Acute ischemic stroke prediction system and prediction method

Abstract

The invention discloses an acute ischemic stroke prediction system and method, and the system comprises a data collection unit, a data processing unit, a model training unit, a detection unit and a prevention unit, and the data collection unit is used for obtaining basic physiological parameters; the data preprocessing unit preprocesses the basic physiological parameters to obtain a training data set, and the model training unit carries out training based on the training data set to generate a cerebral apoplexy prediction model; the detection unit is used for acquiring health data information of a to-be-detected individual, and the prediction unit is used for outputting the probability that the to-be-detected individual suffers from cerebral apoplexy according to the health data information and the cerebral apoplexy prediction model; according to the method, the objective measurement relation is established between the basic physiological parameters and the cerebral apoplexy disease probability, so that the judgment result is quickly output for detecting the physiological parameters of the to-be-detected person, early-stage screening is carried out on high-risk groups of cerebral apoplexy, targeted prevention and care can be conveniently carried out on the high-risk groups, and the incidence rate of cerebral apoplexy is reduced.

Description

Technical Field

[0001] The present application relates to the field of biomedicine technology, and in particular to an acute ischemic stroke prediction system and prediction method. Background Art

[0002] Stroke, also known as "stroke" or "cerebrovascular accident", is an acute cerebrovascular disease. It is a group of diseases that cause brain tissue damage due to sudden rupture of brain blood vessels or blockage of blood vessels, including ischemic and hemorrhagic strokes. The incidence of ischemic stroke is higher than that of hemorrhagic stroke, accounting for 60% to 70% of the total number of strokes.

[0003] Since stroke has the characteristics of rapid onset and sudden onset, prevention is more important than treatment. However, the existing technology lacks the corresponding technical means to conduct initial screening of patients, and prevention work lacks specificity. Summary of the invention

[0004] The main purpose of this application is to provide an acute ischemic stroke prediction system and prediction method, aiming to solve the defect of lack of targetedness in stroke prevention in the prior art.

[0005] This application achieves the above objectives through the following technical solutions:

[0006] An acute ischemic stroke prediction system, comprising:

[0007] A data acquisition unit, used for acquiring basic physiological parameters of a number of stroke patients and a number of non-stroke patients;

[0008] A data preprocessing unit, used for preprocessing the basic physiological parameters to obtain a training data set;

[0009] A model training unit, used for training and generating a stroke prediction model based on the training data set;

[0010] A detection unit, used to obtain health data information of the individual to be detected;

[0011] The prediction unit is used to input the health data information into the stroke prediction model and output the probability that the individual to be detected suffers from a stroke.

[0012] Optionally, the detection unit includes a detection terminal, a blood pressure detector, a blood glucose meter and a heart rate detector, and the control terminal is communicatively connected to the blood pressure detector, the blood glucose meter and the heart rate detector respectively.

[0013] Optionally, the detection terminal includes: a communication module, the communication module is respectively connected to the blood pressure detector, the blood glucose meter and the heart rate detector to obtain physiological parameters such as blood pressure value, blood glucose value, heart rate value of the individual to be detected;

[0014] The questionnaire survey unit is used to obtain data information such as age, weight, smoking history, alcohol abuse history and basic medical history of the individual to be tested.

[0015] Accordingly, the present application also discloses a prediction method based on the above prediction system, comprising the following steps:

[0016] The basic physiological parameters of several stroke patients and several non-stroke patients are obtained; wherein the basic physiological parameters include blood pressure parameters, blood sugar parameters, heart rate parameters, age, weight, smoking history, alcohol abuse history and basic medical history.

[0017] Pre-processing the basic physiological parameters to obtain a training data set;

[0018] Training and generating a stroke prediction model based on the training data set;

[0019] Obtain health data information of the individual to be tested;

[0020] The health data information is input into the stroke prediction model, and the probability of the individual to be tested suffering from stroke is output.

