Disease risk positioning module and method

Abstract

The invention provides a disease risk positioning module and a disease risk positioning method, and the disease risk positioning module is mainly used for carrying out multiple times of grouping and iteration on a sample data set formed by collected physiological data of a plurality of users through a grouping algorithm, generating a disease risk grading table when a grouping result meets a convergence condition, and positioning the disease risk according to the disease risk grading table. Wherein the disease risk grading table can be used for subsequent disease risk grading of the user, calculation of risk scores of the physiological parameters of the user corresponding to the grading result and calculation of the prediction probability of disease occurrence of the user corresponding to the grading result.

Description

Technical Field

[0001] The present invention relates to a disease risk localization module and method, and more particularly to a disease risk localization module and method that inputs a large data sample data set into a clustering algorithm and performs multiple clustering and iterations to obtain a disease risk grading table that meets convergence conditions. Background Art

[0002] With the increase in computing power, various clustering algorithms and deep learning algorithms are being continuously implemented and applied, and are also being used for disease diagnosis and prediction. However, existing technologies cannot accurately classify disease risk based on a user's physiological data. Furthermore, a user's physiological data typically includes multiple physiological parameters, but existing technologies do not generate risk scores corresponding to these physiological parameters, nor do they calculate the predicted probability of a user developing the disease based on the classification results. Therefore, there is a need in the industry for a disease risk positioning module and method. Summary of the Invention

[0003] In order to solve the above-mentioned technical problems, the present invention provides a disease risk positioning module and method, and the disease risk positioning module mainly uses a clustering algorithm to perform multiple clustering and iterations on a sample data set formed by collecting physiological data of multiple users, and generates a disease risk grading table when the clustering results meet the convergence conditions. This disease risk grading table can be used for subsequent disease risk grading of users, calculation of risk scores of users' physiological parameters corresponding to the grading results, and calculation of the predicted probability of disease occurrence of users corresponding to the grading results.

[0004] Based on at least one purpose of the present invention, the technical solution of the present invention is to provide a disease risk positioning module, which includes a disease risk grading table generator. The disease risk grading table generator signal is connected to the sample database to receive a sample data set including a plurality of physiological data of a plurality of users. The disease risk grading table generator is used to perform multiple clustering and iterations on the plurality of physiological data of the sample data set, and check whether the clustering results generated by each clustering and iteration meet the convergence conditions. If the clustering results meet the convergence conditions, a disease risk grading table is generated according to the clustering results; otherwise, the next clustering and iteration is performed. The disease risk grading table is used to perform disease risk grading on the users to be graded to generate a grading result, and the grading result is used to calculate the risk score of each physiological parameter of the users to be graded corresponding to the grading result, and to calculate the predicted probability of disease occurrence of the users to be graded corresponding to the grading result.

[0005] Optionally, in an embodiment of the disease risk positioning module, the disease risk positioning module further includes a disease risk locator. The disease risk locator signal is connected to the disease risk grading table generator to receive the disease risk grading table. The disease risk locator is used to receive physiological data including various physiological parameters of the user to be graded, generate a grading result of the user to be graded based on the various physiological parameters of the user to be graded using the disease risk grading table, determine the starting position of each physiological parameter of the user to be graded using the disease risk grading table, convert the starting position of each physiological parameter of the user to be graded into an adjusted position of each physiological parameter of the user to be graded based on the grading result of the user to be graded, calculate the risk score of each physiological parameter of the user to be graded corresponding to the grading result based on the adjusted position of each physiological parameter of the user to be graded, and calculate the predicted probability of the user to be graded developing a disease within a period of time based on the grading result.

[0006] Optionally, in an embodiment of the disease risk positioning module, the disease risk grading table generator includes an initialization device, a hyperplane clustering device, a clustering value calculation device, an iteration device, a convergence condition checking device, a disease risk grading table generation device, and a parameter updating device. The initialization device is used to initialize the iteration value, the variable conversion value, and the number of clusters. The hyperplane clustering device signal is connected to the initialization device, and is used to obtain at least one hyperplane coordinate value calculation formula based on the clustering results of the plurality of physiological data of the sample data set and at least a portion of the plurality of physiological data, and use each hyperplane coordinate value calculation formula to calculate each hyperplane coordinate value of each physiological data of at least a portion of the plurality of physiological data of the sample data set, wherein the initial clustering result of the clustering result of the plurality of physiological data of the sample data set is a preset clustering result that meets the initial clustering number. The clustering value calculation device signal is connected to the hyperplane clustering device, and is used to calculate the clustering value of the physiological data based on each hyperplane coordinate value of the physiological data. The iterative device signal is connected to the clustering value calculation device, which uses the clustering rules to cluster the physiological data according to the clustering values of each physiological data and updates the clustering results. The convergence condition checking device signal is connected to the iterative device, which is used to determine whether the clustering results meet the convergence conditions. The disease risk grading table generation device signal is connected to the convergence condition checking device, and is used to generate a disease risk grading table based on the clustering results when the clustering results meet the convergence conditions. The parameter updating device signal is connected to the convergence condition checking device and the hyperplane clustering device, and is used to update the iteration value, variable conversion value and number of clusters when the clustering results do not meet the convergence conditions.

[0007] Optionally, in an embodiment of the disease risk positioning module, the disease risk locator includes a grading result calculation device, a prediction probability calculation device, a grading adjustment device and a score calculation device. The grading result calculation device is used to generate a grading result for the user to be graded based on the various physiological parameters of the user to be graded using a disease risk grading table. The prediction probability calculation device signal is connected to the grading result calculation device, and is used to calculate the predicted probability of the user to be graded developing a disease within a period of time based on the grading result. The grading adjustment device signal is connected to the grading result calculation device, and is used to determine the starting position of each physiological parameter of the user to be graded using the disease risk grading table, and convert the starting position of each physiological parameter of the user to be graded into an adjusted position of each physiological parameter of the user to be graded based on the grading result of the user to be graded. The score calculation device signal is connected to the grading adjustment device, and calculates the risk score of each physiological parameter of the user to be graded corresponding to the grading result based on the adjusted position of each physiological parameter of the user to be graded.

[0008] Optionally, in an embodiment of the disease risk positioning module, the hyperplane clustering device uses a support vector clustering algorithm (SVC), and further converts multiple physiological parameters of multiple physiological data based on the iteration value and the variable conversion value to generate at least one extended physiological parameter, and adds each extended physiological parameter to the multiple physiological data to thereby expand and update the multiple physiological data, and the hyperplane clustering device uses at least a portion of the multiple physiological data after expansion and update and the clustering results to calculate the calculation formula of each hyperplane coordinate value.

