Big data based patient respiratory failure risk assessment method

Abstract

The present application relates to the technical field of patient respiratory data analysis, and particularly relates to a patient respiratory failure risk assessment method based on big data. The present application continuously collects multi-dimensional physiological data of a patient, calculates data change factors of each index in a sliding time window, and screens physiological fluctuation data according to the data change factors. A synchronization lag factor is obtained based on the similarity between the fluctuation data and the sampling time difference, and an information correlation factor is obtained by combining the number of fluctuation data and the change characteristics of adjacent fluctuation data. The coupling degree between dimensions is obtained by fusing the two, and a coupling degree matrix is constructed. Then, the central descriptors of each dimension are calculated, and the dynamic assessment of respiratory failure risk is realized by analyzing the change trend of the central descriptors over time. The present application can monitor the patient's condition in real time, identify risks in a timely manner, and thus significantly reduce the delay of early warning and improve the prediction accuracy of acute respiratory failure.

Description

Technical Field

[0001] This invention relates to the technical field of patient respiratory data analysis, and specifically to a method for assessing the risk of respiratory failure in patients based on big data. Background Technology

[0002] Respiratory failure is a serious complication that can occur in the course of many diseases. If it is not identified and intervened in a timely manner, it can easily lead to organ failure. Therefore, establishing a risk assessment system helps to identify potential risks earlier in clinical practice and to implement relevant individualized interventions promptly. With the rapid development of modern medical data acquisition and monitoring technologies, assessing respiratory failure risk using patients' physiological indicators is both reasonable and feasible. Essentially, it objectively reflects the body's state through patients' physiological data, explores the coupling relationships between physiological indicators, and constructs early warning features to improve the timeliness of warnings. Data-driven decision-making provides a comprehensive solution for individualized respiratory monitoring and intervention systems.

[0003] When assessing the risk of respiratory failure, it is difficult to capture dynamic deterioration, especially since conditions such as acute respiratory distress syndrome can deteriorate rapidly in a short period of time. Static models cannot integrate the real-time trends reflected in the data, which may result in situations where alerts cannot be issued at critical moments. Summary of the Invention

[0004] To address the aforementioned technical problems, the present invention aims to provide a method for assessing the risk of respiratory failure in patients based on big data. The specific technical solution adopted is as follows:

[0005] A big data-based method for assessing the risk of respiratory failure in patients, the method comprising:

[0006] Collect physiological data of the patient for each preset dimension within each preset time window;

[0007] Based on the difference between each pair of adjacent physiological data within each preset time window for each preset dimension, a data change factor for each preset dimension within each preset time window is obtained. All physiological data for each preset dimension are then filtered based on the data change factor to obtain physiological fluctuation data for each dimension. Based on the similarity characteristics between physiological fluctuation data of any two preset dimensions within the same preset time window, and the difference in sampling times corresponding to the physiological fluctuation data, a synchronization hysteresis factor for any two preset dimensions within the same preset time window is obtained. Based on the number of physiological fluctuation data of any two preset dimensions within the same preset time window, and the data change characteristics between each pair of adjacent physiological fluctuation data in each preset dimension, an information correlation factor for any two preset dimensions within the same preset time window is obtained. Based on the synchronization hysteresis factor and the information correlation factor between any two dimensions, the coupling degree between any two dimensions is obtained. Based on the coupling degree between any two preset dimensions, a coupling degree matrix is ​​obtained.

[0008] Based on the element distribution in the coupling degree matrix, a central descriptor for each preset dimension is obtained in each preset time window; the risk of respiratory failure in patients is assessed based on the central descriptor for each preset time window.

[0009] Furthermore, the method for obtaining the data change factor includes:

[0010] Select any preset dimension and its physiological data within any preset time window as the physiological basis within the reference time window; calculate the numerical difference between every two adjacent physiological data within the reference time window, and sum all numerical differences to obtain the data change factor within the reference time window; traverse each preset dimension and each preset time window to obtain the data change factor of each preset dimension within each preset time window.

[0011] Furthermore, the method for acquiring the physiological fluctuation data includes:

[0012] The data change factors of each preset dimension in each preset time window are normalized by the maximum and minimum values ​​to obtain the standard change factor.

