Patient physiological data monitoring method and system based on Internet of Things

Abstract

The invention relates to the technical field of physiological data processing, in particular to a patient physiological data monitoring method and system based on the Internet of Things. The method comprises the steps that multiple pieces of historical physiological data and physiological data of a target patient are acquired, all the historical physiological data are clustered through mean shift clustering, and multiple clusters are obtained; and calculating the classification possibility of the physiological data belonging to each cluster, and selecting the cluster corresponding to the maximum value in the normalized values of the classification possibility as the category of the physiological data so as to complete the classification of the physiological data. According to the scheme, classified monitoring of the physiological data, collected in real time, of the target patient can be timely and accurately carried out.

Description

Technical Field

[0001] This invention relates to the field of physiological data processing technology. More specifically, this invention relates to a method and system for monitoring patient physiological data based on the Internet of Things (IoT). Background Technology

[0002] With the rapid development of medical informatization, the scale and complexity of physiological data have increased significantly, giving rise to Internet of Medical Things (IoMT) technology. It supports the real-time collection of patient vital signs and integrates data through cloud platforms or local servers to facilitate precision medicine and personalized treatment.

[0003] The application of the Internet of Things (IoT) in patient physiological data monitoring includes smart wards, telemedicine, and intelligent nurse station management. The IoT connects medical terminals such as smart beds, wearable devices, and monitors to hospital information systems through sensors, communication protocols (such as Wi-Fi 6, Bluetooth, and ZigBee), and cloud platforms, enabling data collection, transmission, and integration. In other words, through the IoT, medical institutions can centrally store and manage physiological data scattered across different systems, rapidly process large-scale data using distributed computing technology, and automatically classify and statistically analyze the data using machine learning algorithms. This not only significantly improves data processing efficiency but also provides strong support for disease prediction, personalized treatment, and epidemiological research.

[0004] In the IoT-based physiological data analysis process, the classification of physiological data is particularly crucial. The proper implementation of this key step not only effectively identifies potential patterns in the data but also lays a solid foundation for subsequent statistical analysis of patient physiological data. Specifically, the main goal of classification is to group cases with similar characteristics into the same category based on multiple variables such as patient physiological data and diagnostic results. This achieves structured organization of large-scale physiological data, facilitating more convenient monitoring. This process helps reveal hidden patterns in the data, providing strong support for medical decision-making and a clear data framework for subsequent statistical analysis.

[0005] In existing technologies, the mean shift algorithm is usually used to classify data. This algorithm is based on the kernel density function to cluster given data. However, after the initial clustering is completed, it is difficult to efficiently cluster subsequent real-time updated data. As a result, the classification of physiological data cannot take into account both real-time performance and accuracy, thereby reducing the reliability of monitoring patients' physiological data.

[0006] Therefore, it is particularly important to be able to classify physiological data efficiently and accurately in order to achieve accurate monitoring of patients' physiological data. Summary of the Invention

[0007] The purpose of this invention is to propose a method and system for monitoring patient physiological data based on the Internet of Things, in order to solve the problem that the classification and statistics of physiological data in the prior art cannot take into account both real-time performance and accuracy, thus affecting the accuracy of physiological data monitoring; to this end, this invention provides solutions in the following two aspects.

[0008] In the first aspect, the present invention provides a method for monitoring patient physiological data based on the Internet of Things, comprising: The system acquires multiple historical physiological data points and real-time physiological data of the target patient. Both the historical and physiological data include multiple data points, which at least include the patient's body temperature, blood pressure, heart rate, and respiratory rate. Mean-shift clustering was used to cluster all historical physiological data, resulting in multiple clusters; The probability of the physiological data belonging to each cluster is calculated. This probability is positively correlated with the reciprocal of the number of neighboring data points of the physiological data, the ratio of trend consistency when any neighboring data point belongs to any cluster to the distance between the physiological data point and that neighboring data point, and a minimum value. The minimum value of 1 is the smaller of the relative dispersion of the physiological data point relative to any cluster. The neighboring data points are historical physiological data whose distance from the physiological data point is less than a set neighborhood radius. The trend consistency represents the iterative change from any historical physiological data point as the initial cluster center to the final cluster center after multiple iterations. The cluster corresponding to the maximum value among the normalized values ​​of classification probability is selected as the category of physiological data to complete the classification of physiological data and realize the monitoring of physiological data of target patients.