[0009] Optionally, in an embodiment of the disease risk positioning module, the clustering value calculation device determines a clustering value calculation formula according to the iteration value, and the clustering value calculation device uses the clustering value calculation formula to calculate the clustering value of the physiological data according to each hyperplane coordinate value of the physiological data.

[0010] Optionally, in an embodiment of the disease risk positioning module, the iteration device is used to determine the grouping rule according to the iteration value and the number of groups.

[0011] Optionally, in an embodiment of the disease risk positioning module, the convergence condition checking device checks whether the risk probability of each risk level in the disease risk grading table corresponding to the clustering result meets the convergence condition.

[0012] Based on at least one purpose of the present invention, the technical solution of the present invention is to provide a disease risk positioning method, which includes the following steps: receiving a sample data set including multiple physiological data of multiple users; performing multiple clustering and iteration on the multiple physiological data of the sample data set, and checking whether the clustering results generated by each clustering and iteration meet the convergence conditions; and if the clustering results meet the convergence conditions, generating a disease risk grading table based on the clustering results, otherwise, performing the next clustering and iteration; wherein the disease risk grading table is used to perform disease risk grading on the users to be graded to generate a grading result, and the grading result is used to calculate the risk score of each physiological parameter of the users to be graded corresponding to the grading result, and to calculate the predicted probability of disease occurrence of the users to be graded corresponding to the grading result.

[0013] Optionally, in an embodiment of the disease risk positioning method, the disease risk positioning method further includes the following steps: receiving physiological data including various physiological parameters of the user to be graded; using a disease risk grading table to generate a grading result of the user to be graded based on the various physiological parameters of the user to be graded; using the disease risk grading table to determine the starting position of each physiological parameter of the user to be graded; converting the starting position of each physiological parameter of the user to be graded into an adjusted position of each physiological parameter of the user to be graded based on the grading result of the user to be graded; calculating the risk score of each physiological parameter of the user to be graded corresponding to the grading result based on the adjusted position of each physiological parameter of the user to be graded; and calculating the predicted probability of the user to be graded developing a disease within a period of time based on the grading result.

[0014] Optionally, in an embodiment of the disease risk positioning method, the steps of performing multiple clustering and iterations on multiple physiological data of the sample data set include: initializing an iteration value, a variable conversion value and a clustering number; obtaining at least one hyperplane coordinate value calculation formula based on the clustering results of the multiple physiological data of the sample data set and at least a part of the multiple physiological data, and using each hyperplane coordinate value calculation formula to calculate each hyperplane coordinate value of each physiological data of at least a part of the multiple physiological data of the sample data set, wherein the initial clustering result of the clustering result of the multiple physiological data of the sample data set is a preset clustering result that meets the initial clustering number; calculating the clustering value of the physiological data according to each hyperplane coordinate value of the physiological data; using the clustering rules to cluster each physiological data according to the clustering value of each physiological data, and updating the clustering result; judging whether the clustering result meets the convergence condition; when the clustering result meets the convergence condition, generating a disease risk grading table according to the clustering result; and when the clustering result does not meet the convergence condition, updating the iteration value, the variable conversion value and the clustering number.

[0015] Optionally, in an embodiment of the disease risk positioning method, the disease risk positioning method further includes: converting multiple physiological parameters of multiple physiological data according to the iteration value and the variable conversion value to generate at least one extended physiological parameter, and adding each extended physiological parameter to the multiple physiological data to thereby expand and update the multiple physiological data; wherein a support vector clustering algorithm is used to obtain each hyperplane coordinate value calculation formula based on at least a part of the multiple physiological data after expansion and update and the clustering result, and each hyperplane coordinate value calculation formula is used to calculate each hyperplane coordinate value of each physiological data of at least a part of the multiple physiological data of the sample data set.

[0016] Optionally, in an embodiment of the disease risk positioning method, a clustering value calculation formula is determined according to the iteration value, and the clustering value calculation formula is used to calculate the physiological data clustering value according to each hyperplane coordinate value of the physiological data.

[0017] Optionally, in an embodiment of the disease risk positioning method, the clustering rule is determined according to the iteration value and the number of clusters.

[0018] Optionally, in an embodiment of the disease risk positioning method, checking whether the clustering result satisfies the convergence condition is checking whether the risk probability of each risk level in the disease risk grading table corresponding to the clustering result satisfies the convergence condition.

[0019] In summary, compared with the prior art, the disease risk localization module and method provided by the present invention have the following beneficial effects:

[0020] Multiple clustering and iterations are performed on a sample dataset of more than 10,000 physiological data points to generate a disease risk grading table that can accurately locate disease risks. Through this disease risk grading table, users' disease risks can be graded based on their subsequent physiological data, so users can clearly understand their risk positioning for the disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 4 is a schematic block diagram of a disease risk positioning module according to an embodiment of the present invention.

[0022] Figure 2 This is a schematic block diagram of a disease risk grading table generator of a disease risk positioning module according to an embodiment of the present invention.

[0023] Figure 3 It is a schematic block diagram of a disease risk locator of a disease risk locating module according to an embodiment of the present invention.

[0024] Figure 4 4 is a schematic flow chart of a disease risk positioning method according to an embodiment of the present invention.

[0025] Description of the figure number:

[0026] 1: Disease risk positioning module

[0027] 10: Sample database

[0028] 11: Disease Risk Rating Table Generator

[0029] 111: Initialize device

[0030] 112: Hyperplane clustering device

[0031] 113: Cluster numerical calculation device

[0032] 114: Iteration Device

[0033] 115: Parameter update device

[0034] 116: Convergence condition check device

[0035] 117: Disease risk grading table generation device

[0036] 12: Disease Risk Locator

[0037] 121: Grading result calculation device

[0038] 122: Prediction Probability Calculation Device

[0039] 123: Grading adjustment device

[0040] 124: Rating Calculation Device

[0041] S401~S411: steps. DETAILED DESCRIPTION

[0042] Please refer to Figure 1 , Figure 1 : is a schematic block diagram of a disease risk positioning module according to an embodiment of the present invention. The disease risk positioning module 1 includes at least a disease risk grading table generator 11, and the disease risk grading table generator 11 is signal-connected to a sample database 10. The sample database 10 stores a sample data set of a plurality of physiological data of a plurality of users. In the present invention, the sample data set comes from a specific health data center. The plurality of physiological data of the sample data set can come from the health questionnaire database, biochemical data database and human biological database of the specific health data center, and the present invention is not limited to this. The number of the plurality of physiological data stored in the specific health data center is at least tens of thousands, or even millions of data, and the number of users corresponding to the plurality of physiological data reaches 300,000, and each user has nearly two hundred or more feature values. For example, in the present invention, approximately 120,000 pieces of physiological data of male users and 120,000 pieces of physiological data of female users can be obtained to generate a subsequent disease risk grading table.