[0013] Based on the standard change factor of each preset dimension in each preset time window, according to 3 The principle is to obtain the multiplication factor of physiological data for each preset dimension in each preset time window;

[0014] Calculate the mean and standard deviation of physiological data for each preset dimension in each preset time window, and obtain the judgment threshold by combining the multiplication factor of each physiological data.

[0015] When the value of each physiological data point is greater than the judgment threshold, the physiological data point is considered to be physiological fluctuation data.

[0016] Furthermore, the method for obtaining the synchronization hysteresis factor includes:

[0017] The synchronization hysteresis factor is obtained according to the synchronization hysteresis factor calculation formula, which is shown below:

[0018]

[0019] In the formula, This represents the synchronization hysteresis factor between any two preset dimensions within the same preset time window. This represents the relationship between any two pre-defined dimensions of physiological fluctuation data. value; This indicates the sampling time corresponding to the first physiological fluctuation data within one of the two preset dimensions; This indicates the sampling time corresponding to the first physiological fluctuation data in one of the two preset dimensions; This indicates the sampling time corresponding to the last physiological fluctuation data in one of the two preset dimensions; This indicates the sampling time corresponding to the last physiological fluctuation data in one of the two preset dimensions; This represents the absolute value function.

[0020] Furthermore, the method for obtaining the information-related factors includes:

[0021] The information relevance factor is obtained according to the information relevance factor calculation formula, which is shown below:

[0022]

[0023] In the formula, This represents the information correlation factor between any two preset dimensions within the same preset time window. This indicates the number of physiological fluctuation data within one of the two preset dimensions; This indicates the number of physiological fluctuation data within one of the two preset dimensions; This represents the maximum number of physiological fluctuation data points between any two preset dimensions; This represents the minimum number of physiological fluctuation data points between any two preset dimensions; This indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; This indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; Indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; Indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; This represents the absolute value function.

[0024] Furthermore, the method for obtaining the degree of coupling includes:

[0025] The sum of the synchronization hysteresis factor and the information correlation factor between any two dimensions is taken as the degree of coupling between any two dimensions.

[0026] Furthermore, the method for obtaining the coupling degree matrix includes:

[0027] Choose any preset dimension as the reference dimension, calculate the coupling degree between the reference dimension and each other preset dimension, arrange the coupling degrees as a row vector of the coupling degree matrix; traverse the other preset dimensions to obtain the row vectors of the coupling degree matrices corresponding to the other dimensions; arrange the row vectors of all coupling degree matrices to obtain the coupling degree matrix.

[0028] Furthermore, the method for obtaining the central descriptor includes:

[0029] Choose any row in the coupling degree matrix as a reference row;

[0030] The central descriptor is obtained according to the central descriptor calculation formula, which is as follows:

[0031]

[0032] In the formula, This represents the center descriptor of the preset dimension corresponding to the reference row; This indicates the number of information-related factors in the reference row; Indicates the first reference row Individual information-related factors; This indicates the maximum value of the information relevance factor within the reference row.

[0033] Furthermore, the risk of respiratory failure in patients is assessed based on the central descriptor, including:

[0034] The central descriptor of each preset dimension is obtained in each preset time window, and arranged in chronological order to obtain the central descriptor sequence;

[0035] Perform the central descriptor sequence Decompose to obtain the trend terms of all central descriptors in the central descriptor sequence;

[0036] Starting with the central descriptor corresponding to the maximum value of each trend term and ending with the central descriptor corresponding to the next nearest minimum value of the next trend term, the number of central descriptors between the maximum and minimum values ​​of each trend term is counted as the number of affected periods.

[0037] When the number of affected periods for each preset dimension exceeds the preset first threshold, it is considered that relevant personnel need to intervene in the patient within that preset dimension.

[0038] Furthermore, the method for obtaining the judgment threshold includes:

[0039] Choose any physiological data point as a reference.

[0040] When the standard variation factor of the reference data is not greater than 0.3, the multiplier of the reference data is set to 1; when the standard variation factor of the reference data is greater than 0.3 and less than 0.9, the multiplier of the reference data is set to 2; when the standard variation factor of the reference data is not less than 0.9, the multiplier of the reference data is set to 3.

[0041] The product of the multiplication factor of the reference data and the standard deviation of all physiological data within the preset time window of the reference data is used as the first product; the sum of the first product and the average value of all physiological data within the preset time window of the reference data is used as the judgment threshold of the reference data.