[0021] Optionally, pre-processing the basic physiological parameters to obtain a training data set includes the following steps:

[0022] Obtain basic physiological parameters (x 11 、x 12 , ..., x 1m ,y1)1,(x 21 、x 22 , ..., x 2m ,y2)2,...,(x n1 、x n2 , ..., x nm ,y n ) n ; where n represents the basic physiological parameter number, m represents the data type number, such as 1 represents age, 2 represents weight, y1 to y n It represents the probability of illness, with a value of 0.999 for stroke patients and a value determined by medical staff for non-stroke patients;

[0023] Reading non-numeric parameters in each of the basic physiological parameters, and converting the non-numeric parameters into digital parameters according to a digital conversion formula, wherein the non-numeric parameters include smoking history, alcohol abuse history and basic medical history; the digital conversion formula is expressed as Q=K*L, wherein K is a constant, K is 0 or 1, 0 represents no, 1 represents yes; L represents age;

[0024] According to the numerical conversion results, the basic physiological parameters are unified and the training data set {(x 11 、x 12 , ..., x 1m ,y1)1,(x 21 、x 22 , ..., x 2m ,y2)2,...,(x n1 、x n2 , ..., x nm ,y n ) n}.

[0025] Optionally, training and generating a stroke prediction model based on the training data set comprises the following steps:

[0026] Construct a linear regression equation and a logistic function, wherein the linear regression equation is expressed as z=α0+α1x n1 +α2x n2 +...+α m x nm ; The expression of the logic function is Where α0 is a constant, α1, α2, ..., α m All represent weight coefficients, and z represents regression value;

[0027] Get the training dataset;

[0028] Calculate the constant and each weight coefficient of the linear regression equation according to the least square method and the training data set;

[0029] Substitute the constant and each weight coefficient into the linear regression equation to obtain the linear regression prediction equation;

[0030] The linear regression prediction equation and the logistic function are combined as a stroke prediction model.

[0031] Optionally, calculating the constant and each weight coefficient of the linear regression equation according to the least squares method and the training data set comprises the following steps:

[0032] Establish an independent variable matrix X according to the training data set;

[0033] Constructing a dependent variable vector Y according to the training data set and the logic function;

[0034] Combine the independent variable matrix X to calculate the independent variable transposed matrix X T ;

[0035] Calculate the weight coefficient vector B, the calculation formula of the weight coefficient vector is B = (X T X) -1 XT Y;

[0036] The constant and each of the weight coefficients are obtained according to the weight coefficient vector.

[0037] Optionally, obtaining health data information of the individual to be tested includes the following steps:

[0038] The data information such as age, weight, smoking history, alcohol abuse history and basic medical history of the individual to be tested is obtained through the questionnaire survey unit;

[0039] Read the non-numeric parameters in each of the data information, and convert the non-numeric parameters into digital parameters according to a digital conversion formula, wherein the non-numeric parameters include smoking history, alcohol abuse history and basic medical history; the digital conversion formula is expressed as Q=K*L, wherein K is a constant, K is 0 or 1, 0 represents no, 1 represents yes; L is the age;

[0040] Setting a sampling frequency, and collecting a number of blood pressure parameters, a number of blood sugar parameters, and a number of heart rate parameters respectively according to the sampling frequency;

[0041] performing noise reduction processing on each of the blood pressure parameters, the plurality of blood sugar parameters and the plurality of heart rate parameters respectively;

[0042] The digitally converted data information and the noise-reduced blood pressure parameters, several blood sugar parameters and several heart rate parameters are integrated into health data information.

[0043] Optionally, performing noise reduction processing on each of the blood pressure parameters, the plurality of blood sugar parameters and the plurality of heart rate parameters respectively comprises the following steps:

[0044] Set the screening threshold;

[0045] Respectively calculating the blood pressure average value of each of the blood pressure parameters, the blood sugar average value of each of the blood sugar parameters, and the heart rate average value of each of the heart rate parameters;

[0046] Respectively comparing the deviations of the blood pressure parameters relative to the blood pressure average value, and screening out the blood pressure parameters with deviations greater than the screening threshold;

[0047] Respectively comparing the deviations of the blood glucose parameters relative to the blood glucose average value, and screening out the blood glucose parameters with deviations greater than the screening threshold;

[0048] Respectively comparing the deviations of the heart rate parameters relative to the heart rate average value, and screening out the heart rate parameters whose deviations are greater than the screening threshold;

[0049] Optionally, inputting the health data information into the stroke prediction model to output the probability of the individual to be detected suffering from stroke comprises the following steps:

[0050] Set the detection threshold P0;

[0051] Obtaining health data information and inputting it into the stroke prediction model to calculate the probability of illness P1;

[0052] The disease probability P1 is compared with the detection threshold P0. If P1≥P0 is satisfied, it is determined that the patient to be detected may suffer from a stroke, and the disease probability P1 is output.

