Dynamic health evaluation method and system based on time series physical measurement data

Abstract

The present invention provides a dynamic health evaluation method and system based on time-series physical measurement data, and relates to the field of medical health evaluation technology. The present invention first determines the evaluation index attribute set of the original physical measurement data set, and differentiates and classifies the original physical measurement data set based on the evaluation index attribute set; then constructs a stepwise regression equation for the evaluation index attribute subset and the data subset at each moment; finally, based on the time series characteristics of the physical measurement data set, the stepwise regression equations of all moments of the data set are merged using the weighted sliding average method to obtain the final weighted regression equation. The final weighted regression equation can then be used to perform dynamic health evaluation on different patients based on the time series physical measurement data. The present invention solves the problem of the influence on the health evaluation results caused by the single source of physical measurement data information and strong subjectivity, and also solves the problem of poor generalization of existing dynamic health evaluation technology.

Description

Technical Field

[0001] The present invention relates to the technical field of medical health evaluation, and in particular to a dynamic health evaluation method and system based on time-series body measurement data. Background Art

[0002] In the medical field, many medical physical test data exist in the form of time series, such as continuous monitoring data of blood sugar, blood pressure, blood lipids, etc. related to chronic diseases, and data of heartbeat, pulse and body position changes continuously monitored by smart wearable devices. For health time series data with strong scale and structured characteristics, different analysis methods can be used to explore potential patterns and valuable information, and then make valuable predictions about the future trend of health time series data based on the potential laws discovered, which can provide valuable data support for patients' health evaluation and prognosis, which is of great significance.

[0003] At present, most health assessment technologies are based on the characteristics of time series data and related health information contained in health time series data, and achieve health assessment by adopting methods such as feature scoring, health degree evaluation or self-assessment health rating sample form (SRHMS) and time series models; there are also health assessment methods that give corresponding health assessments by enumerating features or selecting scores on health assessment scales, and there are also health assessment methods that use SVM algorithms and BP neural networks or use big data analysis algorithms to analyze and obtain corresponding health assessments to predict the patient's condition.

[0004] However, although these methods can provide scientific basis and data support for doctors' diagnosis to a certain extent, the following problems still exist: on the one hand, the subjective survey method and physiological parameter detection method adopted by the existing human health evaluation technology have subjective and single information sources, resulting in inaccurate health evaluation results; on the other hand, the original technical solution only analyzes a certain medical scenario or disease at present, and it is difficult to screen and reuse different characteristics. The health evaluation model lacks usability and generalization. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In view of the deficiencies in the prior art, the present invention provides a dynamic health evaluation method and system based on time-series body measurement data, which solves the problems of inaccurate evaluation results and poor generalization in the prior health evaluation methods.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] In a first aspect, the present invention first proposes a dynamic health evaluation method based on time-series physical measurement data, the method comprising:

[0010] Acquire an evaluation index attribute set based on the acquired original body measurement data set, divide the evaluation index attribute set into a plurality of evaluation index attribute subsets based on the strong correlation between the evaluation index attributes; differentiate and classify the original body measurement data set into a plurality of data subsets based on the evaluation index attribute subsets;

[0011] A stepwise regression equation is constructed based on the attribute subset of the evaluation index and the data subset corresponding to each moment of the original physical measurement data set;

[0012] Based on the time series characteristics of the original physical measurement data set, the stepwise regression equations at all times are merged using the weighted moving average method to obtain the final weighted regression equation;

[0013] Dynamic health assessment is performed using the final weighted regression equation based on the time-series physical measurement data to be evaluated.

[0014] Preferably, determining an evaluation index attribute set of the time series physical measurement data, and acquiring a segmented data set based on the evaluation index attribute set includes:

[0015] S11, determining the evaluation index attributes of the original physical measurement data required for constructing a stepwise regression equation based on the distribution characteristics and correlation of the original physical measurement data set, and forming an evaluation index attribute set, and then dividing the evaluation index attribute set into a plurality of evaluation index attribute subsets according to the strong correlation between the evaluation index attributes;

[0016] S12, calculating the quantitative results of the evaluation index attributes in the evaluation index attribute subset;

[0017] S13. Based on the quantification result, the original body measurement data set is classified to obtain a plurality of data subsets.

[0018] Preferably, the evaluation index attributes include height and weight;

[0019] The quantitative results of the evaluation index attributes include BMI index:

[0020]

[0021] in, represents weight; h represents height.