[0043] The disease risk grading table generator 11 is used to receive a sample data set including a plurality of physiological data of a plurality of users, perform multiple clustering and iterations on the plurality of physiological data of the sample data set, and check whether the clustering results generated by each clustering and iteration meet the convergence conditions. If the clustering results meet the convergence conditions, a disease risk grading table is generated based on the clustering results; otherwise, the next clustering and iteration is performed. The disease risk grading table is used to grade the disease risk of the users to be graded to generate a grading result, and the grading result is used to calculate the risk score of each physiological parameter of the users to be graded corresponding to the grading result, and to calculate the predicted probability of the occurrence of the disease for the users to be graded corresponding to the grading result. Simply put, the disease risk positioning module 1 mainly performs clustering and iteration on the plurality of physiological data of the sample data set to obtain a disease risk grading table that can be used for disease risk positioning.

[0044] The disease risk positioning module 1 further includes a disease risk locator 12. The disease risk locator 12 signal is connected to the disease risk grading table generator 11 to receive the disease risk grading table. The disease risk locator 12 is used to receive physiological data including various physiological parameters of the user to be graded, and use the disease risk grading table to generate a grading result for the user to be graded based on the various physiological parameters of the user to be graded. The disease risk locator 12 can also use the disease risk grading table to determine the starting position of each physiological parameter of the user to be graded, convert the starting position of each physiological parameter of the user to be graded into the adjusted position of each physiological parameter of the user to be graded according to the grading result of the user to be graded, and calculate the risk score of each physiological parameter of the user to be graded corresponding to the grading result based on the adjusted position of each physiological parameter of the user to be graded. In addition, the disease risk locator 12 can also calculate the predicted probability of the user to be graded developing a disease within a period of time (for example, half a year, one year, two years, three years or five years) based on the grading result.

[0045] Please refer to Figure 2 , Figure 2This is a schematic block diagram of a disease risk grading table generator for a disease risk location module according to an embodiment of the present invention. Disease risk grading table generator 11 can be implemented as a pure hardware circuit or by a microcontroller unit executing specific software. Disease risk grading table generator 11 includes an initialization device 111, a hyperplane clustering device 112, a clustering value calculation device 113, an iteration device 114, a parameter update device 115, a convergence condition check device 116, and a disease risk grading table generator 117. Hyperplane clustering device 112 is signal-connected to initialization device 111, clustering value calculation device 113 is signal-connected to hyperplane clustering device 112, iteration device 114 is signal-connected to clustering value calculation device 113, convergence condition check device 116 is signal-connected to iteration device 114, disease risk grading table generator 117 is signal-connected to convergence condition check device 116, and parameter update device 115 is signal-connected to convergence condition check device 116 and hyperplane clustering device 112.

[0046] Please also refer to Figure 2 and Figure 4 , Figure 4 This is a schematic flow chart of a disease risk location method according to an embodiment of the present invention. First, in step S401, the initialization device 111 initializes the iteration value i, variable conversion value n, and number of clusters k for the hyperplane clustering device 112. Next, in step S402, the hyperplane clustering device 112 receives a sample dataset.

[0047] Then, in step S403, the hyperplane clustering device 112 obtains at least one hyperplane coordinate value calculation formula based on the clustering results of the plurality of physiological data of the sample data set and at least a portion of the plurality of physiological data, wherein the initial clustering result of the clustering results of the plurality of physiological data of the sample data set is a preset clustering result that meets the initial clustering number k, that is, the initial preset clustering result is a clustering result of an artificial rule.

[0048] Furthermore, in step S403, before obtaining at least one hyperplane coordinate value calculation formula, the hyperplane clustering device 112 further converts the multiple physiological parameters of the multiple physiological data according to the iteration value i and the variable conversion value n to generate at least one extended physiological parameter, and adds each extended physiological parameter to the multiple physiological data to expand and update the multiple physiological data. Then, the hyperplane clustering device 112 uses a support vector clustering algorithm (SVC, support vector clustering) to obtain each hyperplane coordinate value calculation formula of at least a portion of the multiple physiological data based on the expanded and updated multiple physiological data and the clustering results.

[0049] Next, in step S404, the hyperplane clustering device 112 uses each hyperplane coordinate value calculation formula to calculate each hyperplane coordinate value of each physiological data of a plurality of physiological data of at least a portion of the sample data set. For example, the hyperplane coordinate value may be calculated only for the physiological data under a certain cluster, and the decision of which physiological data to calculate the hyperplane coordinate value is determined by the iteration value i.

[0050] Then, in step S405, the cluster value calculation device 113 calculates the cluster value of the physiological data based on the respective plane coordinate values of the physiological data. Furthermore, the cluster value calculation device 113 determines a cluster value calculation formula based on the iteration value i, and the cluster value calculation device 113 uses the cluster value calculation formula to calculate the cluster value of the physiological data based on the respective hyperplane coordinate values of the physiological data.

[0051] Next, in step S406, the iterative device 114 uses the clustering rules to cluster the physiological data according to the clustering values of each physiological data, and updates the clustering results. Specifically, the iterative device 114 uses the clustering rules to cluster the physiological data according to the clustering values of each physiological data to generate another clustering result. This new clustering result is then integrated with the original clustering result, thereby updating and replacing the original clustering result.

[0052] For example, the data are originally divided into two groups, Group A and Group B. In step S403, the multiple physiological data of Group B are used to train four hyperplane coordinate value calculation formulas (i.e., based on the physiological data of Group B, four hyperplanes can be divided to divide the groups). Next, in step S404, the four hyperplane coordinate values of each physiological data of Group B are calculated. In step S405, the clustering values of each physiological data of Group B are calculated. Then, in step S406, the clustering rule is to classify the physiological data with clustering values in the first range as Group B1, the physiological data with clustering values in the second range as Group B2, and the physiological data with clustering values in the third range as Group B3. Group B1 is merged with Group A to generate Group A*, Group B2 is merged with Group B to generate Group B*, and Group B3 is used as Group C*. Finally, the clustering results originally divided into Group A and Group B are updated to be divided into Group A*, Group B*, and Group C*. In short, steps S403 to S406 are equivalent to one clustering and iteration, and each clustering and iteration will update the clustering result.

[0053] Afterwards, in step S407, the convergence condition checking device determines whether the clustering result updated in step S406 meets the convergence condition. If the clustering result meets the convergence condition, step S408 is executed. If the clustering result does not meet the convergence condition, step S411 is executed. Furthermore, checking whether the clustering result meets the convergence condition is to check whether the risk probability of each risk level in the disease risk grading table corresponding to the clustering result meets the convergence condition. The risk probability of each risk level represents the risk probability of developing a disease within a period of time under this level (that is, the average probability of a disease occurring in a period of time for users with multiple physiological data corresponding to the risk level), and the convergence condition can be, for example, that the risk probability of developing a disease within one year for risk level 1 must be half of the risk probability of developing a disease for all users (that is, the average probability of a disease occurring in one year for users with all physiological data), and the risk probability of developing a disease within one year for risk level 2 must be twice the risk probability of developing a disease for all users, but the present invention is not limited to this.