[0009] The above scheme determines the intrinsic attributes (trend consistency) of the neighboring data of the real-time collected physiological data based on the iterative changes of a large amount of historical physiological data during the clustering process. It then combines this with the distance between the physiological data and the clusters containing its neighboring data to determine the probability of the physiological data belonging to each cluster, thereby achieving the classification of the real-time collected physiological data. In other words, this invention, when classifying real-time collected physiological data, does not require repeated use of clustering algorithms. Instead, it analyzes the attributes and distances between the real-time collected physiological data and the previously classified data's similar neighboring data, thus achieving real-time and efficient classification of physiological data and improving the efficiency of subsequent clinical decision-making by physicians.

[0010] Optionally, the classification possibility for: In the formula, Representing physiological data Compared to the first The relative dispersion of each cluster For physiological data The number of neighboring data, The neighborhood data belongs to the first The number of data points in each cluster For the first The th cluster Consistency of trends in neighborhood data Representing physiological data With the The second distance of the neighboring data, the relative dispersion is the minimum of the ratio and 1, the ratio is the ... first The average distance between all historical physiological data in each cluster and its cluster center and the first cluster... Cluster center and physiological data of individual clusters The ratio of the first distance.

[0011] Optionally, the trend consistency for: ; in, This represents the total number of data items in the historical physiological data. This represents the total number of iterations during clustering. The factorial symbol, It is an exponential function with the natural constant e as its base. The trend ratio of the j-th data point in the k-th iteration is the minimum of 1 and the trend degree, where the trend degree is the value of the initial center after the k-th iteration. Second and third The ratio of the difference between the j-th data point of the final cluster center after K iterations and the ideal degree of change, where the ideal degree of change is the ratio of the difference between the j-th data point of the final cluster center after K iterations and the j-th data point of the initial center, to the total number of iterations during clustering. The initial center is the j-th data point before the iterations begin. Historical physiological data.

[0012] The above scheme provides a method for accurately calculating trend consistency.

[0013] Optionally, the normalized value of the classification probability is: the sum of the classification probabilities of physiological data belonging to each cluster is obtained, and the ratio of the classification probability of physiological data belonging to any cluster to the sum of classification probabilities is used as the normalized value of the corresponding classification probability.

[0014] The above scheme simplifies the calculation by normalizing the classification probability.

[0015] Optionally, the set neighborhood radius is: Obtain the Euclidean distance between any historical physiological data and all remaining historical physiological data. Sort all Euclidean distances in ascending order, select the first H Euclidean distances, and record the sum of the first H Euclidean distances as the neighborhood distance of any historical physiological data. Use the average neighborhood distance of all historical physiological data as the set neighborhood radius, where H is greater than 2.

[0016] The above method of obtaining a neighborhood radius by using historical physiological data provides data support for obtaining neighborhood data of subsequent physiological data.

[0017] Optionally, the first distance is the Euclidean distance between the cluster center of any cluster and the physiological data, and the second distance is the Euclidean distance between the physiological data and each neighboring data in any cluster.

[0018] Optionally, after acquiring multiple historical physiological data and the physiological data of the target patient, a step of standardizing the physiological data and each historical physiological data is also included.

[0019] Optionally, the method further includes using the maximum value as the trend consistency of the corresponding physiological data for the classification of physiological data at the next time step.

[0020] By using the maximum value as the trend consistency of the corresponding physiological data, the above method eliminates the need to cluster all historical physiological data, thus improving the classification efficiency of real-time collected physiological data.

[0021] Optionally, it also includes transmitting the classification results to the monitoring platform in real time after acquiring the classification of physiological data for subsequent reference by doctors.

[0022] In a second aspect, the present invention provides a patient physiological data monitoring system based on the Internet of Things, comprising: processor; The memory stores computer instructions for monitoring patient physiological data based on the Internet of Things (IoT), which, when executed by the processor, cause the system to perform an IoT-based patient physiological data monitoring method as described above.

[0023] The beneficial effects of this invention are as follows: The solution of this invention needs to fully consider the characteristics of physiological data when implementing classification analysis, including the multidimensionality, heterogeneity and potential missing values ​​of the data, so as to group cases with similar characteristics into the same category, realize the structured organization of large-scale physiological data for the monitoring platform to refer to, and provide data support for subsequent case analysis or decision-making on diagnosis and treatment information. Attached Figure Description

[0024] Figure 1The flowchart illustrating the steps of a patient physiological data monitoring method based on the Internet of Things in this embodiment is shown in the illustration. Figure 2 The diagram illustrates the structure of a patient physiological data monitoring system based on the Internet of Things in this embodiment. Detailed Implementation

[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0026] This invention addresses the monitoring of physiological data of a target patient in real time.