[0053] Compared with the prior art, this application has the following beneficial effects:

[0054] The present application includes a data acquisition unit, a data processing unit, a model training unit, a detection unit and a prevention unit, wherein the data acquisition unit is used to obtain basic physiological parameters of a number of stroke patients and a number of non-stroke patients; the data preprocessing unit performs pre-processing on the basic physiological parameters to obtain a training data set, and the model training unit generates a stroke prediction model based on the training data set; the detection unit is used to obtain health data information of the individual to be detected, and the prediction unit outputs the probability of the individual to be detected suffering from stroke according to the health data information and the stroke prediction model;

[0055] The key physiological parameters of stroke patients will change before and after the onset of the disease. Compared with the prior art, the present application constructs a stroke prediction model through the basic physiological parameters of several stroke patients and several non-stroke patients, that is, the probability of the person to be tested suffering from stroke is determined by the comprehensive determination of various physiological parameters, that is, the corresponding probability of disease can be quickly obtained by inputting multiple physiological parameters, which can effectively improve the detection efficiency;

[0056] Secondly, the present application links various factors with the incidence of stroke through a predictive model and quantifies the above relationship, which improves the objectivity of the evaluation of the patient's physical condition, while also avoiding misjudgment caused by medical staff due to experience or personal preference, further improving the accuracy of the prediction.

[0057] Finally, before some key physiological parameters deteriorate but before the disease occurs, the present application can output the probability of disease through the fluctuation of physiological parameters, thereby conducting early screening of people at high risk of stroke, making it convenient for medical staff and family members to provide targeted preventive care for high-risk groups and reduce the incidence of stroke. It is also suitable for families and is convenient for self-testing in daily life. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A schematic diagram of the structure of an acute ischemic stroke prediction system provided in Example 1 of the present application;

[0059] Figure 2 It is a structural schematic diagram of the detection unit;

[0060] Figure 3 It is a structural diagram of the detection terminal;

[0061] Figure 4 A flowchart of the prediction method provided in Example 2 of the present application;

[0062] Figure numerals: 1-detection terminal, 2-blood pressure detector, 3-blood glucose meter, 4-heart rate detector.

[0063] The purpose, features and advantages of this application will be further described in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0065] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0066] In the present invention, unless otherwise clearly specified and limited, the terms "connection", "fixation", etc. should be understood in a broad sense. For example, "fixation" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0067] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the meaning of "and / or" appearing in the full text includes three parallel schemes. Taking "robot coordinate system and / or m" as an example, it includes the robot coordinate system scheme, or the m scheme, or the scheme that the robot coordinate system and m satisfy at the same time. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in this field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0068] Embodiment 1:

[0069] Reference Figure 1 This embodiment, as an optional embodiment of the present application, discloses an acute ischemic stroke prediction system, including a data acquisition unit, a data preprocessing unit, a model training unit, a detection unit and a prediction unit, wherein the prediction system also includes a control terminal, and the data acquisition unit is a data collection program, which is used to obtain basic physiological parameters of a number of stroke patients and a number of non-stroke patients;

[0070] Preferably, the basic physiological parameters of the above-mentioned stroke patients can be obtained from the case database of medical institutions such as hospitals, and the basic physiological parameters of non-stroke patients can be obtained through physical examinations and other methods;

[0071] The control terminal is a computer, and the data preprocessing unit, the model training unit and the prediction unit are all stored in the control terminal in the form of program packages;

[0072] The detection unit includes a control terminal, a blood pressure detector 2, a blood glucose meter and a heart rate detector. The detection terminal 1 includes:

[0073] A communication module, wherein the communication module is respectively connected to the blood pressure detector 2, the blood glucose meter 3 and the heart rate detector 4 to obtain physiological parameters such as blood pressure value, blood glucose value, heart rate value, etc. of the individual to be detected; it should be noted that the above connection can be connected through a local area network or through a wireless communication device such as Bluetooth;

[0074] For hospitals and other medical institutions, local area network and other methods are preferred, while for home users, wireless communication devices such as Bluetooth are preferred to achieve communication connection;

[0075] A questionnaire survey unit, wherein the questionnaire survey module is used to obtain data information such as age, weight, smoking history, alcohol abuse history and basic medical history of the individual to be tested. The control terminal is respectively connected to the blood pressure detector 2, the blood glucose meter 3 and the heart rate detector 4 for communication;

[0076] The questionnaire survey unit collects relevant data in the form of a questionnaire or the like.