[0054] In step S408, the disease risk grading table generating device 117 generates a disease risk grading table based on the clustering results, wherein the grading method of the disease risk grading table corresponds to the clustering results, and each grade has a plurality of average values of physiological parameters, and the average value of the physiological parameter is the average value of a certain type of physiological parameter of the plurality of physiological data of this grade. For example, a certain risk grade has 5 physiological data, and the physiological parameters of the 5 ages in the 5 physiological data are 20, 30, 24, 50, and 46 respectively, then the average value of the physiological parameter of the age of this grade is 34.

[0055] In step S411, the parameter updating device 115 updates the iteration value i, the variable conversion value n and the number of clusters of the hyperplane clustering device 112. Then, steps S402 to S407 are repeated again. In this way, multiple clustering and iterations are performed until the updated clustering result meets the convergence condition, and then step S408 is executed to generate a disease risk grading table. Doctors or nursing staff can use the disease risk grading table to grade the user to be graded according to the physiological parameters of the physiological data of the user to be graded and give medical advice. It can also be used by the disease risk locator 12 to execute steps S409 and S410 to perform risk positioning for the user to be graded, wherein steps S409 and S410 will be described later.

[0056] Please refer to Figure 3 , Figure 3This is a schematic block diagram of the disease risk locator of the disease risk location module according to an embodiment of the present invention. The disease risk locator 12 includes a grading result calculation device 121, a prediction probability calculation device 122, a grading adjustment device 123, and a score calculation device 124. The prediction probability calculation device 122 is signal-connected to the grading result calculation device 121, the grading adjustment device 123 is signal-connected to the grading result calculation device 121, and the score calculation device 124 is signal-connected to the grading adjustment device 123.

[0057] Please refer to Figure 3 and Figure 4 In step S409, the grading result calculation device 121 obtains the physiological data of the user to be graded, which includes multiple physiological parameters, and the grading result calculation device 121 uses the disease risk grading table to generate a grading result of the user to be graded according to each physiological parameter of the user to be graded, wherein the grading result indicates which risk level the user to be graded is graded into.

[0058] Next, in step S410, the disease risk locator 12 outputs the grading result through the grading result calculation device 121, and the disease risk locator 12 further outputs the predicted probability of the user to be graded developing the disease within a period of time and the risk score of each physiological parameter of the user to be graded corresponding to the grading result based on the grading result. Further, in step S410, the predicted probability calculation device 122 calculates the predicted probability of the user to be graded developing the disease within a period of time based on the grading result, the grading adjustment device 123 uses the disease risk grading table to determine the starting position of each physiological parameter of the user to be graded, and converts the starting position of each physiological parameter of the user to be graded into the adjusted position of each physiological parameter of the user to be graded based on the grading result of the user to be graded, and the score calculation device 124 calculates the risk score of each physiological parameter of the user to be graded corresponding to the grading result based on the adjusted position of each physiological parameter of the user to be graded.

[0059] Next, the disease risk assessment module or method according to the above-described embodiment will be described using an example of diabetes risk assessment for a male user who has not yet developed diabetes or hyperglycemia. In this example, the diabetes risk level is predefined as four levels, and the convergence condition is that the risk probability of diabetes for users in risk level 1 is less than 0.5 times that of the overall user, the risk probability for users in risk level 2 is approximately 1 times that of the overall user, the risk probability for users in risk level 2 is 2-3 times that of the overall user, and the risk probability for users in risk level 4 is 5-6 times that of the overall user. Using the disease risk grading table, diabetes risk assessment can be performed for the user to be graded. The predicted probability of the user developing diabetes within two years, given this risk level, is generated, along with the score for each physiological parameter of the user in this risk level (in this example, severity is expressed on a scale of 1 to 10, with higher scores indicating a greater need for control to prevent diabetes).

[0060] The sample data set is based on the physiological data of 123,013 male users. The physiological parameters of each male user include age, gender, height, weight, waist circumference, hip circumference, heart rate, diastolic blood pressure, systolic blood pressure, forced expiratory volume in the first second, vital capacity, whether taking anti-hyperlipidemic drugs, whether suffering from hypertension, whether having a stroke (cerebrovascular disease), whether having cardiovascular disease, whether having asthma, whether third-degree relatives (grandparents, parents, siblings and children) have hypertension, whether third-degree relatives have diabetes, total cholesterol, whether drinking alcohol, fasting blood sugar, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, whether smoking, triglycerides, whether third-degree relatives have cardiovascular disease, and whether third-degree relatives have had a stroke.

[0061] Furthermore, as previously described, before obtaining the hyperplane coordinate calculation formula, the physiological parameters are transformed according to the iteration value i and the variable transformation value n to generate at least one extended physiological parameter. In this example, when the iteration value i is 1 and the variable transformation value n is the initial value, the ratio of triglycerides to high-density lipoprotein cholesterol (kfc1) is calculated as one of the extended physiological parameters, the ratio of triglycerides to low-density lipoprotein cholesterol (kfc2) is calculated as another extended physiological parameter, the standard values of age, BMI, and waist-to-hip ratio are calculated as the other extended physiological parameters, and the positive deviation value dx = max[((x / s)-1)*100,0] and the negative deviation value ddx = min[((x / s)-1)*100,0] of each physiological parameter are calculated, where x represents a physiological parameter, s represents the standard value of the physiological parameter, and max[a,b] and min[a,b] represent the maximum and minimum values within the range between a and b (inclusive).