[0022] Preferably, the stepwise regression equation is constructed based on the evaluation index attribute subset and the data subset corresponding to each moment of the original physical measurement data set, including:

[0023] The stepwise regression equation is constructed with the attribute subset of the evaluation index at each moment as the potential item. The specific process of constructing the stepwise regression equation is as follows:

[0024] S21, constructing the initial stepwise regression equation at the current moment;

[0025] S22, if there is a potential term that has not been added to the model, and the p-value of its F statistic is less than the set addition threshold, then add the term with the smallest p-value of the F statistic among all potential terms that meet the requirements into the current stepwise regression equation, and repeat this step; otherwise, jump to S23;

[0026] S23, if there is a potential item added to the model, and its F statistic p-value is greater than the set deletion threshold, then the item with the largest F statistic p-value among all potential items that meet the requirements is removed from the current stepwise regression equation, and jump to S22; otherwise, end;

[0027] S24, obtaining a stepwise regression equation at the current moment;

[0028] S25. Repeat the above S1-S4 to construct a stepwise regression equation for each moment except the current moment.

[0029] Preferably, the stepwise regression equations at all moments are merged by weighted moving average method based on the time series characteristics of the original body measurement data set to obtain the final weighted regression equation, which includes:

[0030] S31, based on the stepwise regression equation of the original body measurement data set at each moment, obtaining the regression equation matrix at each moment;

[0031] S32, constructing a linear regression model at each moment based on the regression equation matrix of the original body measurement data set at each moment;

[0032] S33. Using the weighted moving average method, the linear regression models of the original body measurement data set at each moment are combined to obtain the final weighted regression equation.

[0033] In a second aspect, the present invention further proposes a dynamic health evaluation system based on time-series physical measurement data, the system comprising:

[0034] A data set differentiation module, used for obtaining an evaluation index attribute set based on the acquired original physical measurement data set, dividing the evaluation index attribute set into a plurality of evaluation index attribute subsets based on the strong correlation between the evaluation index attributes; and differentiating and classifying the original physical measurement data set into a plurality of data subsets based on the evaluation index attribute subsets;

[0035] A stepwise regression equation construction module is used to construct a stepwise regression equation based on the evaluation index attribute subset and the data subset corresponding to each moment of the original physical measurement data set;

[0036] A weighted regression equation construction module is used to merge the stepwise regression equations at all times using a weighted sliding average method based on the time series characteristics of the original body measurement data set to obtain the final weighted regression equation;

[0037] The dynamic health evaluation module is used to perform dynamic health evaluation based on the time series physical measurement data to be evaluated using the final weighted regression equation.

[0038] Preferably, the data set differentiation module obtains an evaluation indicator attribute set based on the acquired original physical measurement data set, and divides the evaluation indicator attribute set into a plurality of evaluation indicator attribute subsets based on the strong correlation between the evaluation indicator attributes; and classifying the original physical measurement data set into a plurality of data subsets based on the evaluation indicator attribute subsets includes:

[0039] S11, determining the evaluation index attributes of the original physical measurement data required for constructing a stepwise regression equation based on the distribution characteristics and correlation of the original physical measurement data set, and forming an evaluation index attribute set, and then dividing the evaluation index attribute set into a plurality of evaluation index attribute subsets according to the strong correlation between the evaluation index attributes;

[0040] S12, calculating the quantitative results of the evaluation index attributes in the evaluation index attribute subset;

[0041] S13. Based on the quantification result, the original body measurement data set is classified to obtain a plurality of data subsets.

[0042] Preferably, the evaluation index attributes include height and weight;

[0043] The quantitative results of the evaluation index attributes include BMI index:

[0044]

[0045] in, represents weight; h represents height.

[0046] Preferably, the stepwise regression equation is constructed based on the evaluation index attribute subset and the data subset corresponding to each moment of the original physical measurement data set, including:

[0047] The stepwise regression equation is constructed with the attribute subset of the evaluation index at each moment as the potential item. The specific process of constructing the stepwise regression equation is as follows:

[0048] S21, constructing the initial stepwise regression equation at the current moment;

[0049] S22, if there is a potential term that has not been added to the model, and the p-value of its F statistic is less than the set addition threshold, then add the term with the smallest p-value of the F statistic among all potential terms that meet the requirements into the current stepwise regression equation, and repeat this step; otherwise, jump to S23;

[0050] S23, if there is a potential item added to the model, and its F statistic p-value is greater than the set deletion threshold, then the item with the largest F statistic p-value among all potential items that meet the requirements is removed from the current stepwise regression equation, and jump to S22; otherwise, end;

[0051] S24, obtaining a stepwise regression equation at the current moment;

[0052] S25. Repeat the above S1-S4 to construct a stepwise regression equation for each moment except the current moment.

[0053] Preferably, the weighted regression equation construction module merges the stepwise regression equations at all times using the weighted sliding average method based on the time series characteristics of the original body measurement data set to obtain the final weighted regression equation, which includes:

[0054] S31, based on the stepwise regression equation of the original body measurement data set at each moment, obtaining the regression equation matrix at each moment;

[0055] S32, constructing a linear regression model at each moment based on the regression equation matrix of the original body measurement data set at each moment;

[0056] S33. Using the weighted moving average method, the linear regression models of the original body measurement data set at each moment are combined to obtain the final weighted regression equation.