[0077] Example 2

[0078] Reference Figure 2 This embodiment is an optional embodiment of the present application, which discloses a method for predicting acute ischemic stroke, comprising the following steps:

[0079] S1. Obtain basic physiological parameters of several stroke patients and several non-stroke patients; wherein the basic physiological parameters include blood pressure parameters, blood sugar parameters, heart rate parameters, age, weight, smoking history, alcohol abuse history and basic medical history.

[0080] The medical records of some stroke patients are selected from the case database of medical institutions, and basic physiological parameters are obtained from the medical records; for non-stroke patients, their basic physiological parameters can be obtained through volunteers or other channels;

[0081] To further improve the accuracy of prediction, the basic physiological parameters of some patients before the onset of the disease can also be used as the parameters of non-stroke patients;

[0082] The basic physiological parameters include blood pressure parameters, blood sugar parameters, heart rate parameters, age, weight, smoking history, alcohol abuse history and basic medical history. The above parameters are arranged in no particular order and can be flexibly adjusted according to actual conditions.

[0083] It should be noted that the basic diseases that need to be inquired about in the basic medical history include cardiovascular disease history, family history of stroke, history of hypertension, history of diabetes, history of heart disease, previous history of stroke and history of hyperhomocysteinemia, etc.

[0084] S2. Pre-processing the basic physiological parameters to obtain a training data set;

[0085] S21. Obtain basic physiological parameters (x 11 、x 12 , ..., x 1m ,y1)1,(x 21 、x 22 , ..., x 2m ,y2)2,...,(x n1 、x n2 , ..., x nm ,y n ) n; where n represents the basic physiological parameter number, m represents the data type number, such as 1 represents age, 2 represents weight, y1 to y n It represents the probability of illness, with a value of 0.999 for stroke patients and a value determined by medical staff for non-stroke patients;

[0086] Retrieve all basic physiological parameters in step S1, group the relevant data of the same person together, and complete the sorting of all parameters to obtain basic physiological parameters (x 11 、x 12 , ..., x 1m ,y1)1,(x 21 、x 22 , ..., x 2m ,y2)2,...,(x n1 、x n2 , ..., x nm ,y n ) n ; Where n represents the basic physiological parameter number, which can also be understood as the patient number;

[0087] m represents the data type number, such as 1 for age, 2 for weight, y1 to y n Indicates the probability of illness. The value for stroke patients is 0.999. In order to reduce the difficulty of subsequent data calculation, the value of stroke patients cannot be 1. However, in order to ensure the accuracy of the prediction, its value cannot be lower than 0.999.

[0088] The values ​​of non-stroke patients are taken by medical staff. The above values ​​are taken by medical staff based on comprehensive judgment of basic physiological parameters. The values ​​can also be independently given by multiple medical staff and their average value is taken as the probability of disease.

[0089] S22, reading non-numeric parameters in each of the basic physiological parameters, and converting the non-numeric parameters into digital parameters according to a digital conversion formula, wherein the non-numeric parameters include smoking history, alcohol abuse history and basic medical history; the digital conversion formula is expressed as Q=K*L, wherein K is a constant, K is 0 or 1, 0 represents no, 1 represents yes; L is the age;

[0090] For example, when reading a set of basic physiological parameters, if it is found that the person has a history of smoking for 12 years, the constant K is 1. At the same time, since the person has a smoking history of 12 years, the converted digital parameter is 12.

[0091] Similarly, if there is no history of alcohol abuse, the converted parameter is 0;

[0092] It should be noted that the non-numeric parameters in the basic physiological parameters include smoking history, alcoholism history and basic medical history; each disease in the basic medical history generates a separate parameter. For example, if the basic medical history includes cardiovascular disease history and family stroke history, it will include two parameters after conversion. The cardiovascular disease history also needs to investigate the corresponding years of illness.