[0062] The sample dataset was initially clustered into two groups, Group A and Group B. Group A included the physiological data of all men aged 20 to 49 who were not hypertensive at the time of the data collection, while Group B included the physiological data of all men aged 50 to 100 who were not hypertensive at the time of the data collection. The first clustering and iteration involved training a hyperplane to cluster the multiple physiological data points in Group A. Five hyperplanes were obtained, and the corresponding hyperplane coordinate calculation formulas for these five hyperplanes are as follows:

[0063] HP1=-10.12+5.87*g_whr+0.09*d_dbp-0.03*d_fvc+0.21*d_fg-0.0042*d_kfc2+0.69*psick14+0.47*rsick11;

[0064] HP2=-13.28+0.074*bmi+7.21*g_whr+0.015*d_dbp+0.018*d_fev1-0.028*d_fvc+0.166*d_fg-0.295*drinkornot_98+0.9915*mdrug07+0.4321*rsick10;

[0065] HP3=-12.94+0.105*bmi+6.89*g_whr+0.016*d_fev1-0.025*d_fvc+0.186*d_fg+0.109*smokeornot_98-0.239*drinkornot_98+0.914*mdrug07+0.547*rsick10;

[0066] HP4 = -12.83 + 0.079*bmi + 6.95*g_whr + 0.017*d_dbp + 0.018*d_fev1 - 0.027*d_fvc + 0.164*d_fg - 0.301*drinkornot_98 + 0.9876*mdrug07 + 0.417*rsick10; and

[0067] HP5=-12.95+0.165*bmi+6.84*g_whr+0.017*d_fev1-0.026*d_fvc+0.171*d_fg+0.0075*d_chol-0.0021* d_kfc2+0.0027*d_kfc1+0.062*smokeornot_98-0.168*drinkornot_98+0.776*mdrug07+0.694*rsick10;

[0068] HP1 to HP5 are five hyperplane coordinate values. The physiological parameters used in the above formula are as follows: g_whr is waist-to-hip ratio, d_dbp is the positive deviation of diastolic blood pressure, d_fvc is the positive deviation of maximum breathing, d_fg is the positive deviation of blood glucose, d_kfc2 is the positive deviation of the index value of the aforementioned ratio kfc2, psick14 is whether there is asthma (1 means yes, 0 means no), rsick11 is whether a third-degree relative has had a stroke (1 means yes, 0 means no), bmi is BMI, d_fev1 is the positive deviation of the first second of forced exhalation, drinkornot_98 is whether to drink alcohol (1 means yes, 0 means no), where 1 indicates no drinking or less than once a week, 2 indicates former drinking but now quit drinking, 3 indicates 1-2 times a week, 4 indicates 3-4 times a week, and 5 indicates daily drinking), mdrug07 indicates whether anti-hyperlipidemic drugs are taken (1 indicates yes, 0 indicates no), rsick10 indicates whether a third-degree relative has diabetes (1 indicates yes, 0 indicates no), smokeornot_98 indicates whether a person smokes (1 indicates no smoking, 2 indicates no smoking but often smells secondhand smoke, 3 indicates former smoking but currently quit smoking, 4 indicates occasional smoking, and 5 indicates daily smoking), d_chol is a positive deviation of total cholesterol, and d_kfc1 is a positive deviation of the index value of the aforementioned ratio kfc1.

[0069] Next, the first clustering and iteration are performed to calculate the clustering value, and the calculation formula is as follows:

[0070] HPcount_1 = 1 (if HP1 is greater than -7.1) or 0 (if HP1 is less than or equal to -7.1);

[0071] HPcount_2 = 1 (if HP2 is greater than -5.3) or 0 (if HP2 is less than or equal to -5.3);

[0072] HPcount_3 = 1 (if HP3 is greater than -3.2) or 0 (if HP3 is less than or equal to -3.2);

[0073] HPcount_4=1 (if HP4 is greater than -2.7) or 0 (if HP4 is less than or equal to -2.7);

[0074] HPcount_5 = 1 (if HP5 is greater than -2.1) or 0 (if HP5 is less than or equal to -2.1); and

[0075] HPscore=HPcount_1+HPcount_2+HPcount_3+HPcount_4+HPcount_5;

[0076] HPscore is the clustering value of the physiological data.

[0077] Afterward, the first clustering and iteration process uses a clustering rule to cluster each physiological data point based on its HPscore. Here, the clustering rule assigns physiological data points with an HPscore of 0 to group A*, and assigns physiological data points with HPscores of 1 to 5, along with those originally in group B, to group B*. Next, the first clustering and iteration process checks that the updated clustering results (multiple physiological data points divided into groups A* and B*) do not meet convergence criteria, necessitating a second clustering and iteration process. The iteration value i, variable conversion value n, and number of clusters k are updated.

[0078] The second clustering and iteration does not convert the physiological parameters of the physiological data to expand and update the physiological data, and trains the hyperplane used to cluster the multiple physiological data of the B* group. A total of 5 hyperplanes are obtained, and the calculation formulas for the hyperplane coordinate values corresponding to these 5 hyperplanes are as follows:

[0079] HP1*=-16.4692+0.028*age+7.29381*g_whr+0.1827*d_fg+0.0981*smokeornot_98-0.119 4*drinkornot_98+0.6596*psick09+0.6942*psick14+0.4895*rsick10-0.4935*rsick12;

[0080] HP2*=-14.0532+0.0351*age+0.0707*bmi+4.42*g_whr+0.0072*d_hr+0.1596*d_fg+0.001*d_kfc2+0.152*smokeor not_98-0.1293*drinkornot_98+0.637*psick09+0.56*psick14+0.3994*mdrug07+0.647*rsick10-0.335*rsick12;

[0081] HP3*=-13.7849+0.0276*age+0.0936*bmi+4.8561*g_whr-0.0074*d_sbp+0.0101*d_hr+0.1672*d_fg+0.0015* d_kfc2+0.1561*smokeornot_98-0.1539*drinkornot_98+0.5745*psick09+0.5468*mdrug07+0.5693*rsick10;

[0082] HP4*=-14.0325+0.0235*age+0.1138*bmi+5.8238*g_whr-0.0105*d_sbp+0.0093*d_hr+0.1583*d_fg+0.0019*d_kfc2+0.0033*dd and

[0083] HP5*=-14.9094+0.02323*age+0.1304*bmi+6.9417*g_whr+0.0075*d_hr+0.1581*d_fg+0.0089*d_chol+0.0019*d_kfc1 +0.1133*smokeornot_98-0.1613*drinkornot_98+0.4597*psick09+0.3958*mdrug07+0.67*rsick10-0.2281*rsick12;

[0084] Among them, HP1*~HP5* are 5 hyperplane coordinate values respectively. The physiological parameters used in the above formula are explained as follows: age is age, psick09 is whether there is hypertension (1 for yes, 0 for no), rsick12 is whether the third-degree relative has hypertension (1 for yes, 0 for no), d_hr is the positive deviation value of blood pressure, d_sbp is the positive deviation value of systolic blood pressure, and dd_kfc2 is the negative deviation value of the index value of the aforementioned ratio kfc2.

[0085] Next, the second clustering and iteration are performed to calculate the clustering value, and the calculation formula is as follows:

[0086] HPcount_1*=1 (if HP1* is greater than -6.3) or 0 (if HP1* is less than or equal to -6.3);

[0087] HPcount_2*=1 (if HP2* is greater than -4.7) or 0 (if HP2* is less than or equal to -4.7);

[0088] HPcount_3*=1 (if HP3* is greater than -3.7) or 0 (if HP3* is less than or equal to -3.7);

[0089] HPcount_4*=1 (if HP4* is greater than -2.7) or 0 (if HP4* is less than or equal to -2.7);

[0090] HPcount_5*=1 (if HP5* is greater than -1.9) or 0 (if HP5* is less than or equal to -1.9); and

[0091] HPscore*=HPcount_1*+HPcount_2*+HPcount_3*+HPcount_4*+HPcount_5*;

[0092] HPscore* is the clustering value of the physiological data.