[0057] (III) Beneficial effects

[0058] The present invention provides a dynamic health evaluation method and system based on time series physical measurement data. Compared with the prior art, it has the following beneficial effects:

[0059] 1. The present invention first determines the evaluation index attribute set of the time series physical measurement data, and obtains the partitioned data set based on the evaluation index attribute set, and then constructs a stepwise regression equation based on the evaluation index attribute subset and the data subset at each moment; finally, based on the time series characteristics, the stepwise regression equations of all moments of the data set are merged using the weighted sliding average method to obtain the final weighted regression equation, and the final weighted regression equation is used to perform dynamic health evaluation based on the time series physical measurement data. The present invention establishes the evaluation index attribute set and partitions the data set based on a wide range of medical physical measurement data, and screens and reuses the establishment of the evaluation index attribute set and the partitioning of the data set through a stepwise regression algorithm, thereby significantly improving the usability and versatility of the model itself, and at the same time improving the accuracy of the evaluation model compared to the traditional model;

[0060] 2. The present invention determines the evaluation index attribute set of the time series physical measurement data, and differentiates and classifies the time series physical measurement data set based on the attribute set, and then establishes a dynamic health evaluation model suitable for different populations and different characteristics based on the evaluation index attribute set and the data set, thereby realizing multi-level segmentation and integration of the sample set, so that the dynamic health evaluation of the model realizes the flexibility and robustness of being built according to the characteristics of the population and changing according to the characteristics of the sample;

[0061] 3. The present invention integrates the temporal characteristics of time-series physical measurement data and reflects the dynamic characteristics of physical measurement data. Compared with traditional health evaluation models, the present invention fully considers the influence of time-series factors on evaluation results on the basis of classical fitting and evaluation, and integrates the characteristics of time-series physical measurement data sets in different dimensions through weighted sliding average. Therefore, the temporal characteristics are used to add more comprehensive dynamic characteristics to the model without losing the advantages of the original evaluation method, which is more in line with the actual situation of time-series physical measurement data and makes the evaluation results more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0063] Figure 1 This is a flow chart of a dynamic health assessment method based on time-series body measurement data in an embodiment of the present invention;

[0064] Figure 2 It is a schematic diagram of the segmentation of health indicators corresponding to the BMI index in an embodiment of the present invention. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are 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.

[0066] The embodiments of the present application provide a dynamic health assessment method and system based on time-series physical measurement data, thereby solving the problems of inaccurate assessment results and poor generalization in existing health assessment methods, and achieving the generalization and robustness of the health assessment method based on population characteristics and adapted to sample characteristics.

[0067] The technical solution in the embodiment of the present application is to solve the above technical problems, and the overall idea is as follows:

[0068] In order to solve the problem that the existing health evaluation technology is affected by the subjective and single source of physical measurement data information, and at the same time, in order to solve the problem that the existing technology can only analyze a certain specific medical scenario or specific disease at present, and cannot screen and reuse different features, the present invention first determines the evaluation index attribute set of the original physical measurement data set, and differentiates and classifies the original physical measurement data set based on the evaluation index attribute set to obtain a data subset; then, a stepwise regression equation is constructed for the evaluation index attribute subset and the data subset at each moment; finally, based on the time series characteristics of the original physical measurement data, the stepwise regression equations of all moments of the original physical measurement data set are merged using the weighted sliding average method to obtain the final weighted regression equation. In this way, the final weighted regression equation can be used to perform dynamic health assessment on different patients based on the time series physical measurement data to be evaluated.

[0069] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0070] Embodiment 1:

[0071] In the first aspect, the present invention first proposes a dynamic health assessment method based on time series physical measurement data, see Figure 1 , the method comprising:

[0072] S1. acquiring an evaluation index attribute set based on the acquired original body measurement data set, dividing the evaluation index attribute set into a plurality of evaluation index attribute subsets based on the strong correlation between the evaluation index attributes; and differentiating and classifying the original body measurement data set into a plurality of data subsets based on the evaluation index attribute subsets;

[0073] S2, constructing a stepwise regression equation based on the evaluation index attribute subset and data subset corresponding to each moment of the original physical measurement data set;

[0074] S3, based on the time series characteristics of the original physical measurement data set, the stepwise regression equations at all times are merged using the weighted moving average method to obtain the final weighted regression equation;

[0075] S4. Perform dynamic health assessment based on the time-series physical measurement data to be evaluated using the final weighted regression equation.

[0076] It can be seen that the present invention first determines the evaluation index attribute set of the time series physical measurement data, and obtains the partitioned data set based on the evaluation index attribute set, and then constructs a stepwise regression equation based on the evaluation index attribute subset and the data subset at each moment; finally, based on the time series characteristics, the stepwise regression equations of all moments of the data set are merged using the weighted sliding average method to obtain the final weighted regression equation, and the final weighted regression equation is used to perform dynamic health evaluation based on the time series physical measurement data. The present invention establishes the evaluation index attribute set and divides and classifies the data set based on a wide range of medical physical measurement data, and screens and reuses the establishment of the evaluation index attribute set and the division and classification of the data set through the stepwise regression algorithm, thereby significantly improving the availability and versatility of the model itself, and at the same time improving the accuracy of the evaluation model compared to the traditional model.

