Method and system for indexing biological functions based on health data

Abstract

A method and system for quantifying bio-function based on health data are disclosed. The method for quantifying bio-function based on health data according to the present invention comprises: receiving a plurality of health data from a health information database; calculating a bio-function index and a detailed function index for a specific user through a prediction model modeled by deep learning the plurality of health data; and outputting the bio-function index and the detailed function index to the user.

Description

Technology Field

[0001] The present invention relates to a method and system for quantifying biological functions based on health data, which calculates the degree of aging of biological functions relative to age based on health data. Background Technology

[0002] Registration No. 10-2686984 (2024.07.16), "System, method, and program for quantifying medical examination data based on data volume amplification"

[0003] Registration No. 10-2686984 discloses a screening data classification system and a classification method that classifies screening data through an ensemble natural language processing model that reflects both word and context features.

[0004] Registration No. 2625827 (2024.01.10), "System and method for predicting biological age by analyzing aging index through artificial intelligence"

[0005] Registration No. 2625827 discloses a system and method for calculating various biological ages of a person using multiple aging indices using artificial intelligence.

[0006] A method for calculating biological age based on physical indicators, performed by a server according to registration number 2625827, comprises: (a) receiving physical indicators from a user terminal; (b) calculating the biological ages of a plurality of bodies, each of which is pre-set, through the physical indicators and a pre-set machine learning model; and (c) generating a report based on the biological ages and providing it to the user terminal, wherein the machine learning model is composed of multimodals corresponding to a plurality of bodies, and is a model that utilizes a plurality of physical indicators and the biological ages as training data, and the machine learning model comprises: (b1) a step in a first layer, based on the physical indicators, inputting physical indicators corresponding to each modal into a first artificial intelligence model of each modal and extracting first feature values ​​for each modal; and (b2) a step in a second layer, inputting the plurality of first feature values ​​into a second artificial intelligence model and extracting second feature values ​​representing the first feature values. and (b3) in the third layer, the second feature values ​​are separated into tasks corresponding to each modal and input into the third artificial intelligence model for each task, thereby outputting multiple body ages for each modal.

[0007] However, these conventional technologies have the problem of failing to reflect the weighting of physical indicators that vary by age, failing to reflect the diverse and complex interactions of biological functions, and failing to provide interpretation of analysis results or solutions.

[0008] Therefore, there is an urgent need for an improved model that numerically displays the degree of aging in biological functions relative to the user's age, based on health data acquired from the user. The problem to be solved

[0009] The embodiment of the present invention aims to provide a method and system for quantifying bio-functions based on health data, which is applied to healthcare services to help users intuitively understand their current health status and degree of aging, and accordingly provide customized solutions needed by the user. means of solving the problem

[0010] A bio-function indexing method based on health data according to the present invention may include: a step of receiving a plurality of health data from a health information database; a step of calculating a bio-function index and a detailed function index for a specific user through a prediction model modeled by deep learning the plurality of health data; and a step of outputting the bio-function index and the detailed function index to the user.

[0011] In addition, the bio-function indexing system based on health data according to the present invention may be configured to include: an input unit that receives a plurality of health data from a health information database; a calculation unit that calculates a bio-function index and a detailed function index for a specific user through a prediction model modeled by deep learning the plurality of health data; and an output unit that outputs the bio-function index and the detailed function index to the user. Effects of the invention

[0012] According to the present invention, a method and system for quantifying bio-functions based on health data can be provided, which is applied to healthcare services to help users intuitively understand their current health status and degree of aging, and accordingly provide customized solutions needed by the user. Brief explanation of the drawing

[0013] FIG. 1 is a block diagram illustrating the internal configuration of a bio-function indexing system based on health data according to the present invention. FIG. 2 is a figure showing an example of the configuration of the bio-function indexing system of the present invention. Figure 3 is a diagram illustrating the calculation of the bio-functional index and the detailed functional index. Figure 4 is a figure illustrating an example of outlier removal using the IQR method. Figures 5 and 6 are figures to illustrate an example of mapping severity scores to abnormal findings. Figure 7 is a figure illustrating the calculation of a density distribution-based function index. Figure 8 is a diagram summarizing the detailed functional indices of 'carbohydrate metabolism' calculated in Figure 7. FIG. 9 is a flowchart illustrating the sequence of a bio-function indexing method based on health data according to the present invention. Specific details for implementing the invention

[0014] Hereinafter, embodiments are described in detail with reference to the attached drawings. However, various modifications may be made to the embodiments, and thus the scope of the patent application is not limited or restricted by these embodiments. It should be understood that all modifications, equivalents, and substitutions to the embodiments are included within the scope of the rights.