[0042] The present invention has the following beneficial effects:

[0043] This invention collects physiological data of a patient for each preset dimension within each preset time window. Since acute respiratory failure manifests as significant changes in physiological data across different preset dimensions, a data change factor for each preset dimension within each preset time window is obtained based on the difference between any two adjacent physiological data points within that window. This data change factor is then used to filter out physiological data points showing the most significant changes within that time window. Because physiological data from different preset dimensions may interact—a sudden change in one preset dimension may lead to changes in another—the synchronicity and correlation between any two preset dimension physiological data points are analyzed. The invention obtains synchronization lag factors and information correlation factors; using these factors within the same preset time window, it describes the correlation between any two preset dimensions, thus obtaining the coupling degree between them; to reflect the steady-state relationship between physiological indicators of different preset dimensions, it obtains the central descriptor for each preset dimension in each preset time window based on the element distribution in the coupling degree matrix; and it assesses the risk of respiratory failure in patients based on the central descriptor for each preset time window. This invention enables real-time monitoring of patient conditions and timely risk identification, thereby significantly reducing warning delays and improving the prediction accuracy of acute respiratory failure. Attached Figure Description

[0044] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 The flowchart illustrates a method for assessing the risk of respiratory failure in patients based on big data, as provided in one embodiment of the present invention. Detailed Implementation

[0046] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a big data-based patient respiratory failure risk assessment method proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0048] The following description, in conjunction with the accompanying drawings, details a specific scheme for a patient respiratory failure risk assessment method based on big data provided by the present invention.

[0049] Please see Figure 1 This illustrates an embodiment of the present invention providing a method for assessing the risk of respiratory failure in patients based on big data, the method comprising:

[0050] Step S1: Collect physiological data of the patient for each preset dimension within each preset time window.

[0051] This invention is primarily applied to the monitoring and assessment of the risk of respiratory failure in patients. Therefore, it first acquires physiological data on different physiological indicators related to respiratory failure, i.e., physiological data across different preset dimensions. Since respiratory failure is related to the patient's respiratory, circulatory, and metabolic systems, blood oxygen saturation and respiratory rate are used as preset dimensions related to the respiratory system, heart rate and blood pressure as preset dimensions related to the circulatory system, and arterial blood lactate and arterial blood pH as preset dimensions related to the metabolic system.

[0052] In one embodiment of the present invention, the preset time window is set to 30 minutes, and the interval between any two adjacent preset time windows is 5 minutes. It should be noted that the preset time window can be set arbitrarily and is not limited here.

[0053] Since the physiological data sources for different preset dimensions are different, the sampling frequencies are also different. For example, the sampling frequency of heart rate is once per second, and the sampling frequency of blood oxygen saturation is once per minute. Therefore, the amount of data for different preset dimensions may be different within each preset time window. Therefore, in one embodiment of the present invention, the Newton interpolation method is used to fit a polynomial for preset dimensions with less data and the missing physiological data is obtained from the polynomial to complete the alignment of physiological data for different preset dimensions.

[0054] Step S2: Based on the difference between each pair of adjacent physiological data within each preset time window for each preset dimension, obtain the data change factor for each preset dimension within each preset time window; filter all physiological data for each preset dimension based on the data change factor to obtain physiological fluctuation data for each dimension; based on the similarity characteristics between physiological fluctuation data of any two preset dimensions within the same preset time window, and the difference in sampling time corresponding to the physiological fluctuation data, obtain the synchronization hysteresis factor for any two preset dimensions within the same preset time window; based on the number of physiological fluctuation data of any two preset dimensions within the same preset time window, and the data change characteristics between each pair of adjacent physiological fluctuation data in each preset dimension, obtain the information correlation factor for any two preset dimensions within the same preset time window; based on the synchronization hysteresis factor and information correlation factor between any two dimensions, obtain the coupling degree between any two dimensions; based on the coupling degree between any two preset dimensions, obtain the coupling degree matrix.

[0055] Because acute respiratory failure in patients manifests as significant variations in physiological data across different preset dimensions, this embodiment of the invention obtains a data variation factor for each preset dimension within each preset time window based on the difference between every two adjacent physiological data points within each preset time window. Then, using the data variation factor for each preset time window, physiological data points exhibiting more pronounced changes within that time window are selected.