[0093] Family history of stroke does not need to be investigated in terms of years, but the number of people can be investigated. That is, if no family member has suffered from stroke, the numerical parameter is zero; if two people have suffered from stroke, the numerical parameter is 2.

[0094] S23, unify the basic physiological parameters according to the numerical conversion results, and output the training data set {(x 11 、x 12 , ..., x 1m ,y1)1,(x 21 、x 22 , ..., x 2m ,y2)2,...,(x n1 、x n2 , ..., x nm ,y n ) n}.

[0095] Through step S22, all non-numeric parameters are adjusted and replaced, and the acquired parameters are sorted to obtain the training data set {(x 11 、x 12 , ..., x 1m ,y1)1,(x 21 、x 22 , ..., x 2m ,y2)2,...,(x n1 、x n2 , ..., x nm ,y n ) n}.

[0096] S3, training and generating a stroke prediction model based on the training data set;

[0097] S31, construct a linear regression equation and a logistic function, the linear regression equation is expressed as z=α0+α1x n1 +α2x n2 +...+α m x nm ; The expression of the logic function is Where α0 is a constant, α1, α2, ..., α m All represent weight coefficients, and z represents regression value;

[0098] First, a linear regression equation and a logistic function are constructed, wherein the linear regression equation is expressed as z=α0+α1x n1 +α2x n2 +...+α m x nm ; The expression of the logic function is Where α0 is a constant, α1, α2, ..., α m All represent weight coefficients, and z represents regression value;

[0099] S32, obtaining a training data set;

[0100] Retrieve the training data set {(x 11 、x 12 , ..., x 1m ,y1)1,(x 21 、x 22 , ..., x 2m ,y2)2,...,(x n1 、x n2 , ..., x nm ,y n ) n};

[0101] S33, calculating the constant and each weight coefficient of the linear regression equation according to the least square method and the training data set;

[0102] S331, establishing an independent variable matrix X according to the training data set;

[0103] In the training data set {(x 11 、x 12 , ..., x 1m ,y1)1,(x 21 、x 22 , ..., x 2m ,y2)2,...,(x n1 、x n2 , ..., x nm ,y n ) n}, x 11 to x nm are all independent variable parameters, and y n is the probability of illness, which is the dependent variable parameter;

[0104] By extracting all the independent variable parameters, the independent variable matrix X can be constructed. Its specific expression is:

[0105] In the above, 1 corresponds to the constant α0.

[0106] S332, constructing a dependent variable vector Y according to the training data set and the logic function;

[0107] First call the logic function Then, the disease probabilities in the training data set are retrieved, and the regression indices z1, ..., z corresponding to the disease probabilities are calculated according to the logistic function. n ;

[0108] Then, the dependent variable vector Y is constructed by the calculated regression value; the expression of the dependent variable vector is:

[0109] S333, calculate the independent variable transposition matrix X in combination with the independent variable matrix X T ;

[0110] S334, calculate the weight coefficient vector B, the calculation formula of the weight coefficient vector is B=(X T X) -1 X T Y;

[0111] After calculating the independent variable transpose matrix X T After that, first calculate the inverse matrix, that is, (X T X) -1 ;

[0112] Then according to the formula B = (X T X) -1 X T Y calculates the weight coefficient vector B;

[0113] S335. Obtain the constant and each of the weight coefficients according to the weight coefficient vector.

[0114] Each data in the weight coefficient vector B is the weight coefficient, that is:

[0115]

[0116] S34, bringing the constant and each weight coefficient into the linear regression equation to obtain a linear regression prediction equation;

[0117] Substitute the above parameters into the linear regression equation z = α0 + α1x n1 +α2x n2 +...+α m x nm The linear regression prediction equation can be obtained;

[0118] S35, combining the linear regression prediction equation and the logistic function as a stroke prediction model.