[0015] The terms used in the embodiments are for illustrative purposes only and should not be interpreted as intended to be limiting. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0016] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the embodiments pertain. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application.

[0017] In addition, when describing with reference to the attached drawings, identical components are assigned the same reference numeral regardless of drawing symbols, and redundant descriptions thereof are omitted. In describing the embodiments, if it is determined that a detailed description of related prior art could unnecessarily obscure the essence of the embodiments, such detailed description is omitted.

[0018] FIG. 1 is a block diagram illustrating the internal configuration of a bio-function indexing system based on health data according to the present invention.

[0019] Referring to FIG. 1, a bio-function indexing system based on health data according to the present invention (hereinafter abbreviated as ‘bio-function indexing system’, 100) may be configured to include an input unit (110), a calculation unit (120) including a detailed function calculation unit (122) and a bio-function calculation unit (124), and an output unit (130). Additionally, the bio-function indexing system (100) may be configured to optionally include a processing unit (140) according to an embodiment.

[0020] First, the input unit (110) receives multiple health data from a health information database. That is, the input unit (110) can play the role of collecting multiple health data for deep learning from a health information database maintained at a contracted hospital, clinic, examination center, etc.

[0021] In collecting health data, the input unit (110) can input health data after excluding key information (name, address, occupation, etc.) that can identify the examinee from the health data in order to prevent leakage of personal information.

[0022] The prediction model may receive multiple health data as input and be modeled by deep learning these health data. The prediction model in the present invention can be exemplified as a learning model that learns from multiple health data and outputs a learned value for a detailed function (e.g., a density estimate described below).

[0023] The calculation unit (120) calculates the bio-function index and detailed function index for a specific user through a prediction model that is modeled by deep learning the plurality of health data. That is, the calculation unit (120) can numerically calculate the degree of activation of the user's bio-function and detailed functions of the bio-function as the bio-function index / detailed function index.

[0024] Biological function refers to an important biological ability that reflects the deterioration of a user's health status due to aging, and in the present invention, it can be classified into eight functions: 'musculoskeletal function', 'endocrine function', 'detoxification function', 'cardiovascular function', 'immune function', 'digestive function', 'metabolic function', and 'excretion function'.

[0025] Detailed functions refer to detailed capabilities for each biological function, and in the present invention, they can be classified into 18 functions: 'bone function, muscle maintenance function, body fat control function' (detailed function of 'musculoskeletal function'), 'thyroid function, blood sugar control function' (detailed function of 'endocrine function'), 'liver metabolic function, bile excretion function' (detailed function of 'detoxification function'), 'heart function, blood vessel function' (detailed function of 'cardiovascular function'), 'immune cell generation function, antibody generation function' (detailed function of 'immune function'), 'pancreatic digestive function, gastrointestinal digestive function' (detailed function of 'digestive function'), 'carbohydrate metabolism, lipid metabolism, protein metabolism, thyroid function' (detailed function of 'metabolic function'), and 'waste excretion function' (detailed function of 'excretion function').

[0026] The calculation unit (120) may be configured to include a detailed function calculation unit (122) and a bio-function calculation unit (124) for calculating each of the detailed function index and the bio-function index.

[0027] The detailed function calculation unit (122) can identify the biological function defined by the prediction model and the utilization data for each detailed function of the biological function in the health data input by the user. That is, the detailed function calculation unit (122) can identify the utilization data of the detailed function by analyzing the health data collected from a user who wants to know their function index through the health data.

[0028] For example, the detailed function calculation unit (122) can check the detailed function of the biological function 'metabolic function', 'carbohydrate metabolism', in conjunction with the input of health data, and by referring to Table 7 described later, etc., it can check 'fasting blood glucose (FBS)', 'triglycerides (TG)', 'urinary glucose (Gluco)', and 'glycated hemoglobin (HbA1c)' as utilization data for the detailed function 'carbohydrate metabolism'.