[0093] Afterward, a second clustering and iteration process uses the clustering rule to cluster each physiological data point based on its HPscore*. Here, the clustering rule is to classify physiological data points in group A* into group A**, physiological data points with an HPscore* value of 0 into group B**, and physiological data points with HPscore* values of 1 to 5 into group C**. Next, the updated clustering results (multiple physiological data points divided into groups A**, B**, and C**) from the second clustering and iteration process are checked and do not meet the convergence criteria. Therefore, a third clustering and iteration process is performed, and the iteration value i, variable conversion value n, and number of clusters k are updated.

[0094] The third clustering and iteration does not convert the physiological parameters of the physiological data to expand and update the physiological data, and trains the hyperplane used to cluster the multiple physiological data of the C** group. A total of 5 hyperplanes are obtained, and the calculation formulas for the hyperplane coordinate values corresponding to these 5 hyperplanes are as follows:

[0095] HP1**=-16.5355+0.0265*age+7.5791*g_whr+0.1827*d_fg-0.0083*dd_kfc2-0.1532* drinkornot_98+0.5918*psick09+0.8583*psick14+0.538*rsick10-0.6683*rsick12;

[0096] HP2**=-12.7436+0.0313*age+0.0778*bmi+2.9073*g_whr+0.0144*d_dbp-0.0154*d_sbp+0.01*dd_fev1+0.1544*d_fg+0.0013*d_kfc2+0 .1497*smokeornot_98-0.1591*drinkornot_98+0.6095*psick09+0.4848*psick14+0.4309*mdrug07-0.3359*rsick09-0.7551*rsick10;

[0097] HP3**=-12.9493+0.0311*age+0.1041*bmi+3.2428*gwhr+0.0128*d_dbp-0.0191*d_sbp+0.0089*d_hr+0.071*d_fvc+0.1617*d_fev1+0.0087*d _chol+0.0015*d_kfc1+0.1527*smokeornot_98-0.1826*drinkornot_98+0.5436*psick09+0.3812*psick14+0.5507*mdrug07+0.6068*rsick10;

[0098] HP4**=-13.4635+0.0253*age+0.1221*bmi+4.7293*g_whr-0.0124*d_sbp+0.0076*d_hr+0.0054*d_fev1+0.1551*d_fg+0.0125* d_chol+0.0021*d_kfc1+0.1153*smokeornot_98-0.1775*drinkornot_98+0.5508*psick09+0.389*mdrug07+0.5955*rsick10; and

[0099] HP5**=-13.328+0.023*age+0.1323*bmi+5.1874*g_whr+0.0061*d_fev1+0.1527*d_fg+0.0111*d_chol+0.0024*d_kfc1 +0.0953*smokeornot_98-0.1818*drinkornot_98+0.478*psick09+0.3364*mdrug07-0.1756*rsick09+0.7175*rsick10;

[0100] Among them, HP1**~HP5** are the coordinate values of 5 hyperplanes respectively.

[0101] Next, the third clustering and iteration are performed to calculate the clustering value, and the calculation formula is as follows:

[0102] HPcount_1**=1 (if HP1** is greater than -5.5) or 0 (if HP1** is less than or equal to -5.5);

[0103] HPcount_2**=1 (if HP2** is greater than -3.9) or 0 (if HP2** is less than or equal to -3.9);

[0104] HPcount_3**=1 (if HP3** is greater than -2.5) or 0 (if HP3** is less than or equal to -2.5);

[0105] HPcount_4**=1 (if HP4** is greater than -1.6) or 0 (if HP4** is less than or equal to -1.6);

[0106] HPcount_5**=1 (if HP5** is greater than -0.8) or 0 (if HP5** is less than or equal to -0.8); and

[0107] HPscore**=HPcount_1**+HPcount_2**+HPcount_3**+HPcount_4**+HPcount_5**;

[0108] HPscore** is the clustering value of the physiological data.

[0109] Afterward, the third clustering and iteration uses the clustering rule to cluster each physiological data according to its clustering value HPscore**. Here, the clustering rule is to classify physiological data in group A** into group A#, the clustering rule is to classify physiological data in group B** into group B#, physiological data with a clustering value HPscore** of 0 into group C#, physiological data with a clustering value HPscore** of 1 to 4 into group D#, and physiological data with an HPscore** of 5 into group E#. Next, the updated clustering results (multiple physiological data points divided into groups A#, B#, C#, D#, and E#) from the third clustering and iteration check do not meet the convergence criteria, so a fourth clustering and iteration is performed, and the iteration value i, variable conversion value n, and number of clusters k are updated.

[0110] During the fourth clustering and iteration, the iteration value i and the variable conversion value n meet the default requirements for expanding and updating the physiological parameters. Therefore, the fourth clustering and iteration first performs a physiological parameter conversion to generate at least one expanded physiological parameter. In this example, the physiological parameter DKAbio is calculated as follows: 0.9783*LevelB#+2.5003*LevelC#+3.4812*LevelD#+4.6959*LevelE#, where LevelB# indicates whether the physiological data belongs to the B# group (LevelB# is 1 when the physiological data belongs to the B# group, otherwise it is 0), LevelC# indicates whether the physiological data belongs to the C# group (LevelC# is 1 when the physiological data belongs to the C# group, otherwise it is 0), LevelD# indicates whether the physiological data belongs to the D# group (LevelD# is 1 when the physiological data belongs to the D# group, otherwise it is 0), and LevelE# indicates whether the physiological data belongs to the E# group (LevelE# is 1 when the physiological data belongs to the E# group, otherwise it is 0).

[0111] Afterwards, the fourth round of clustering and iterative training is used to cluster the multiple physiological data of group A#, group B#, group C#, group D#, and group E#. A total of one hyperplane is obtained, and the hyperplane coordinate calculation formula corresponding to this hyperplane is as follows:

[0112] HP=-5.2191+DKAbio+0.0032*d_dbp+0.0023*d_fg-0.003*HDL+0.0015*d_ldlc+0.2656*psick14+0.1147*psick16+0.152*psick09+0.1618*mdrug07;

[0113] Where HP is the hyperplane coordinate value. The physiological parameters used in the above formula are described as follows: d_ldlc is the positive deviation value of high-density lipoprotein cholesterol, and psick16 is whether the third-degree relative has cardiovascular disease (1 for yes, 0 for no).