[0119] The above calculation formula can establish a connection between multiple independent variables (physiological parameters such as age, weight, blood pressure value, etc.) and the dependent variable (the probability of suffering from stroke), thus simplifying the calculation procedure;

[0120] Secondly, in the process of model construction, multiple independent variables were introduced, and weight parameters were used to reflect the influence of each parameter on stroke. The determination of the above weight parameters needs to be combined with a large number of sample parameters. That is, if heart rate imbalance has a significant impact on stroke, then heart rate abnormalities must be common in the patient samples. The only difference is the severity of heart rate abnormalities. Therefore, the weight parameter corresponding to heart rate must be higher.

[0121] Therefore, the calculation of this application can truly reflect the influence of various factors on stroke and improve the accuracy of prediction as much as possible;

[0122] Finally, since the prediction model of the present application contains multiple different independent variables, and during the model construction process, they can be added and adjusted according to actual needs, it can effectively reduce the impact of a single factor on the final judgment, which is conducive to improving the accuracy of the prediction.

[0123] S4. Obtaining health data information of the individual to be tested;

[0124] S41. Obtaining data information such as age, weight, smoking history, alcohol abuse history, and basic medical history of the individual to be tested through a questionnaire survey unit;

[0125] S42, reading non-numeric parameters in each of the data information, and converting the non-numeric parameters into digital parameters according to a digital conversion formula, wherein the non-numeric parameters include smoking history, alcohol abuse history and basic medical history; the digital conversion formula is expressed as Q=K*L, wherein K is a constant, K is 0 or 1, 0 represents no, 1 represents yes; L is the age;

[0126] The processing method for non-numeric parameters in this step is exactly the same as the processing method in step S22;

[0127] S43, setting a sampling frequency, and collecting a number of blood pressure parameters, a number of blood sugar parameters, and a number of heart rate parameters respectively according to the sampling frequency;

[0128] The sampling frequency is set according to the characteristics of the blood pressure monitor, blood glucose meter, heart rate monitor and other equipment. The sampling frequencies can be the same or different.

[0129] Then, each device is controlled to collect a number of blood pressure parameters, a number of blood sugar parameters and a number of heart rate parameters according to the set sampling frequency;

[0130] S44, performing noise reduction processing on each of the blood pressure parameters, the plurality of blood sugar parameters and the plurality of heart rate parameters respectively;

[0131] S441, setting a screening threshold;

[0132] The screening threshold can be determined according to the actual situation. Preferably, the screening threshold is selected as 5%-10%;

[0133] S442, respectively calculating the blood pressure average of each of the blood pressure parameters, the blood sugar average of each of the blood sugar parameters, and the heart rate average of each of the heart rate parameters;

[0134] Assume that the collected blood pressure parameters are H1, H2, ..., H i , where i represents the number of blood pressure parameters;

[0135] First, calculate the average blood pressure, which is expressed as H0 = (H1 + H2 + ... + H i ) / i;

[0136] Similarly, calculate the average blood sugar value G0 and the average heart rate HR0;

[0137] S443, respectively comparing the deviations of the blood pressure parameters relative to the blood pressure average value, and screening out the blood pressure parameters with deviations greater than the screening threshold;

[0138] The deviation of each blood pressure parameter relative to the average blood pressure value is calculated respectively, and the calculation formula is (H i -H0)*100%, screening out blood pressure parameters with a deviation greater than the screening threshold;

[0139] S444, respectively comparing the deviations of the blood glucose parameters relative to the blood glucose average value, and screening out the blood glucose parameters with deviations greater than the screening threshold;

[0140] S445, respectively comparing the deviations of the heart rate parameters relative to the heart rate average value, and screening out the heart rate parameters whose deviations are greater than the screening threshold;

[0141] S45, integrating the digitally converted data information and the blood pressure parameters, several blood sugar parameters and several heart rate parameters that have undergone noise reduction processing into health data information.

[0142] After the screening is completed, the average values ​​of the remaining parameters are recalculated and used as blood pressure parameters, blood glucose parameters, and heart rate parameters;

[0143] Finally, all parameters are integrated to obtain health data information (A1, A2, ..., A m );

[0144] Through the above screening, some distorted parameters can be removed, thereby improving the accuracy of the final detection parameters;

[0145] Secondly, the present application can effectively avoid the impact of short-term fluctuations in physiological parameters by collecting multiple data and calculating average values, thereby ensuring that all parameters can truly reflect body functions and improve the accuracy of detection.