[0077] The following takes height and weight as evaluation indicators in the evaluation indicator attribute set as an example, combined with the explanation of the specific steps S1-S3, to describe in detail the implementation process of an embodiment of the present invention.

[0078] S1. Acquire an evaluation index attribute set based on the acquired original body measurement data set, divide the evaluation index attribute set into several evaluation index attribute subsets based on the strong correlation between the evaluation index attributes; and differentiate and classify the original body measurement data set into several data subsets based on the evaluation index attribute subsets.

[0079] Obtain the original physical measurement data, and select a part of it as a sample to form the original physical measurement data data set. The data set contains the physical measurement items and physical measurement data values ​​of the original physical measurement data, as well as related time series information, etc. For the original physical measurement data set, first determine the evaluation index attribute set of the original physical measurement data set (the evaluation index will be used as the regression index for the subsequent construction of the stepwise regression model), and then based on the determined evaluation index attribute set, combined with the distribution characteristics and correlation of different physical measurement data evaluation indicators, classify the original physical measurement data set to obtain the divided data subsets. The specific steps are as follows:

[0080] S11. Determine the evaluation index attributes of the original physical measurement data required for constructing a stepwise regression equation based on the distribution characteristics and correlation of the original physical measurement data set, and form an evaluation index attribute set. Then, divide the evaluation index attribute set into several evaluation index attribute subsets according to the strong correlation between the evaluation index attributes.

[0081] There are many kinds of evaluation indicators, which can be determined according to actual needs. For example, height, weight, blood sugar, blood pressure, blood lipids, heart rate, pulse, etc. can all be used as evaluation indicators. In particular, according to the strong correlation that may exist within each regression independent variable and the discrimination criteria in the relevant medical field, a complete partition of the evaluation indicator attribute set S is made in the initial stage of regression data selection, that is, S is divided into n evaluation indicator attribute subsets, namely {S1, S2, ..., S n The evaluation index attribute subsets are multiple subsets obtained by dividing the evaluation index attributes according to the medical field discrimination of each physical measurement item. The n evaluation index attribute subsets must meet the following conditions at the same time:

[0082] And S1∪S2∪...∪S n =S

[0083] In this embodiment, height and weight are used as two evaluation index attributes. These two evaluation index attributes form an evaluation index attribute subset, which is defined as S j , then S j = {height, weight}, and

[0084] According to the set evaluation index attribute subset, a corresponding data subset vector is imported, and the data subset vector is a data sequence in a certain specified order according to the set evaluation index attribute subset.

[0085] S12, calculate the quantitative results of the evaluation index attributes in the evaluation index attribute subset. Calculate the evaluation index attribute subset S j In this embodiment, the quantitative results corresponding to the evaluation index attributes of height and weight are calculated. Specifically, the corresponding BMI index is calculated based on the height and weight attributes:

[0086]

[0087] in, is the weight in kilograms; h is the height in meters.

[0088] S13. Based on the quantification result, the original body measurement data set is classified to obtain a plurality of data subsets.

[0089] In this embodiment, the data set is classified based on the above BMI calculation results. Figure 2 , the data set can be divided into four types according to the size of the BMI index: "underweight", "normal weight", "overweight" and "obese", and the corresponding data sets are recorded as N1, N2, N3, and N4 respectively.

[0090] Each evaluation index attribute in the data set is used as a regression independent variable, and the quantified result corresponding to the evaluation index attribute is used as a regression dependent variable as the source data for the following regression analysis. In this embodiment, the height and weight are used as regression independent variables, and the calculated BMI value is used as the regression dependent variable.

[0091] In addition, for the data set, each evaluation index attribute subset of the data set at k time points (i.e., starting from the current observation time point k (including the current time point k) and selecting k time points in the past) is obtained respectively, and a total of k groups of corresponding data are selected, which are recorded as {T1, T2, ..., T k}.

[0092] S2. Construct a stepwise regression equation based on the evaluation index attribute subset and data subset corresponding to each moment of the original physical measurement data set.

[0093] When calculating the stepwise regression equation corresponding to a certain moment in the data set, the differentiated data subset at that moment and its corresponding evaluation index attribute subset are used to establish a stepwise regression equation with a single differentiated data subset and a single evaluation index attribute subset as units using the stepwise regression method. Specifically, taking the construction of the stepwise regression equation at time i as an example, for the data subset N at time i i (i∈{1,2,3,4}) and the corresponding evaluation index attribute subset S at time i j (j∈{1,2,...,n}) constructs a stepwise regression equation.