[0056] Preferably, in one embodiment of the present invention, the method for obtaining the data change factor includes:

[0057] Select any preset dimension and its physiological data within any preset time window as the physiological basis within the reference time window; calculate the numerical difference between every two adjacent physiological data within the reference time window, and sum all numerical differences to obtain the data change factor within the reference time window, which reflects the overall change of physiological data within the reference time window; traverse each preset dimension and each preset time window to obtain the data change factor of each preset dimension within each preset time window.

[0058] Preferably, in one embodiment of the present invention, the method for acquiring physiological fluctuation data includes:

[0059] The data change factors of each preset dimension in each preset time window are normalized by the maximum and minimum values ​​to obtain the standard change factor.

[0060] Choose any physiological data point as a reference.

[0061] Since the distribution of physiological data conforms to a normal distribution, based on the standard variation factor of each preset dimension in each preset time window, according to 3 The principle is to obtain the multiplication factor of physiological data for each preset dimension in each preset time window. Specifically, since the standard variation factor is small, the physiological data values ​​within the preset time window will be concentrated near the mean. At this time, if there are physiological data with a small deviation, they can be considered as physiological fluctuation data. If the standard variation factor is large, it indicates that the overall fluctuation within the preset time window is more obvious. At this time, only physiological data with a large deviation can be considered as physiological fluctuation data. Accordingly, in one embodiment of the present invention, the following settings can be made: when the standard variation factor of the reference data is not greater than 0.3, the multiplication factor of the reference data is set to 1; when the standard variation factor of the reference data is greater than 0.3 and less than 0.9, the multiplication factor of the reference data is set to 2; when the standard variation factor of the reference data is not less than 0.9, the multiplication factor of the reference data is set to 3.

[0062] The product of the multiplication factor of the reference data and the standard deviation of all physiological data within the preset time window of the reference data is used as the first product; the sum of the first product and the average value of all physiological data within the preset time window of the reference data is used as the judgment threshold of the reference data.

[0063] When the value of each physiological data point is greater than the judgment threshold, the physiological data point is considered to be physiological fluctuation data.

[0064] Since physiological data from different preset dimensions of a patient may influence each other—that is, a sudden change in physiological data from one preset dimension may cause a change in physiological data from another preset dimension—in this embodiment of the invention, within the same preset time window, the similarity characteristics between physiological fluctuation data from any two preset dimensions and the difference in sampling time corresponding to the physiological fluctuation data are first used to analyze the synchronicity between physiological data from any two preset dimensions. Secondly, based on the number of physiological fluctuation data from any two preset dimensions and the data change characteristics between each pair of adjacent physiological fluctuation data in each preset dimension, information correlation factors between any two preset dimensions are obtained to analyze the correlation between any two preset dimensions.

[0065] Preferably, in one embodiment of the present invention, the method for obtaining the synchronization hysteresis factor includes:

[0066] The synchronization hysteresis factor is obtained according to the formula for calculating the synchronization hysteresis factor, which is shown below:

[0067]

[0068] In the formula, This represents the synchronization hysteresis factor between any two preset dimensions within the same preset time window. This represents the relationship between any two pre-defined dimensions of physiological fluctuation data. value; This indicates the sampling time corresponding to the first physiological fluctuation data within one of the two preset dimensions; This indicates the sampling time corresponding to the first physiological fluctuation data in one of the two preset dimensions; This indicates the sampling time corresponding to the last physiological fluctuation data in one of the two preset dimensions; This indicates the sampling time corresponding to the last physiological fluctuation data in one of the two preset dimensions; This represents the absolute value function.

[0069] In the formula for calculating the synchronization hysteresis factor, if the physiological data of two preset dimensions are within each preset time window... The smaller the value, the stronger the similarity between the two preset dimensions. In this case, the stronger the correlation between the two preset dimensions, the larger the synchronization lag factor. The smaller the difference between the earliest physiological fluctuation data corresponding to the sampling time of the two preset dimensions, and the smaller the difference between the latest physiological fluctuation data corresponding to the sampling time, the stronger the synchronization between the two preset dimensions in each preset time window. In this case, the larger the synchronization lag factor.