[0146] S5. Input the health data information into the stroke prediction model, and output the probability of the individual to be tested suffering from stroke.

[0147] S51, setting a detection threshold P0;

[0148] The detection threshold can be determined according to the actual situation, such as 50% or 0.5;

[0149] S52, obtaining health data information and inputting it into the stroke prediction model to calculate the probability of illness P1;

[0150] The health data information (A1, A2, ..., A m ) is brought into the stroke prediction model to calculate the probability of disease P1;

[0151] S53, comparing the disease probability P1 with the detection threshold P0, if P1≥P0 is satisfied, determining that the patient to be detected may suffer from stroke, and outputting the disease probability P1;

[0152] Compared with the prior art, the present application constructs a stroke prediction model through the basic physiological parameters of several stroke patients and several non-stroke patients, that is, the probability of the person to be tested suffering from stroke is determined by the comprehensive determination of various physiological parameters, that is, the corresponding probability of suffering from stroke can be quickly obtained by inputting multiple physiological parameters, which can effectively improve the detection efficiency;

[0153] Secondly, the present application links various factors with the incidence of stroke through a predictive model and quantifies the above relationship, which improves the objectivity of the evaluation of the patient's physical condition, while also avoiding misjudgment caused by medical staff due to experience or personal preference, further improving the accuracy of the prediction.

[0154] Finally, before some key physiological parameters deteriorate but before the disease occurs, the present application can output the probability of disease through the fluctuation of physiological parameters, thereby conducting early screening of people at high risk of stroke, making it convenient for medical staff and family members to provide targeted preventive care for high-risk groups and reduce the incidence of stroke. It is also suitable for families and is convenient for self-testing in daily life.

[0155] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. An acute ischemic stroke prediction system, characterized in that: include: A data acquisition unit, used for acquiring basic physiological parameters of a number of stroke patients and a number of non-stroke patients; A data preprocessing unit, used for preprocessing the basic physiological parameters to obtain a training data set; A model training unit, used for training and generating a stroke prediction model based on the training data set; A detection unit, used to obtain health data information of the individual to be detected; The prediction unit is used to input the health data information into the stroke prediction model and output the probability that the individual to be detected suffers from a stroke.

2. The acute ischemic stroke prediction system according to claim 1, characterized in that: The detection unit comprises a detection terminal (1), a blood pressure detector (2), a blood glucose meter (3) and a heart rate detector (4), and the control terminal is respectively connected to the blood pressure detector (2), the blood glucose meter (3) and the heart rate detector (4) for communication.

3. The acute ischemic stroke prediction system according to claim 2, characterized in that: The detection terminal (1) comprises: a communication module, wherein the communication module is respectively connected to the blood pressure detector (2), the blood glucose meter (3) and the heart rate detector (4) for communication, so as to obtain physiological parameters such as blood pressure value, blood glucose value, heart rate value, etc. of the individual to be detected; The questionnaire survey unit is used to obtain data information such as age, weight, smoking history, alcohol abuse history and basic medical history of the individual to be tested.

4. The prediction method of the acute ischemic stroke prediction system according to any one of claims 1 to 3, characterized in that: The following steps are involved: The basic physiological parameters of several stroke patients and several non-stroke patients are obtained; wherein the basic physiological parameters include blood pressure parameters, blood sugar parameters, heart rate parameters, age, weight, smoking history, alcohol abuse history and basic medical history. Pre-processing the basic physiological parameters to obtain a training data set; Training and generating a stroke prediction model based on the training data set; Obtain health data information of the individual to be tested; The health data information is input into the stroke prediction model, and the probability of the individual to be tested suffering from stroke is output.