[0094] The stepwise regression algorithm is a systematic method for adding and removing terms from a multilinear model based on statistical significance in the regression. The method starts with an initial model and then compares the explanatory power of the gradually increasing and decreasing models, and makes updates and decisions based on this. At each step of the update, the p-value of the F statistic is calculated to test the possible effect of the current potential term on the model. If the current potential term has not yet been added to the model, the following null hypothesis is first set:

[0095] After the potential term is added to the current model, the corresponding coefficient is zero. Then, if there is sufficient evidence to reject the null hypothesis, the term is added to the current model;

[0096] On the contrary, if the current potential term already exists in the current model, the corresponding null hypothesis is that its coefficient is zero. If there is no sufficient evidence to reject the null hypothesis, the potential term is removed from the model.

[0097] Specifically, the algorithm is explained as follows:

[0098] Algorithm 1: Stepwise regression model algorithm

[0099] Input: Data subset N at time i i And the current evaluation index attribute subset S j ;

[0100] Result: Stepwise regression equation

[0101] Among them, s j Represents the evaluation index attribute subset S j The cardinality of s j =|S j |,α k represents the weight of the kth evaluation index attribute in the evaluation index attribute subset; S j [k] represents the evaluation index attribute subset S j The kth element of y ij Represents the total records of the weights of each evaluation indicator attribute in the evaluation indicator attribute subset.

[0102] The construction process of the stepwise regression equation is as follows:

[0103] S21. Construct the initial stepwise regression equation at the current moment.

[0104] S22. If there is a potential term that has not been added to the model, and the p-value of its F statistic is less than the set addition threshold (that is, if it is added to the model, there is sufficient evidence to believe that its corresponding coefficient is not zero), then add the term with the smallest p-value of the F statistic among all potential terms that meet the requirements to the current stepwise regression equation, and repeat this step; otherwise, S23.

[0105] S23. If there is a potential term that has been added to the model and the p-value of its F statistic is greater than the set deletion threshold (that is, there is insufficient evidence to deny the hypothesis that its corresponding coefficient is zero), then the term with the largest p-value of the F statistic among all potential terms that meet the requirements is removed from the current stepwise regression equation, S22; otherwise, end.

[0106] S24. Obtain the stepwise regression equation at the current moment.

[0107] S25. Repeat the above S1-S4 to construct a stepwise regression equation for each moment except the current moment.

[0108] It is worth noting that the method can build different models from the same set of potential terms, depending on the terms included in the initial model and the order in which terms are moved in and out. The method terminates when no single step improves the model. However, there is no guarantee that a different initial model or a different sequence of steps would not lead to a better fit. In this sense, the stepwise model is locally optimal, but may not be globally optimal.

[0109] S3. Based on the time series characteristics of the original physical measurement data set, the stepwise regression equations at all times are merged using the weighted moving average method to obtain the final weighted regression equation.

[0110] S31. Based on the stepwise regression equation of the original body measurement data set at each moment, a regression equation matrix at each moment is obtained.

[0111] The stepwise regression equations corresponding to all moments of the original body measurement data set constitute a stepwise regression equation group, and a stepwise regression equation matrix can be constructed based on the stepwise regression equation group.

[0112] The construction of the stepwise regression matrix and the solution process of the linear regression model are as follows:

[0113] Based on the above stepwise regression equations, we can get the corresponding regression equation matrix M:

[0114]

[0115] S32. Based on the regression equation matrix of the original body measurement data set at each moment, a linear regression model at each moment is constructed.

[0116] Based on the above regression matrix, the following linear regression model is constructed:

[0117] y=Γ*[β1β2β3β4] T *M X

[0118] Where X is the set of independent regression variables; Γ is the parameter matrix of linear regression; T represents the transpose of the matrix; β i (i∈{1,2,3,4}) is a characterization factor used to characterize whether X belongs to the data subset N at time i i , β i The following constraints are met:

[0119]

[0120]

[0121] And M XIt is calculated by inputting independent variable X through regression equation matrix M. In addition, element-by-element matrix multiplication is set as the calculation scheme of the above regression equation, and element-by-element matrix multiplication is expressed by the formula as follows:

[0122]

[0123] In the above parameter solving process, the gradient descent method is used to regress the parameters in Γ. The specific steps are as follows:

[0124] 1) Select Γ0 as the initial value and calculate the gradient of the loss function F at this point It can be seen that the function F is in the opposite direction of the gradient at this point The fastest decline;

[0125] 2) If you select This is true for γ>0 when it is a sufficiently small real value, then F(Γ1)<F(Γ0);

[0126] 3) Repeat the above descending steps and adjust the iteration step size to make the sequence {Γ i}converges to the corresponding extreme value.

[0127] Calculate the corresponding linear regression model according to the stepwise regression equation matrix, and use the above method to calculate T1, T2, ..., T k Establish a linear regression model and record it as F1, F2, ..., F k .