[0027] Specifically, such as Figure 1 As shown in this embodiment, a method for monitoring patient physiological data based on the Internet of Things includes the following steps: Step S1: Collect physiological data of the target patient in real time.

[0028] Specifically, the physiological data in this embodiment is the vital sign monitoring data of the target patient, which includes multiple data of the target patient, such as the patient's body temperature, blood pressure, heart rate, respiratory rate, systolic blood pressure, tensile blood pressure, and blood oxygen saturation, etc.

[0029] The physiological data mentioned above were collected using non-contact sensors on the smart bed. Since smart beds are existing technology, they will not be discussed further here.

[0030] The above data were collected using discrete sampling. When collecting physiological data from the target patient, the sampling interval was 5 or 10 minutes. Of course, it can also be determined according to the actual situation.

[0031] In this embodiment, the collected physiological data is also preprocessed, such as through standardization.

[0032] Step S2: Calculate the classification probability of the physiological data.

[0033] The process of obtaining classification probabilities includes steps S21-S24, as follows: S21, acquire multiple historical physiological data.

[0034] In this embodiment, historical physiological data of the target patient is acquired through an Internet of Things (IoT) device. Specifically, the sampling interval for the historical physiological data is 5 or 10 minutes, collecting physiological data from the patient throughout the day to construct a historical physiological dataset.

[0035] Of course, if the amount of historical physiological data of the target patient is small, the historical physiological data of multiple different patients can be obtained directly through big data technology to ensure that a large amount of historical physiological data can be obtained, thus avoiding the problem of inaccurate classification of physiological data due to insufficient sample data, which in turn leads to inaccurate data monitoring of patients.

[0036] Among them, the disease conditions of multiple different patients can be the same or different; the disease conditions of multiple different patients can also be the same or different from those of the target patient.

[0037] In this embodiment, the first The j-th data item collected in the second time is denoted as... The total number of historical physiological data collected is denoted as The number of items in the historical physiological data is recorded as .

[0038] It should be noted that the historical physiological data was preprocessed before being used, such as data denoising and standardization.

[0039] S22, cluster all historical physiological data to obtain multiple clusters and the trend consistency of historical physiological data in each cluster.

[0040] In this case, considering that the mean-shift algorithm is designed for iterative clustering of fixed data, while physiological data is often updated in real time and requires real-time clustering, directly applying mean-shift clustering to real-time updated physiological data would require a large amount of data for clustering, resulting in low efficiency. Therefore, in this embodiment, the probability of real-time collected physiological data belonging to each cluster is determined only by analyzing the clustering results of historical physiological data, in order to cluster the real-time updated physiological data and improve the classification efficiency of real-time collected physiological data.

[0041] The Mean Shift Algorithm is a non-parametric clustering algorithm commonly used for tasks such as image segmentation, object tracking, and density estimation. Based on the principle of density estimation, this algorithm iteratively updates the positions of data points, causing them to move towards areas of higher density and eventually cluster together. The core idea of ​​the Mean Shift Algorithm is to calculate the density gradient direction around each data point in its feature space and move the data points in the direction of increasing density along the gradient until they converge to a region of higher density.

[0042] In this embodiment, the mean shift algorithm is used to cluster all historical physiological data to obtain multiple clusters. Since the mean shift algorithm is existing technology, the specific steps will not be described in detail here.

[0043] After acquiring multiple clusters, these clusters are also labeled. For example, the severity level of a cluster can be classified according to the physiological data within the cluster, or the cluster can be labeled according to the diagnostic results corresponding to the physiological data within the cluster.

[0044] In this embodiment, obtaining multiple clusters and their corresponding annotation results provides a data reference for the clustering of physiological data collected in real time.

[0045] During the clustering process, the number of clustering iterations is denoted as... , with the first Using historical physiological data as the initial center, the j-th data point before the iteration begins is denoted as... The initial center is at the 1st The j-th data point of the center after the nth iteration is denoted as... Let the final cluster centers after K iterations from the initial centers be denoted as . .