5. The prediction method according to claim 4, characterized in that: The pre-processing of the basic physiological parameters to obtain a training data set includes the following steps: Obtain basic physiological parameters (x 11 、x 12 , ..., x 1m ,y1)1,(x 21 、x 22 , ..., x 2m ,y2)2,...,(x n1 、x n2 , ..., x nm ,y n ) n ; where n represents the basic physiological parameter number, m represents the data type number, such as 1 represents age, 2 represents weight, y1 to y n It represents the probability of illness, with a value of 0.999 for stroke patients and a value determined by medical staff for non-stroke patients; Reading non-numeric parameters in each of the basic physiological parameters, and converting the non-numeric parameters into digital parameters according to a digital conversion formula, wherein the non-numeric parameters include smoking history, alcohol abuse history and basic medical history; the digital conversion formula is expressed as Q=K*L, wherein K is a constant, K is 0 or 1, 0 represents no, 1 represents yes; L represents age; According to the numerical conversion results, the basic physiological parameters are unified and the training data set {(x 11 、x 12 , ..., x 1m ,y1)1,(x 21 、x 22 , ..., x 2m ,y2)2,...,(x n1 、x n2 , ..., x nm ,y n ) n }.

6. The prediction method according to claim 4, characterized in that: The training and generating of the stroke prediction model based on the training data set comprises the following steps: Construct a linear regression equation and a logistic function, wherein the linear regression equation is expressed as z=α0+α1x n1 +α2x n2 +...+α m x nm ; The expression of the logic function is Where α0 is a constant, α1, α2, ..., α m All represent weight coefficients, and z represents regression value; Get the training dataset; Calculate the constant and each weight coefficient of the linear regression equation according to the least square method and the training data set; Substitute the constant and each weight coefficient into the linear regression equation to obtain the linear regression prediction equation; The linear regression prediction equation and the logistic function are combined as a stroke prediction model.

7. The prediction method according to claim 6, characterized in that: The step of calculating the constant and each weight coefficient of the linear regression equation according to the least squares method and the training data set comprises the following steps: Establish an independent variable matrix X according to the training data set; Constructing a dependent variable vector Y according to the training data set and the logic function; Combine the independent variable matrix X to calculate the independent variable transposed matrix X T ; Calculate the weight coefficient vector B, the calculation formula of the weight coefficient vector is B = (X T X) -1 X T Y; The constant and each of the weight coefficients are obtained according to the weight coefficient vector.

8. The prediction method according to claim 4, characterized in that: The step of obtaining the health data information of the individual to be detected includes the following steps: The data information such as age, weight, smoking history, alcohol abuse history and basic medical history of the individual to be tested is obtained through the questionnaire survey unit; Read the non-numeric parameters in each of the data information, and convert the non-numeric parameters into digital parameters according to a digital conversion formula, wherein the non-numeric parameters include smoking history, alcohol abuse history and basic medical history; the digital conversion formula is expressed as Q=K*L, wherein K is a constant, K is 0 or 1, 0 represents no, 1 represents yes; L is the age; Setting a sampling frequency, and collecting a number of blood pressure parameters, a number of blood sugar parameters, and a number of heart rate parameters respectively according to the sampling frequency; performing noise reduction processing on each of the blood pressure parameters, the plurality of blood sugar parameters and the plurality of heart rate parameters respectively; The digitally converted data information and the noise-reduced blood pressure parameters, several blood sugar parameters and several heart rate parameters are integrated into health data information.

9. The prediction method according to claim 8, characterized in that: The noise reduction process is performed on each of the blood pressure parameters, the plurality of blood sugar parameters and the plurality of heart rate parameters, respectively, comprising the following steps: Set the screening threshold; Respectively calculating the blood pressure average value of each of the blood pressure parameters, the blood sugar average value of each of the blood sugar parameters, and the heart rate average value of each of the heart rate parameters; Respectively comparing the deviations of the blood pressure parameters relative to the blood pressure average value, and screening out the blood pressure parameters with deviations greater than the screening threshold; Respectively comparing the deviations of the blood glucose parameters relative to the blood glucose average value, and screening out the blood glucose parameters with deviations greater than the screening threshold; The deviations of the heart rate parameters relative to the heart rate average are compared respectively, and the heart rate parameters with deviations greater than the screening threshold are screened out.

10. The prediction method according to claim 4, characterized in that: The step of inputting the health data information into the stroke prediction model and outputting the probability of the individual to be detected suffering from a stroke comprises the following steps: Set the detection threshold P0; Obtaining health data information and inputting it into the stroke prediction model to calculate the probability of illness P1; The disease probability P1 is compared with the detection threshold P0. If P1≥P0 is satisfied, it is determined that the patient to be detected may suffer from a stroke, and the disease probability P1 is output.

Images