[0128] S33, using the weighted moving average method, and based on the time series characteristics of the data, the linear regression model of each moment of the original body measurement data set is merged to obtain the final weighted regression equation. The specific steps are as follows:

[0129] The linear regression models obtained at different times of the data set are combined using the weighted moving average method to obtain the final weighted regression equation. When using the weighted moving average method to set the relevant weights corresponding to the regression equation at each moment, the principle of the closer to the current moment, the greater the weight is, and the weight distribution formula is established in sequence, in which the weight gradually decreases over time. At the same time, the weight distribution fully adopts and respects the opinions of experts, and makes appropriate adjustments to the weights to ensure the objectivity of the results.

[0130] Based on the above linear regression model group, a weighted moving average equation is established:

[0131] F=w1F1+w2F2+...+w k F k

[0132] Among them, the weights in the sliding average equation are calculated and normalized according to the following formula.

[0133]

[0134] w i Represents the weight assigned to the i-th linear regression equation.

[0135] The weight ratio of the current weighted moving average equation meets the exponential decay condition. In practical applications, a linear decreasing decay strategy is adopted for different time periods and the distribution characteristics of the corresponding data sets. Specifically, after obtaining the time distribution of known physical measurement features, a comparative evaluation can be performed through the following different weight calculation modes. The weight distribution applicable to this technical solution is shown in Table 1 below.

[0136]

[0137] Based on this, a weighted moving average equation of the corresponding decay form can be established, in which the training mode of the hyperparameters can be processed by referring to the above-mentioned gradient descent method. The expected risk minimization principle is adopted to evaluate each average equation, so as to establish the final weighted regression equation.

[0138] S4. Perform dynamic health assessment based on the time-series physical measurement data to be evaluated using the final weighted regression equation.

[0139] According to different groups of people with different characteristics, their unique time-series physical test data are used, different indicators are used to evaluate attributes, and the final weighted regression equation is used to perform dynamic health assessment.

[0140] At this point, the entire process of the dynamic health evaluation method based on time-series body measurement data of the present invention is completed.

[0141] Embodiment 2:

[0142] In a second aspect, the present invention further provides a dynamic health evaluation system based on time series physical measurement data, the system comprising:

[0143] A data set differentiation module, used for obtaining an evaluation index attribute set based on the acquired original physical measurement data set, dividing the evaluation index attribute set into a plurality of evaluation index attribute subsets based on the strong correlation between the evaluation index attributes; and differentiating and classifying the original physical measurement data set into a plurality of data subsets based on the evaluation index attribute subsets;

[0144] A stepwise regression equation construction module is used to construct a stepwise regression equation based on the evaluation index attribute subset and the data subset corresponding to each moment of the original physical measurement data set;

[0145] A weighted regression equation construction module is used to merge the stepwise regression equations at all times using a weighted sliding average method based on the time series characteristics of the original body measurement data set to obtain the final weighted regression equation;

[0146] The dynamic health evaluation module is used to perform dynamic health evaluation based on the time series physical measurement data to be evaluated using the final weighted regression equation.

[0147] Optionally, the data set differentiation module obtains an evaluation indicator attribute set based on the acquired original physical measurement data set, and divides the evaluation indicator attribute set into a plurality of evaluation indicator attribute subsets based on a strong correlation between the evaluation indicator attributes; and differentiating and classifying the original physical measurement data set into a plurality of data subsets based on the evaluation indicator attribute subsets includes:

[0148] S11, determining the evaluation index attributes of the original physical measurement data required for constructing a stepwise regression equation based on the distribution characteristics and correlation of the original physical measurement data set, and forming an evaluation index attribute set, and then dividing the evaluation index attribute set into a plurality of evaluation index attribute subsets according to the strong correlation between the evaluation index attributes;

[0149] S12, calculating the quantitative results of the evaluation index attributes in the evaluation index attribute subset;

[0150] S13. Based on the quantification result, the original body measurement data set is classified to obtain a plurality of data subsets.

[0151] Preferably, the evaluation index attributes include height and weight;

[0152] The quantitative results of the evaluation index attributes include BMI index:

[0153]

[0154] in, represents weight; h represents height.

[0155] Optionally, constructing a stepwise regression equation based on the evaluation index attribute subset and the data subset corresponding to each moment of the original physical measurement data set includes:

[0156] The stepwise regression equation is constructed with the attribute subset of the evaluation index at each moment as the potential item. The specific process of constructing the stepwise regression equation is as follows:

[0157] S21, constructing the initial stepwise regression equation at the current moment;

[0158] S22, if there is a potential term that has not been added to the model, and the p-value of its F statistic is less than the set addition threshold, then add the term with the smallest p-value of the F statistic among all potential terms that meet the requirements into the current stepwise regression equation, and repeat this step; otherwise, jump to S23;

[0159] S23, if there is a potential item added to the model, and its F statistic p-value is greater than the set deletion threshold, then the item with the largest F statistic p-value among all potential items that meet the requirements is removed from the current stepwise regression equation, and jump to S22; otherwise, end;

[0160] S24, obtaining a stepwise regression equation at the current moment;

[0161] S25. Repeat the above S1-S4 to construct a stepwise regression equation for each moment except the current moment.

