disorders) can be provided without waiting for the test results of internal indicators (such as blood and milk data), enabling early intervention. When A represents a high-temperature and low-activity event for a cow, and B represents an abnormal heart rate event for a cow, if the lift is greater than 1, it indicates that event A is positively correlated with event B; if the lift is equal to 1, it indicates that there is no correlation between A and B; and if the lift is less than 1, it indicates that there is a negative correlation between A and B.

[0021] A health data prediction model is built based on correlation. This deep learning model receives a 128-dimensional dynamic feature vector of cow health data at the input layer, including multi-dimensional time series data such as heart rate data, blood data, milk data, and exercise data from the past few days. The hidden layer consists of a three-layer fully connected network with 64, 32, and 16 neurons, respectively, all using the RELU activation function for nonlinear transformation. The output layer passes The function outputs the probability distribution of cow movement types, for example, the probability of "heart rate fluctuation" and the probability of "regular movement." The loss function uses cross-entropy. The model is trained using the Adam optimizer with a learning rate of 0.001 and 100 iterations. The training set and test set are split in a 7:3 ratio, with 70% of the cow health data used for training and 30% for testing. The probability distribution output by the health data prediction model is hashed. First, the probability value data is converted into a UTF-8 encoded string. Then, a 256-bit hash value is calculated using the SHA-256 algorithm. The hash value serves as the unique identifier of the data. Data is transmitted through the MQTT protocol. Each data packet contains a hash value and a timestamp. To ensure transmission security, it is encrypted using the TLS1.3 protocol. Authentication is used to ensure the legitimacy of the identities of both parties in the data transmission, preventing data from being tampered with or stolen during transmission.

[0022] In step S4, the Long Short-Term Memory (LSTM) network algorithm, a deep learning model that excels at processing time series data, optimizes the health data prediction model by iteratively adjusting network parameters to reduce prediction errors and improve the ability to capture long-term dependencies on cow health dynamics data. The specific optimization process can be divided into the following key steps: The core goal of optimization is to minimize the prediction error so that the model's prediction results for the cow's health dynamic data are closer to the true value. Therefore, the prediction error function is defined as the root mean square error (RMSE), and its calculation formula is: in, is the real health data value, is the model prediction value, n is the number of samples, and the objective function is clearly "minimizing RMSE", that is, making the error value smaller through iterative optimization; To adapt to the time series characteristics of health data, the LSTM network parameters were set as follows: an input dimension of 32, corresponding to the 32 key features extracted from the cow health dynamics data, ensuring that the model focuses on core influencing factors; a hidden layer dimension of 64, as the number of hidden layer neurons determines the model's feature learning ability. The 64-dimensional setting ensures learning ability while avoiding data overfitting; an output dimension of 1, with the final output being a single predicted value; and a time step of 24 hours. The model uses the past 24 hours of cow health data as input, which aligns with the diurnal cycle of cow physiological rhythms and improves the ability to capture short-term health trends. The LSTM algorithm optimization process effectively addresses the problems of "insufficient capture of time dependence" and "high prediction error" in health data prediction through reasonable network design, refined parameter control, and targeted weight adjustment, laying the foundation for the model's practical value. Based on the optimized health data prediction model, a health data sharing network with a consortium chain architecture is constructed. Data sharing is achieved through the construction of the consortium chain, node network deployment, PBFT configuration phase, and smart contracts. The specific steps are as follows: By analyzing the model optimized by the LSTM algorithm, it was determined that the core goal is to achieve secure sharing of cow health data and controllable permissions. The network architecture was determined to be an "alliance chain + multi-node model". Node types include biological type nodes, medical institution nodes, and cow health data nodes. The data interaction path between nodes was clarified. Biological node deployment is based on the number of individual cows, and a lightweight consortium chain terminal is installed through a mobile phone or home network terminal. The number of medical institution node deployment is set to three, which are used for community veterinary hospitals, physical examination centers, and regional private diagnostic offices. It is recommended to use a dedicated server, install a consortium chain full-node terminal, and configure the data storage module. Medical institutions need to provide corresponding qualification certificates, and the server accesses the network based on the qualification certificates. Cow health data nodes are developed based on the types of health data collected by IoT devices and are used in the IoT shared health data analysis system. The consortium chain master node is installed by deploying a high-performance server, integrating data synchronization and consensus modules. Consensus parameters are configured based on the PBFT algorithm to ensure consistency in data transmission between nodes. Basic parameter settings enable five core nodes to participate in consensus, including three medical institution nodes, one biological type node, and one cow health data representative node, meeting the requirement of more than four nodes. Three rounds of consensus are conducted on the five nodes to ensure data consistency even when some nodes fail. The correctness of the PBFT configuration is determined by simulating node data interactions. The cow health data representative initiates the interaction, such as the hash value of the cow health data. The five core nodes reach consensus according to the PBFT process, determining the consistency of each node's data and avoiding data redundancy caused by excessive data. The smart contract is compiled into bytecode and deployed to the consortium chain. Test interactions are initiated through the cow health data representative node, such as simulating authorized medical points to access data to verify the accuracy of the smart contract rule execution; the cow health data is SHA-256 hashed to generate a hash value, which is transmitted to the consortium chain for local storage through hash encryption; when a medical institution node initiates a data query during access control, it must pass the permission rules defined in the smart contract to access the biological type node. The access request must be verified by the PBFT consensus before data sharing between nodes can be achieved.