[0070] Preferably, in one embodiment of the present invention, the method for obtaining information relevance factors includes:

[0071] The information relevance factor is obtained according to the formula for calculating the information relevance factor, which is shown below:

[0072]

[0073] In the formula, This represents the information correlation factor between any two preset dimensions within the same preset time window. This indicates the number of physiological fluctuation data within one of the two preset dimensions; This indicates the number of physiological fluctuation data within one of the two preset dimensions; This represents the maximum number of physiological fluctuation data points between any two preset dimensions; This represents the minimum number of physiological fluctuation data points between any two preset dimensions; This indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; This indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; Indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; Indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; This represents the absolute value function.

[0074] In the formula for calculating the information relevance factor, within the same preset time window, the smaller the difference in physiological fluctuation data between two preset dimensions, the better. The smaller the value, the more obvious the overall correlation between the two preset dimensions, and the greater the information correlation factor between any two preset dimensions within the same preset time window. Using the minimum number of physiological fluctuation data points in the two preset dimensions as the standard, each pair of adjacent physiological fluctuation data points is analyzed, and the rate of change between each pair of adjacent physiological fluctuation data points within one preset dimension is determined. Rate of change between each two adjacent physiological fluctuation data points within another preset dimension The smaller the difference between the two, the more obvious the interaction between the two preset dimensions is, that is, the greater the information correlation factor between the two preset dimensions within the same preset time window.

[0075] By using the synchronization lag factor and information correlation factor of any two preset dimensions within the same preset time window, the correlation between any two preset dimensions within the same preset window can be described, thus obtaining the coupling degree between any two dimensions.

[0076] Preferably, in one embodiment of the present invention, the sum of the synchronization lag factor and the information correlation factor between any two dimensions is taken as the degree of coupling between any two dimensions.

[0077] Furthermore, in order to intuitively describe the degree of coupling between any two preset dimensions in different preset time windows, a coupling degree matrix is ​​obtained.

[0078] Preferably, in one embodiment of the present invention, the method for obtaining the coupling degree matrix includes:

[0079] Choose any preset dimension as the reference dimension, calculate the coupling degree between the reference dimension and each other preset dimension, arrange the coupling degrees as a row vector of the coupling degree matrix; traverse the other preset dimensions to obtain the row vectors of the coupling degree matrices corresponding to the other dimensions; arrange the row vectors of all coupling degree matrices to obtain the coupling degree matrix.

[0080] Step S3: Based on the element distribution in the coupling degree matrix, obtain the central descriptor for each preset dimension in each preset time window; assess the patient's risk of respiratory failure based on the central descriptor for each preset time window.

[0081] In the coupling degree matrix, each row can be viewed as a correlation row between a preset dimension and other preset dimensions. The larger the value of a certain element in each row, the closer the correlation between the preset dimension corresponding to that row and the other preset dimensions corresponding to that element. When a patient is at risk of respiratory failure, the homeostasis of the previously closely correlated indicators will be disrupted. This is because when entering a state of respiratory failure risk, the body's self-regulation mechanism may become unbalanced, thus affecting the physiological indicators of the preset dimensions reflecting homeostasis. Therefore, in this embodiment of the invention, in order to reflect the homeostatic relationship between physiological indicators of different preset dimensions, a central descriptor for each preset dimension in each preset time window is obtained based on the element distribution in the coupling degree matrix.

[0082] Preferably, in one embodiment of the present invention, the method for obtaining the central descriptor includes:

[0083] Choose any row in the coupling degree matrix as a reference row;

[0084] The central descriptor is obtained based on the central descriptor calculation formula, which is shown below:

[0085]

[0086] In the formula, This represents the center descriptor of the preset dimension corresponding to the reference row; This indicates the number of information-related factors in the reference row; Indicates the first reference row Individual information-related factors; This indicates the maximum value of the information relevance factor within the reference row.

[0087] In the central descriptor calculation formula, the physiological indicators of the preset dimension corresponding to each row are analyzed, each element in the row vector is normalized and summed, and the correlation between the physiological indicators of each preset dimension and the physiological indicators of other preset dimensions is described.