[0114] Next, the fourth clustering and iteration are performed to calculate the clustering value, and the calculation formula is as follows:

[0115] HP# = 0 (if HP is less than or equal to -4.5), 1 (if HP# is greater than -4.5 and less than or equal to -3), 2 (if HP# is greater than -3 and less than or equal to -1.5), or 3 (if HP# is greater than -1.5);

[0116] HP# is the clustering value of the physiological data.

[0117] Afterward, the fourth clustering and iteration process uses the clustering rule to cluster each physiological data point based on its clustering value HP#. Here, the clustering rule is to classify physiological data points with clustering values HP# between 0 and 3 into groups A##, B##, C##, and D##, respectively. Next, the fourth clustering and iteration process checks that the updated clustering results (multiple physiological data points divided into groups A##, B##, C##, and D##) meet the convergence criteria (the corresponding risk probabilities for each risk level in the disease risk grading table are shown in Table 1. These meet the previously mentioned convergence criteria: "The risk probability of diabetes for users in risk level 1 is less than 0.5 times that of all users, the risk probability of diabetes for users in risk level 2 is approximately 1 times that of all users, the risk probability of diabetes for users in risk level 2 is 2-3 times that of all users, and the risk probability of diabetes for users in risk level 4 is 5-6 times that of all users"). Therefore, no further clustering and iteration are performed, and the disease risk grading table (see Table 2) is generated.

[0118] Table 1

[0119]

[0120]

[0121] Table 2

[0122]

[0123]

[0124] Next, the physiological data of the user to be classified is input, and each physiological parameter of the physiological data of the user to be classified is compared with each physiological parameter of each risk level in the disease risk grading table, and the risk level corresponding to the closest one is found as the risk level of the user to be classified. Furthermore, the user to be classified can be classified by calculating the difference between the BMI, waist-to-hip ratio, fasting blood glucose, triglycerides, total cholesterol, and high-density cholesterol of the physiological data of the user to be classified and the BMI, waist-to-hip ratio, fasting blood glucose, triglycerides, total cholesterol, and high-density cholesterol of each risk level. Assuming that the BMI, waist-to-hip ratio, fasting blood glucose, triglycerides, total cholesterol, and high-density cholesterol of the physiological data of the user to be classified are 25.0021, 0.9100, 124.5560, 160.5978, 204.1111, and 124.6045 respectively, then the user to be classified will be classified into risk level 3 (i.e., the grading result is expressed as risk level 3).

[0125] Next, the risk score of each physiological parameter of the user to be graded in risk level 3 is calculated. Taking fasting blood glucose as an example, the fasting blood glucose is 124.5560, and the average value of the fasting blood glucose closest to risk level 4 is (124.4486). Therefore, the starting position of fasting blood glucose is 3 (i.e., risk level minus one), and the starting position will be multiplied by the adjustment factor corresponding to the risk level to generate the adjusted position. The adjustment factors of risk levels 1 to 4 are set to 1.5 / 3, 2 / 3, 2.5 / 3, and 3 / 3 respectively. If the user to be tested already has diabetes, the starting position is directly multiplied by the adjustment factor of 4 / 3. After that, the final position is calculated. The final position is the adjusted position multiplied by 2 plus 1 and rounded off. The final position is the risk score of the physiological parameter of the user to be graded under the corresponding risk level. In this example, the risk score will fall between 1 and 10. The higher the risk score of the physiological parameter, the worse the physiological parameter is, and the user to be graded needs to pay attention to improvement. The grading results of the user to be graded and the risk score of the physiological parameter are presented in a table, as shown in Table 3 below.

[0126] Table 3

[0127]

[0128]

[0129] Finally, the formula for calculating the predicted probability of a user under risk classification developing diabetes within two years can be expressed as follows:

[0130] Risk level 1: 0.6875*EXP(Z) / (1+EXP(Z)), Z=-0.3892+0.8267*HP;

[0131] Risk level 2: 0.6875*EXP(Z) / (1+EXP(Z)), Z=-2.07769+10.0329*HP 2 / 100;

[0132] Risk level 3: 0.6875*EXP(Z) / (1+EXP(Z)), Z=0.06045+1.0137*HP; and

[0133] Risk level 4: 0.6875*EXP(Z) / (1+EXP(Z)), Z=-0.3518-49.2577*HP 2 / 100.

[0134] In summary, the disease risk positioning module and method provided by this invention performs multiple clustering and iterations on a sample dataset of more than 10,000 physiological data points to generate a disease risk grading table that accurately locates disease risk. This disease risk grading table can then be used to stratify the user's disease risk based on their subsequent physiological data. Based on this grading result, the predicted probability of the user developing the disease within a certain period of time and the score of each physiological parameter of the user can be calculated. In this way, the user can clearly understand their risk positioning for the disease.

Claims

1. A disease risk positioning module, characterized in that: include: A disease risk grading table generator (11) is signal-connected to a sample database (10) to receive a sample data set including a plurality of physiological data of a plurality of users; The disease risk grading table generator (11) is used to perform multiple grouping and iteration on the plurality of physiological data of the sample data set, and check whether a grouping result generated by each grouping and iteration satisfies a convergence condition. If the grouping result satisfies the convergence condition, a disease risk grading table is generated according to the grouping result; otherwise, the next grouping and iteration is performed; The disease risk grading table is used to perform a disease risk grading on a user to be graded to generate a grading result, and the grading result is used to calculate a risk score corresponding to each physiological parameter of the user to be graded under the grading result, and to calculate a predicted probability of the occurrence of the disease in the user to be graded corresponding to the grading result.

2. The disease risk positioning module according to claim 1, wherein: Also includes: a disease risk locator (12) connected to the disease risk grading table generator (11) to receive the disease risk grading table; The disease risk locator (12) is used to receive physiological data including each physiological parameter of the user to be graded, generate the grading result of the user to be graded according to each physiological parameter of the user to be graded using the disease risk grading table, determine a starting position of each physiological parameter of the user to be graded using the disease risk grading table, convert the starting position of each physiological parameter of the user to be graded into an adjusted position of each physiological parameter of the user to be graded according to the grading result of the user to be graded, calculate the risk score of each physiological parameter of the user to be graded corresponding to the grading result according to the adjusted position of each physiological parameter of the user to be graded, and calculate the predicted probability of the user to be graded developing a disease within a period of time according to the grading result.