[0023] The embodiment of the present invention provides a health data analysis system based on Internet of Things sharing, and the specific modules include the following: Data collection module: sets a predetermined biological group as the target biological group, and collects health data of the target biological group in real time by deploying IoT devices, including ear tag sensors, neck collar motion sensors, milk composition analyzers, blood testing equipment, and environmental sensors; Data processing module: Health data includes in-vivo health data and in-vivo movement data. By analyzing the in-vivo health data and in-vivo movement data, in-vivo health indicators and in-vivo health indicators are obtained. The confidence and lift are calculated based on the two health indicators to generate the correlation degree. Prediction model establishment and transmission module: establishes a health data prediction model through correlation data and transmits the encoded prediction results to the IoT terminal, using hash coding to transmit the data; Model optimization and sharing module: Use the LSTM algorithm to optimize the health data prediction model, build a health data sharing network with a consortium chain architecture by setting up network nodes, manage access rights through the smart contract engine, and realize the sharing of health data in the Internet of Things.

[0024] Secondly, the drawings of the embodiments disclosed in the present invention only involve structures related to the embodiments disclosed in the present invention. Other structures can refer to common designs. In the absence of conflicts, the same embodiment and different embodiments of the present invention can be combined with each other. Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A health data analysis method based on Internet of Things sharing, characterized in that: include: S1: Set the predetermined biological group as the target biological group, and collect the health data of the target biological group in real time by deploying IoT devices. The health data includes the in vivo health data and the in vitro movement data. S2: Dynamically analyze the in vivo health data of the organism to obtain in vivo health indicators; divide the in vitro motion data of the organism to obtain biological motion data and in vitro environmental data, and perform cluster analysis on the biological motion data and in vitro environmental data to obtain in vitro health indicators; S3: By performing group analysis on internal health indicators and external health indicators, the correlation between internal health data and external exercise data is obtained; Establish a health data prediction model based on the correlation, encode the prediction results of the health data prediction model and transmit them to the Internet of Things terminal; S4: Optimize the health data prediction model and build a consortium chain architecture network node based on the optimized model. The network nodes include biological type nodes, medical institution nodes, and cow health data nodes. The Internet of Things health data sharing is achieved through data transmission and access between nodes.

2. The health data analysis method based on Internet of Things sharing according to claim 1 is characterized in that: The health data: A predetermined number of cows are set as the target biological group. The cows need to meet the same breed attributes, and the cow health data is collected in real time through IoT devices. The health data includes heart rate data, blood data, milk data, exercise steps, temperature and humidity data.