[0088] Preferably, in one embodiment of the present invention, assessing the risk of respiratory failure in a patient based on a central descriptor includes:

[0089] The central descriptor of each preset dimension is obtained in each preset time window, and arranged in chronological order to obtain the central descriptor sequence;

[0090] Perform the central descriptor sequence Decompose to obtain the trend terms of all central descriptors in the central descriptor sequence;

[0091] Starting with the central descriptor corresponding to the maximum value of each trend term and ending with the central descriptor corresponding to the next nearest minimum value of the next trend term, the number of central descriptors between the maximum and minimum values ​​of each trend term is counted as the number of affected periods.

[0092] When the number of affected periods for each preset dimension exceeds the preset first threshold, it is considered that relevant personnel need to intervene in the patient within that preset dimension.

[0093] In one embodiment of the present invention, the preset first threshold is set to 30%. It should be noted that the preset first threshold can be set by itself and is not limited here.

[0094] In summary, physiological data for each preset dimension of the patient is collected within each preset time window. Based on the differences between any two adjacent physiological data points within each preset time window, a data variation factor for each preset dimension is obtained. All physiological data for each preset dimension are then filtered based on the data variation factor to obtain physiological fluctuation data for each dimension. Based on the similarity characteristics between physiological fluctuation data points of any two preset dimensions within the same preset time window, and the differences in sampling times corresponding to the physiological fluctuation data, a synchronization lag factor for any two preset dimensions within the same preset time window is obtained. Based on the number of physiological fluctuation data points of any two preset dimensions within the same preset time window, and the data variation characteristics between any two adjacent physiological fluctuation data points within each preset dimension, an information correlation factor for any two preset dimensions within the same preset time window is obtained. Based on the synchronization lag factor and information correlation factor between any two dimensions, the coupling degree between any two dimensions is obtained. Based on the coupling degree between any two preset dimensions, a coupling degree matrix is ​​obtained. Based on the element distribution in the coupling degree matrix, a central descriptor for each preset dimension within each preset time window is obtained. Finally, the risk of respiratory failure in the patient is assessed based on the central descriptor for each preset time window.

[0095] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0096] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for assessing the risk of respiratory failure in patients based on big data, characterized in that, The method includes: Collect physiological data of the patient for each preset dimension within each preset time window; Based on the difference between each pair of adjacent physiological data within each preset time window for each preset dimension, a data change factor for each preset dimension within each preset time window is obtained. All physiological data for each preset dimension are then filtered based on the data change factor to obtain physiological fluctuation data for each dimension. Based on the similarity characteristics between physiological fluctuation data of any two preset dimensions within the same preset time window, and the difference in sampling times corresponding to the physiological fluctuation data, a synchronization hysteresis factor for any two preset dimensions within the same preset time window is obtained. Based on the number of physiological fluctuation data of any two preset dimensions within the same preset time window, and the data change characteristics between each pair of adjacent physiological fluctuation data in each preset dimension, an information correlation factor for any two preset dimensions within the same preset time window is obtained. Based on the synchronization hysteresis factor and the information correlation factor between any two dimensions, the coupling degree between any two dimensions is obtained. Based on the coupling degree between any two preset dimensions, a coupling degree matrix is ​​obtained. Based on the element distribution in the coupling degree matrix, a central descriptor for each preset dimension is obtained in each preset time window; the risk of respiratory failure in patients is assessed based on the central descriptor for each preset time window.

2. The method for assessing the risk of respiratory failure in patients based on big data according to claim 1, characterized in that, The method for obtaining the data change factor includes: Select any preset dimension and its physiological data within any preset time window as the physiological basis within the reference time window; calculate the numerical difference between every two adjacent physiological data within the reference time window, and sum all numerical differences to obtain the data change factor within the reference time window; traverse each preset dimension and each preset time window to obtain the data change factor of each preset dimension within each preset time window.

3. The method for assessing the risk of respiratory failure in patients based on big data according to claim 1, characterized in that, The methods for acquiring the physiological fluctuation data include: The data change factors of each preset dimension in each preset time window are normalized by the maximum and minimum values ​​to obtain the standard change factor. Based on the standard change factor of each preset dimension in each preset time window, according to 3 The principle is to obtain the multiplication factor of physiological data for each preset dimension in each preset time window; Calculate the mean and standard deviation of physiological data for each preset dimension in each preset time window, and obtain the judgment threshold by combining the multiplication factor of each physiological data. When the value of each physiological data point is greater than the judgment threshold, the physiological data point is considered to be physiological fluctuation data.