[0046] In this embodiment, clustering requires traversing all historical physiological data, i.e., iteratively clustering each historical physiological data point as an initial center. It should be noted that during clustering, a historical physiological data point can be treated as a single data point, and the position of the data point is determined by multiple corresponding data items, with different data items corresponding to the data point's position in different dimensions.

[0047] In this embodiment, the trend consistency of each historical physiological data is analyzed based on the iterative update position of all historical physiological data in the clustering process. If the historical physiological data always points to its final cluster center during the iteration process (regarding the pointing issue here, this embodiment considers that a historical physiological data is actually a data point in a multi-dimensional space, and the data point can be represented by a vector, so the difference of each component of the vector represents the pointing in the N-dimensional space), then the historical physiological data is considered to have a high trend consistency. Then, based on the classification of other historical physiological data around the real-time collected physiological data, the classification probability of the physiological data is obtained to achieve real-time and accurate classification of physiological data.

[0048] The consistency of trends among the historical physiological data is as follows: ; in, This represents the total number of data items in the historical physiological data. This represents the total number of iterations during clustering. The factorial symbol, It is an exponential function with the natural constant e as its base. This represents the trend ratio of the j-th data item in the k-th iteration.

[0049] The process of obtaining the trend ratio is as follows: Use any historical physiological data point before the iteration begins as the initial center; Calculate the initial center after the first Second and third The difference between the j-th data item at the center after the next iteration; The ideal degree of change is defined as the ratio of the difference between the j-th data point in the final cluster center after K iterations from the initial center and the j-th data point in the initial center to the total number of iterations K during clustering. Passing the initial center through the first Second and third The ratio of the difference between the j-th data point at the center after the next iteration and the ideal degree of change is used as the degree of trend. The minimum value of the trend degree compared to 1 is selected as the trend ratio.

[0050] Specifically, the trend ratio is: ; and The initial center after the first Second and third The j-th data point of the center after the nth iteration Let j be the j-th data point representing the final cluster center after K iterations from the initial centers. Let j be the j-th data item in the initial center, where the initial center is the j-th data item before the iteration begins. There are 10 historical physiological data points, and min() is used to take the minimum value.

[0051] It should be noted that in the formula For the first The degree of change of historical physiological data (initial center) between the k-th and k-1-th iterations; the ratio of the degree of change to the ideal degree of change, i.e., the trend degree, the larger the ratio, the more the change of the j-th data in the k-th iteration conforms to the trend of moving towards its final cluster center; conversely, the smaller the ratio, the less it conforms to the trend of uniquely moving towards the final cluster center. If the ratio is greater than 1, it means that the change of the j-th data in the k-th iteration has exceeded the ideal state, and the excess part is meaningless in the calculation of the degree of change (here we only consider the trend inconsistency; for a certain displacement that particularly conforms to the displacement towards the final cluster center, the value is directly taken as 1). Therefore, the minimum value of the ratio and 1 is taken in the trend consistency formula; then through... The function normalizes the minimum value to the range of 0-1.

[0052] The aforementioned trend consistency combines the distance differences of each data point to represent the trend of the corresponding historical physiological data moving towards the cluster center. It should be noted that for any given historical physiological data point, if its distance from its final cluster center decreases during the iteration process, it indicates that the historical physiological data point is consistently moving towards its corresponding final cluster center during the iteration process, demonstrating a high degree of trend consistency.

[0053] S23, Obtain neighborhood data of physiological data, determine the trend consistency of neighborhood data belonging to each cluster and the corresponding number of data.

[0054] It should be noted that for real-time collected physiological data, the classification result of the real-time collected physiological data can be estimated based on the distance of the data to other historical physiological data in the surrounding area.

[0055] Specifically, the Euclidean distance between each historical physiological data point and the real-time physiological data is obtained. When the Euclidean distance is less than the set neighborhood radius, the corresponding historical physiological data is considered to be the neighborhood data of the real-time physiological data.

[0056] The process of obtaining the neighborhood radius is as follows: obtain the Euclidean distance between any historical physiological data and all remaining historical physiological data, sort the Euclidean distances in ascending order, select the Euclidean distances of the first H historical physiological data, and record the sum of the Euclidean distances of the first H historical physiological data as the neighborhood distance. The average of the neighborhood distances of all historical physiological data is used as the neighborhood radius, where H is greater than 2.