3. The health data analysis method based on Internet of Things sharing according to claim 1 is characterized in that: The in vivo health indicators: Through the health data of organisms Perform dynamic analysis and dynamic feature extraction to obtain the health data feature vector of the organism, and construct a dynamic weighted fusion function through the feature vector Calculate the time change rate of the biological body's health data and the standard deviation of the body's health data to obtain the body's health index The in-vivo health data includes heart rate data, blood data, and milk data.

4. The health data analysis method based on Internet of Things sharing according to claim 1 is characterized in that: The biological motion data and in vitro environmental data: According to the sensor type and collection location, the motion sensor data attached to the organism is used as the basis for classification as biological motion data, and the environmental sensor data of environmental fixed sensors or organisms is used as the basis for classification as in vitro environmental data.

5. The health data analysis method based on Internet of Things sharing according to claim 1 is characterized in that: The in vitro health indicators: The K-means clustering algorithm is used for biological motion data and in vitro environmental data. According to the biological cluster center and the number of cluster categories, the Euclidean distance between the biological motion data and in vitro environmental data and the cluster center is calculated to obtain the in vitro health index. The biological motion data and in vitro environmental data include motion steps, temperature and humidity data.

6. The health data analysis method based on Internet of Things sharing according to claim 1 is characterized in that: The correlation between the in-vivo health data and the in-vivo exercise data: Health indicators in the body and in vitro health indicators Perform group analysis and use association rule mining algorithms to extract item sets and calculate the confidence and lift of the item sets to obtain the association degree, where A represents an in vitro motion event and B represents an in vivo physiological state event. The confidence degree is the probability of event B occurring under the condition that event A occurs. Lift is the ratio of the probability of event B occurring under the condition that event A occurs to the probability of event B occurring alone.

7. The health data analysis method based on Internet of Things sharing according to claim 1 is characterized in that: The health data prediction model: Model building process: The model matching mechanism is constructed by inputting correlation data. The model receives the health data of the organism in real time, outputs the probability value of the biological movement type, takes the probability value as the predicted matching result, and transmits it to the IoT terminal through hash coding. The predicted matching result output by the model is hashed and SHA-256 is used to generate a 256-bit hash value.

8. The health data analysis method based on Internet of Things sharing according to claim 1 is characterized in that: The health data prediction model is iteratively optimized: The LSTM algorithm is used to optimize the health data prediction model. First, the LSTM network parameters are set, the model learning rate and number of iterations are used to calibrate the error, and the gradient descent method is used to adjust the LSTM weights in each iteration.

9. The method for analyzing health data based on Internet of Things sharing according to claim 1, characterized in that: The alliance chain architecture network nodes include: Through the node access mechanism, biological type nodes are digitally authenticated for access, medical institution nodes need to provide qualification certificates reviewed by CA institutions for access, and cow health data nodes are the core nodes of the alliance chain, which are operated and accessed by institutions with data storage and processing qualifications. The PBFT consensus algorithm is used to conduct three rounds of consensus on the part with more than 4 nodes, and access rights are defined through smart contracts. The sharing of IoT health data is realized through data transmission and access between nodes.

10. A health data analysis system based on Internet of Things sharing, according to the health data analysis method based on Internet of Things sharing according to any one of claims 1-9, characterized in that: Includes the following modules: Data collection module: Set the predetermined biological group as the target biological group, and collect the health data of the target biological group in real time by deploying IoT devices; Data processing module: by analyzing the health data, obtain the biological body's internal health data and the biological body's external movement data; by calculating the biological body's internal and external data, obtain the internal health indicators and the external health indicators; Prediction model establishment and transmission module: establishes a health data prediction model through correlation analysis and transmits the encoded prediction results to the IoT terminal; Model optimization and sharing module: Use the LSTM algorithm to optimize the health data prediction model, build a health data sharing network with a consortium chain architecture by setting up network nodes, manage access rights through the smart contract engine, and realize the sharing of health data in the Internet of Things.

Images