4. The method for assessing the risk of respiratory failure in patients based on big data according to claim 1, characterized in that, The method for obtaining the synchronization hysteresis factor includes: The synchronization hysteresis factor is obtained according to the synchronization hysteresis factor calculation formula, which is shown below: In the formula, This represents the synchronization hysteresis factor between any two preset dimensions within the same preset time window. This represents the relationship between any two pre-defined dimensions of physiological fluctuation data. value; This indicates the sampling time corresponding to the first physiological fluctuation data within one of the two preset dimensions; This indicates the sampling time corresponding to the first physiological fluctuation data in one of the two preset dimensions; This indicates the sampling time corresponding to the last physiological fluctuation data in one of the two preset dimensions; This indicates the sampling time corresponding to the last physiological fluctuation data in one of the two preset dimensions; This represents the absolute value function.

5. The method for assessing the risk of respiratory failure in patients based on big data according to claim 1, characterized in that, The method for obtaining the information-related factors includes: The information relevance factor is obtained according to the information relevance factor calculation formula, which is shown below: In the formula, This represents the information correlation factor between any two preset dimensions within the same preset time window. This indicates the number of physiological fluctuation data within one of the two preset dimensions; This indicates the number of physiological fluctuation data within one of the two preset dimensions; This represents the maximum number of physiological fluctuation data points between any two preset dimensions; This represents the minimum number of physiological fluctuation data points between any two preset dimensions; This indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; This indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; Indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; Indicates the first element within one of two preset dimensions. Data values ​​of individual physiological fluctuations; This represents the absolute value function.

6. The method for assessing the risk of respiratory failure in patients based on big data according to claim 1, characterized in that, The method for obtaining the degree of coupling includes: The sum of the synchronization hysteresis factor and the information correlation factor between any two dimensions is taken as the degree of coupling between any two dimensions.

7. The method for assessing the risk of respiratory failure in patients based on big data according to claim 1, characterized in that, The method for obtaining the coupling degree matrix includes: Choose any preset dimension as the reference dimension, calculate the coupling degree between the reference dimension and each other preset dimension, arrange the coupling degrees as a row vector of the coupling degree matrix; traverse the other preset dimensions to obtain the row vectors of the coupling degree matrices corresponding to the other dimensions; arrange the row vectors of all coupling degree matrices to obtain the coupling degree matrix.

8. The method for assessing the risk of respiratory failure in patients based on big data according to claim 1, characterized in that, The method for obtaining the central descriptor includes: Choose any row in the coupling degree matrix as a reference row; The central descriptor is obtained according to the central descriptor calculation formula, which is as follows: In the formula, This represents the center descriptor of the preset dimension corresponding to the reference row; This indicates the number of information-related factors in the reference row; Indicates the first reference row Individual information-related factors; This indicates the maximum value of the information relevance factor within the reference row.

9. The method for assessing the risk of respiratory failure in patients based on big data according to claim 1, characterized in that, The risk of respiratory failure in patients is assessed based on the central descriptor, including: The central descriptor of each preset dimension is obtained in each preset time window, and arranged in chronological order to obtain the central descriptor sequence; Perform the central descriptor sequence Decompose to obtain the trend terms of all central descriptors in the central descriptor sequence; Starting with the central descriptor corresponding to the maximum value of each trend term and ending with the central descriptor corresponding to the next nearest minimum value of the next trend term, the number of central descriptors between the maximum and minimum values ​​of each trend term is counted as the number of affected periods. When the number of affected periods for each preset dimension exceeds the preset first threshold, it is considered that relevant personnel need to intervene in the patient within that preset dimension.

10. The method for assessing the risk of respiratory failure in patients based on big data according to claim 3, characterized in that, The method for obtaining the judgment threshold includes: Choose any physiological data point as a reference. When the standard variation factor of the reference data is not greater than 0.3, the multiplier of the reference data is set to 1; when the standard variation factor of the reference data is greater than 0.3 and less than 0.9, the multiplier of the reference data is set to 2; when the standard variation factor of the reference data is not less than 0.9, the multiplier of the reference data is set to 3. The product of the multiplication factor of the reference data and the standard deviation of all physiological data within the preset time window of the reference data is used as the first product; the sum of the first product and the average value of all physiological data within the preset time window of the reference data is used as the judgment threshold of the reference data.

Images