[0057] The value of H can be 10, and the neighborhood radius can be set according to the actual implementation.

[0058] By acquiring neighborhood data of the physiological data, the classification of the physiological data collected in real time can be determined.

[0059] S24, based on the trend consistency of neighborhood data, calculate the probability of physiological data belonging to each cluster.

[0060] Among them, classification probability for: ; In the formula, Representing physiological data Compared to the first The relative dispersion of each cluster For physiological data The number of neighboring data, The neighborhood data belongs to the first The number of data points in each cluster For the first The th cluster Consistency of trends in neighborhood data Representing physiological data With the The second distance of the neighborhood data.

[0061] The relative dispersion is the minimum of the ratio and 1, where the ratio is the th. The average distance between all historical physiological data in each cluster and its cluster center and the first cluster... Cluster center and physiological data of individual clusters The ratio of the first distance is used to determine the physiological data. When this ratio is greater than or equal to 1, it indicates that the physiological data is closer to the cluster center than the average distance within the cluster or that the physiological data is consistent with the average distribution within the cluster. When the ratio is less than 1, it indicates that the physiological data is far from the cluster center and may be an edge point or an outlier.

[0062] The first distance mentioned above is the Euclidean distance between the calculated cluster center and the real-time collected physiological data. The second distance is the Euclidean distance between the physiological data and each neighboring data in any cluster; the average distance is the mean of the Euclidean distances from each historical physiological data in each cluster to the cluster center.

[0063] Before calculating the Euclidean distance, the data in each historical physiological dataset are standardized, and then the Euclidean distance is calculated. Since the calculation of the Euclidean distance is a current technique, it will not be elaborated upon here.

[0064] Step S3: Select the cluster corresponding to the maximum value among the normalized values ​​of classification probability as the category of physiological data to complete the classification of physiological data.

[0065] In this embodiment, after obtaining the classification probability of physiological data belonging to each cluster, the classification probability is also normalized.

[0066] In one embodiment, the normalization process can employ Min-Max Normalization (Min-Max Normalization, deviation standardization) or Z-score normalization.

[0067] In another embodiment, the normalization process specifically involves: obtaining the sum of the classification probabilities of physiological data belonging to each cluster, and using the ratio of the classification probability of physiological data belonging to any cluster to the sum of classification probabilities as the normalized value of the corresponding classification probability.

[0068] In this embodiment, the cluster corresponding to the maximum value among the normalized values ​​of classification probability is selected as the category of the physiological data collected in real time.

[0069] After classifying the real-time collected physiological data, the physiological data can be stored as historical physiological data. At this time, the maximum value corresponding to the physiological data is recorded as the trend consistency of the physiological data, which provides data support for the classification of physiological data acquired in the next moment.

[0070] The reason for the above normalization is that the sum of the directly obtained classification possibilities may not be one, so it needs to be adjusted to simplify the calculation.

[0071] In this embodiment, the classification analysis of the real-time collected physiological data is performed based on all acquired historical physiological data. This allows the physiological data of the target patient to be classified with similar historical physiological data, providing data reference for doctors to make subsequent diagnoses of the target patient.

[0072] For example, by classifying real-time collected physiological data and transmitting the classification results to a monitoring platform in real time, audible and visual alarms or push notifications can be triggered. This allows doctors to obtain more accurate information in a timely manner, enabling real-time monitoring of patient information. For instance, by classifying patients' physiological data according to their severity, alarm information can be automatically triggered at the monitoring terminal of the smart nurse station. Doctors can then promptly access relevant patient diagnosis and treatment information, improving the efficiency of clinical decision-making. Simultaneously, this can provide more efficient data support for medical research and offer rich data for subsequent case analysis.

[0073] The solution of this invention can effectively and timely classify real-time collected physiological data, enabling doctors to obtain patient diagnosis and treatment information more efficiently and improving the efficiency of subsequent clinical decision-making.

[0074] This invention also provides a system for evaluating the operating status of a current transformer. For example... Figure 2 As shown, the system includes a processor and a memory, the memory storing computer program instructions, which, when executed by the processor, implement the current transformer operating status evaluation method according to the present invention.

[0075] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and therefore will not be described in detail here.

[0076] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented by computer-readable / executable instructions stored or otherwise maintained on such a computer-readable medium.