[0162] Optionally, the weighted regression equation construction module merges the stepwise regression equations at all times using a weighted sliding average method based on the time series characteristics of the original body measurement data set to obtain the final weighted regression equation, including:

[0163] S31, based on the stepwise regression equation of the original body measurement data set at each moment, obtaining the regression equation matrix at each moment;

[0164] S32, constructing a linear regression model at each moment based on the regression equation matrix of the original body measurement data set at each moment;

[0165] S33. Using the weighted moving average method, the linear regression models of the original body measurement data set at each moment are combined to obtain the final weighted regression equation.

[0166] It can be understood that the dynamic health evaluation system based on time-series physical measurement data provided in the embodiment of the present invention corresponds to the above-mentioned dynamic health evaluation method based on time-series physical measurement data. The explanations, examples, beneficial effects and other parts of the relevant contents can refer to the corresponding contents in the dynamic health evaluation method based on time-series physical measurement data, and will not be repeated here.

[0167] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0168] 1. The present invention first determines the evaluation index attribute set of the time series physical measurement data, and obtains the partitioned data set based on the evaluation index attribute set, and then constructs a stepwise regression equation based on the evaluation index attribute subset and the data subset at each moment; finally, based on the time series characteristics, the stepwise regression equations of all moments of the data set are merged using the weighted sliding average method to obtain the final weighted regression equation, and the final weighted regression equation is used to perform dynamic health evaluation based on the time series physical measurement data. The present invention establishes the evaluation index attribute set and partitions the data set based on a wide range of medical physical measurement data, and screens and reuses the establishment of the evaluation index attribute set and the partitioning of the data set through a stepwise regression algorithm, thereby significantly improving the usability and versatility of the model itself, and at the same time improving the accuracy of the evaluation model compared to the traditional model;

[0169] 2. The present invention determines the evaluation index attribute set of the time series physical measurement data, and differentiates and classifies the time series physical measurement data set based on the attribute set, and then establishes a dynamic health evaluation model suitable for different populations and different characteristics based on the evaluation index attribute set and the data set, thereby realizing multi-level segmentation and integration of the sample set, so that the dynamic health evaluation of the model realizes the flexibility and robustness of being built according to the characteristics of the population and changing according to the characteristics of the sample;

[0170] 3. The present invention integrates the temporal characteristics of time-series physical measurement data and reflects the dynamic characteristics of physical measurement data. Compared with traditional health evaluation models, the present invention fully considers the influence of time-series factors on evaluation results on the basis of classical fitting and evaluation, and integrates the characteristics of time-series physical measurement data sets in different dimensions through weighted sliding average. Therefore, the temporal characteristics are used to add more comprehensive dynamic characteristics to the model without losing the advantages of the original evaluation method, which is more in line with the actual situation of time-series physical measurement data and makes the evaluation results more accurate.

[0171] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0172] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dynamic health assessment method based on time series physical measurement data, characterized in that: The method comprises: Acquire an evaluation index attribute set based on the acquired original body measurement data set, divide the evaluation index attribute set into a plurality of evaluation index attribute subsets based on the strong correlation between the evaluation index attributes; differentiate and classify the original body measurement data set into a plurality of data subsets based on the evaluation index attribute subsets; A stepwise regression equation is constructed based on the attribute subset of the evaluation index and the data subset corresponding to each moment of the original physical measurement data set; Based on the time series characteristics of the original physical measurement data set, the stepwise regression equations at all times are merged using the weighted moving average method to obtain the final weighted regression equation; Based on the time series physical test data to be evaluated, dynamic health evaluation is performed using the final weighted regression equation; The stepwise regression equation is constructed based on the evaluation index attribute subset and the data subset corresponding to each moment of the original physical measurement data set, including: The stepwise regression equation is constructed with the attribute subset of the evaluation index at each moment as the potential item. The specific process of constructing the stepwise regression equation is as follows: S21, constructing the initial stepwise regression equation at the current moment; S22, if there is a potential term that has not been added to the model, and the p-value of its F statistic is less than the set addition threshold, then add the term with the smallest p-value of the F statistic among all potential terms that meet the requirements into the current stepwise regression equation, and repeat this step; otherwise, jump to S23; S23, if there is a potential item added to the model, and its F statistic p-value is greater than the set deletion threshold, then the item with the largest F statistic p-value among all potential items that meet the requirements is removed from the current stepwise regression equation, and jump to S22; otherwise, end; S24, obtaining a stepwise regression equation at the current moment; S25, repeat the above S1-S4 to construct a stepwise regression equation for each moment except the current moment; The stepwise regression equations at all moments are merged using the weighted moving average method based on the time series characteristics of the original body measurement data set to obtain the final weighted regression equation, which includes: S31, based on the stepwise regression equation of the original body measurement data set at each moment, obtaining the regression equation matrix at each moment; S32, constructing a linear regression model at each moment based on the regression equation matrix of the original body measurement data set at each moment; S33. Using the weighted moving average method, the linear regression models of the original body measurement data set at each moment are combined to obtain the final weighted regression equation.