3. The disease risk positioning module according to claim 1, wherein: The disease risk grading table generator (11) includes: An initialization device (111) for initializing an iteration value, a variable conversion value and a grouping number; a hyperplane clustering device (112), signal-connected to the initialization device (111), for obtaining at least one hyperplane coordinate value calculation formula based on the clustering result of the plurality of physiological data of the sample data set and at least a portion of the plurality of physiological data, and using each hyperplane coordinate value calculation formula to calculate each hyperplane coordinate value of each physiological data of at least a portion of the plurality of physiological data of the sample data set, wherein an initial clustering result of the clustering result of the plurality of physiological data of the sample data set is a preset clustering result that meets the initial clustering number; a clustering value calculation device (113), signal-connected to the hyperplane clustering device (112), for calculating a clustering value of the physiological data according to each hyperplane coordinate value of the physiological data; an iterative device (114), signal-connected to the grouping value calculation device (113), using a grouping rule to group each physiological data according to the grouping value of each physiological data, and updating the grouping result; a convergence condition checking device (116), signal-connected to the iterative device (114), for determining whether the clustering result satisfies the convergence condition; a disease risk grading table generating device (117), signal-connected to the convergence condition checking device (116), for generating the disease risk grading table based on the clustering result when the clustering result satisfies the convergence condition; and A parameter updating device (115) is signal-connected to the convergence condition checking device (116) and the hyperplane clustering device (112), and is used to update the iteration value, the variable conversion value and the number of clusters when the clustering result does not meet the convergence condition.

4. The disease risk positioning module according to claim 2, wherein: The disease risk locator (12) includes: a grading result calculation device (121) for generating the grading result of the user to be graded based on the physiological parameters of the user to be graded using the disease risk grading table; a predicted probability calculation device (122), signal-connected to the classification result calculation device (121), for calculating the predicted probability of the user to be classified developing a disease within a period of time based on the classification result; a grading adjustment device (123), signal-connected to the grading result calculation device (121), for determining a starting position of each physiological parameter of the user to be graded using the disease risk grading table, and converting the starting position of each physiological parameter of the user to be graded into an adjusted position of each physiological parameter of the user to be graded according to the grading result of the user to be graded; and A score calculation device (124) is signal-connected to the grading adjustment device (123) and calculates the risk score of each physiological parameter of the user to be graded corresponding to the grading result based on the adjusted positioning of each physiological parameter of the user to be graded.

5. The disease risk positioning module according to claim 3, wherein: The hyperplane clustering device (112) uses a support vector clustering algorithm, and further converts the multiple physiological parameters of the multiple physiological data according to the iteration value and the variable conversion value to generate at least one extended physiological parameter, and adds each of the extended physiological parameters to the multiple physiological data to thereby expand and update the multiple physiological data, and the hyperplane clustering device (112) uses at least a portion of the multiple physiological data after expansion and update and the clustering result to calculate the calculation formula of each hyperplane coordinate value.

6. The disease risk positioning module according to claim 5, characterized in that: The clustering value calculation device (113) determines a clustering value calculation formula according to the iteration value, and the clustering value calculation device (113) uses the clustering value calculation formula to calculate the clustering value of the physiological data according to each hyperplane coordinate value of the physiological data.

7. The disease risk positioning module according to claim 6, characterized in that: The iteration device (114) is used to determine the grouping rule according to the iteration value and the number of groups.

8. The disease risk positioning module according to claim 7, wherein: The convergence condition checking device (116) checks whether a risk probability of each risk level of the disease risk grading table corresponding to the clustering result satisfies the convergence condition.

9. A disease risk positioning method, characterized in that: include: receiving a sample data set including a plurality of physiological data of a plurality of users; Performing multiple grouping and iterations on the plurality of physiological data of the sample data set, and checking whether the grouping results generated by each grouping and iteration meet a convergence condition; as well as If the clustering result meets the convergence condition, a disease risk grading table is generated according to the clustering result; otherwise, the next clustering and iteration is performed; The disease risk grading table is used to perform a disease risk grading on a user to be graded to generate a grading result, and the grading result is used to calculate a risk score corresponding to each physiological parameter of the user to be graded under the grading result, and to calculate a predicted probability of the occurrence of the disease in the user to be graded corresponding to the grading result.

10. The disease risk positioning method according to claim 9, wherein: Also includes: receiving physiological data including the physiological parameters of the user to be graded; Generating the grading result of the user to be graded according to each of the physiological parameters of the user to be graded using the disease risk grading table; Determine a starting position of each physiological parameter of the user to be graded using the disease risk grading table; converting the initial position of each physiological parameter of the user to be graded into an adjusted position of each physiological parameter of the user to be graded according to the grading result of the user to be graded; Calculating the risk score of each physiological parameter of the user to be classified corresponding to the classification result according to the adjusted positioning of each physiological parameter of the user to be classified; and The predicted probability of the user to be classified developing a disease within a period of time is calculated based on the classification result.

11. The disease risk positioning method according to claim 9, wherein: The steps of performing multiple grouping and iteration on the plurality of physiological data of the sample data set include: Initialize an iteration value, a variable conversion value and a number of clusters; obtaining at least one hyperplane coordinate value calculation formula based on the clustering result of the plurality of physiological data of the sample data set and at least a portion of the plurality of physiological data, and calculating each hyperplane coordinate value of each physiological data of at least a portion of the plurality of physiological data of the sample data set using each hyperplane coordinate value calculation formula, wherein an initial clustering result of the clustering result of the plurality of physiological data of the sample data set is a predetermined clustering result that meets the initial clustering number; Calculating a clustering value of the physiological data according to each of the hyperplane coordinate values of the physiological data; using a grouping rule to group the physiological data according to the grouping value of the physiological data, and updating the grouping result; Determine whether the clustering result meets the convergence condition; When the clustering result satisfies the convergence condition, generating the disease risk grading table according to the clustering result; and When the clustering result does not satisfy the convergence condition, the iteration value, the variable conversion value and the number of clusters are updated.

12. The disease risk positioning method according to claim 11, wherein: Also includes: Converting a plurality of physiological parameters of the plurality of physiological data according to the iteration value and the variable conversion value to generate at least one extended physiological parameter, and adding each of the extended physiological parameters to the plurality of physiological data to thereby expand and update the plurality of physiological data; A support vector clustering algorithm is used to obtain a calculation formula for each hyperplane coordinate value based on at least a portion of the multiple physiological data after expansion and update and the clustering result, and each hyperplane coordinate value calculation formula is used to calculate each hyperplane coordinate value of each physiological data of at least a portion of the multiple physiological data of the sample data set.

13. The disease risk positioning method according to claim 12, wherein: A clustering value calculation formula is determined according to the iteration value, and the clustering value of the physiological data is calculated according to each hyperplane coordinate value of the physiological data using the clustering value calculation formula.

14. The disease risk positioning method according to claim 13, wherein: The grouping rule is determined according to the iteration value and the number of groups.

15. The disease risk positioning method according to claim 14, wherein: The checking of whether the clustering result satisfies the convergence condition is to check whether a risk probability of each risk level of the disease risk grading table corresponding to the clustering result satisfies the convergence condition.

Images