[0077] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A method for monitoring patient physiological data based on the Internet of Things, characterized in that, include: The system acquires multiple historical physiological data and real-time physiological data of the target patient. The historical physiological data and the physiological data both include multiple data, including at least the patient's body temperature, blood pressure, heart rate, and respiratory rate. Mean-shift clustering was used to cluster all historical physiological data, resulting in multiple clusters; Calculate the probability that the physiological data belongs to each cluster. The classification probability is positively correlated with the reciprocal of the number of neighboring data of the physiological data, the ratio of the trend consistency when any neighboring data belongs to any cluster to the distance between the physiological data and any neighboring data, and the minimum value; the minimum value of 1 is the smaller of the relative dispersion of the physiological data relative to any cluster; the neighboring data is historical physiological data whose distance from the physiological data is less than a set neighborhood radius; the trend consistency characterizes the iterative change from any historical physiological data as the initial cluster center to the final cluster center after multiple iterations; The cluster corresponding to the maximum value among the normalized values ​​of classification probability is selected as the category of physiological data to complete the classification of physiological data and realize the monitoring of physiological data of target patients.

2. The method for monitoring patient physiological data based on the Internet of Things according to claim 1, characterized in that, The classification possibility for: In the formula, Representing physiological data Compared to the first The relative dispersion of each cluster For physiological data The number of neighboring data, The neighborhood data belongs to the first The number of data points in each cluster For the first The th cluster Consistency of trends in neighborhood data Representing physiological data With the The second distance of the neighboring data, the relative dispersion is the minimum of the ratio and 1, the ratio is the ... first The average distance between all historical physiological data in each cluster and its cluster center and the first cluster... Cluster center and physiological data of individual clusters The ratio of the first distance.

3. The method for monitoring patient physiological data based on the Internet of Things according to claim 2, characterized in that, The trend consistency for: ； in, This represents the total number of data items in the historical physiological data. This represents the total number of iterations during clustering. The factorial symbol, It is an exponential function with the natural constant e as its base. The trend ratio of the j-th data point in the k-th iteration is the minimum of 1 and the trend degree, where the trend degree is the value of the initial center after the k-th iteration. Second and third The ratio of the difference between the j-th data point of the final cluster center after K iterations and the ideal degree of change, where the ideal degree of change is the ratio of the difference between the j-th data point of the final cluster center after K iterations and the j-th data point of the initial center, to the total number of iterations during clustering. The initial center is the j-th data point before the iterations begin. Historical physiological data.

4. The method for monitoring patient physiological data based on the Internet of Things according to claim 1, characterized in that, The normalized value of the classification probability is: the sum of the classification probabilities of physiological data belonging to each cluster, and the ratio of the classification probability of physiological data belonging to any cluster to the sum of classification probabilities is used as the normalized value of the corresponding classification probability.

5. The method for monitoring patient physiological data based on the Internet of Things according to claim 1, characterized in that, The defined neighborhood radius is: Obtain the Euclidean distance between any historical physiological data and all remaining historical physiological data. Sort all Euclidean distances in ascending order, select the first H Euclidean distances, and record the sum of the first H Euclidean distances as the neighborhood distance of any historical physiological data. Use the average neighborhood distance of all historical physiological data as the set neighborhood radius, where H is greater than 2.

6. The method for monitoring patient physiological data based on the Internet of Things according to claim 1, characterized in that, The first distance is the Euclidean distance between the cluster center of any cluster and the physiological data, and the second distance is the Euclidean distance between the physiological data and each neighboring data in any cluster.

7. The method for monitoring patient physiological data based on the Internet of Things according to claim 6, characterized in that, After acquiring multiple historical physiological data and the physiological data of the target patient, the physiological data and each historical physiological data are further standardized.

8. The method for monitoring patient physiological data based on the Internet of Things according to claim 1, characterized in that, It also includes using the maximum value as the trend consistency of the corresponding physiological data for the classification of physiological data at the next moment.

9. A method for monitoring patient physiological data based on the Internet of Things according to claim 1, characterized in that, It also includes transmitting the classification results to the monitoring platform in real time after acquiring and classifying the physiological data, so that doctors can refer to them later.

10. A patient physiological data monitoring system based on the Internet of Things, characterized in that, include: processor; A memory storing computer instructions for monitoring patient physiological data based on the Internet of Things (IoT), which, when executed by the processor, cause the system to perform an IoT-based method for monitoring patient physiological data according to any one of claims 1-9.

Images