2. The method according to claim 1, characterized in that Determining an evaluation index attribute set of the time series physical measurement data, and acquiring a partitioned data set based on the evaluation index attribute set includes: S11. Determine the evaluation index attributes of the original physical measurement data required for constructing a stepwise regression equation based on the distribution characteristics and correlation of the original physical measurement data set, and form an evaluation index attribute set, and then divide the evaluation index attribute set into a plurality of evaluation index attribute subsets according to the strong correlation between the evaluation index attributes; S 12. Calculate the quantitative results of the evaluation indicator attributes in the evaluation indicator attribute subset; S 13. Based on the quantification result, the original body measurement data set is classified to obtain a plurality of data subsets.

3. The method according to claim 2, characterized in that The evaluation index attributes include height and weight; The quantitative results of the evaluation index attributes include BMI index: in, represents weight; h represents height.

4. A dynamic health assessment system based on time series physical measurement data, characterized in that: The system comprises: A data set differentiation module, used for obtaining an evaluation index attribute set based on the acquired original physical measurement data set, dividing the evaluation index attribute set into a plurality of evaluation index attribute subsets based on the strong correlation between the evaluation index attributes; and differentiating and classifying the original physical measurement data set into a plurality of data subsets based on the evaluation index attribute subsets; A stepwise regression equation construction module is used to construct a stepwise regression equation based on the evaluation index attribute subset and the data subset corresponding to each moment of the original physical measurement data set; A weighted regression equation construction module is used to merge the stepwise regression equations at all times using a weighted sliding average method based on the time series characteristics of the original body measurement data set to obtain the final weighted regression equation; A dynamic health evaluation module is used to perform dynamic health evaluation based on the time series physical measurement data to be evaluated using the final weighted regression equation; The stepwise regression equation is constructed based on the evaluation index attribute subset and the data subset corresponding to each moment of the original physical measurement data set, including: The stepwise regression equation is constructed with the attribute subset of the evaluation index at each moment as the potential item. The specific process of constructing the stepwise regression equation is as follows: S21, constructing the initial stepwise regression equation at the current moment; S22, if there is a potential term that has not been added to the model, and the p-value of its F statistic is less than the set addition threshold, then add the term with the smallest p-value of the F statistic among all potential terms that meet the requirements into the current stepwise regression equation, and repeat this step; otherwise, jump to S23; S23, if there is a potential item added to the model, and its F statistic p-value is greater than the set deletion threshold, then the item with the largest F statistic p-value among all potential items that meet the requirements is removed from the current stepwise regression equation, and jump to S22; otherwise, end; S24, obtaining a stepwise regression equation at the current moment; S25, repeat the above S1-S4 to construct a stepwise regression equation for each moment except the current moment; The weighted regression equation construction module merges the stepwise regression equations at all times using the weighted sliding average method based on the time series characteristics of the original body measurement data set to obtain the final weighted regression equation, including: S31, based on the stepwise regression equation of the original body measurement data set at each moment, obtaining the regression equation matrix at each moment; S32, constructing a linear regression model at each moment based on the regression equation matrix of the original body measurement data set at each moment; S33. Using the weighted moving average method, the linear regression models of the original body measurement data set at each moment are combined to obtain the final weighted regression equation.

5. The system according to claim 4, characterized in that The data set differentiation module obtains an evaluation index attribute set based on the acquired original body measurement data set, and divides the evaluation index attribute set into a plurality of evaluation index attribute subsets based on the strong correlation between the evaluation index attributes; Differentiating and classifying the original physical measurement data set into several data subsets based on the evaluation index attribute subset includes: S11. Determine the evaluation index attributes of the original physical measurement data required for constructing a stepwise regression equation based on the distribution characteristics and correlation of the original physical measurement data set, and form an evaluation index attribute set, and then divide the evaluation index attribute set into a plurality of evaluation index attribute subsets according to the strong correlation between the evaluation index attributes; S 12. Calculate the quantitative results of the evaluation indicator attributes in the evaluation indicator attribute subset; S 13. Based on the quantification result, the original body measurement data set is classified to obtain a plurality of data subsets.

6. The system according to claim 5, characterized in that The evaluation index attributes include height and weight; The quantitative results of the evaluation index attributes include BMI index: in, represents weight; h represents height.

Images