[0029] Additionally, the detailed function calculation unit (122) can input the result value measured by the above utilization data into the prediction model. That is, the detailed function calculation unit (122) can find the result value of the utilization data identified in the user's health data and input it as an input into the prediction model.

[0030] In the example described above where data on the utilization of 'carbohydrate metabolism' is confirmed, the detailed function calculation unit (122) can find the result value of 'fasting blood glucose (FBS)' '105', the result value of 'triglycerides (TG)' '9.34', the result value of 'urinary glucose (Gluco)' '0.9', and the result value of 'glycated hemoglobin (HbA1c)' '5.80' from the health data and input them into the prediction model.

[0031] The detailed function calculation unit (122) can derive a density estimate of the detailed function from the prediction model learned from the result value. That is, the detailed function calculation unit (122) can derive a density estimate as an output from the prediction model learned from the result value.

[0032] The above prediction model can convert each of the above result values ​​into an estimate by applying density estimation according to the user's gender and age. That is, the prediction model can convert result values ​​obtained in different ranges into an estimate obtained within a normalized range between min 0 and max 200 by applying density estimation.

[0033] For example, regarding the result value '105' of 'Fasting Blood Sugar (FBS)' described above, the prediction model can apply density estimation to convert it into an estimate of '65.93' within a normalized range between min 0 and max 200. Similarly, the prediction model can convert the result value '9.34' of 'Triglycerides (TG)' into an estimate of '41.15', the result value '0.9' of 'Gluco' into an estimate of '58.03', and the result value '5.80' of 'Glycated Hemoglobin (HbA1c)' into an estimate of '67.01'.

[0034] In addition, the prediction model may provide the density estimate as a sum obtained by reflecting density-based weights according to the user's gender and age to the transformed estimate. That is, the prediction model can calculate the density estimate by considering the user's gender and age, reflecting density-based weights to each estimate, and summing all estimates after reflecting the density-based weights.

[0035] For example, regarding the above-mentioned estimates of ‘fasting blood glucose (FBS)’ (65.93), ‘triglycerides (TG)’ (41.15), ‘glucose in urine (Gluco)’ (58.03), and ‘glycated hemoglobin (HbA1c)’ (67.01), the prediction model can provide a value of ‘59.34’ as a density estimate, which is the sum of values ​​after reflecting density-based weights according to user gender ‘male’ and age ‘55 years’.

[0036] Afterward, the detailed function calculation unit (122) can calculate a detailed function index for the detailed function through the density estimate. That is, the detailed function calculation unit (122) can calculate a detailed function index by interpreting the density estimate derived from the prediction model as a value within a Gaussian distribution where the median is 100.

[0037] In the calculation of the detailed function index, the detailed function calculation unit (122) can create a Gaussian distribution such that the median of the distribution is 100, as a distribution regarding the total density estimate of the group of gender and age to which the user belongs. That is, the detailed function calculation unit (122) can create a Gaussian distribution for each group corresponding to a specific gender and age, and create a Gaussian distribution such that the median of the total density estimate is 100.

[0038] Additionally, the detailed function calculation unit (122) can calculate a detailed function index for the detailed function by substituting the density estimate derived for the user with a value in the created Gaussian distribution. That is, the detailed function calculation unit (122) can find a value corresponding to the density estimate derived from the prediction model in the Gaussian distribution created for the group to which the user belongs, and calculate the found value as the detailed function index.

[0039] For example, in the case where the density estimate '59.34' described above is derived, the detailed function calculation unit (122) can calculate the detailed function index for the detailed function 'carbohydrate metabolism' by replacing the density estimate '59.34' with a value '79' that makes the median value 100.

[0040] The bio-function calculation unit (124) can calculate the value obtained by assigning a selected weight to each of the detailed function indices and summing them as the bio-function index for the bio-function. That is, the bio-function calculation unit (124) can apply a weight to the detailed function indices and sum the detailed function indices after application to calculate the bio-function index of the bio-function to which these detailed functions belong.

[0041] At this time, the above weights may be selected for each detailed function index by considering the contribution of each detailed function to the biological function, and selected so that the total sum is 1.

[0042] For example, the weights can be set to 0.266 for the detailed function 'carbohydrate metabolism', 0.245 for the detailed function 'lipid metabolism', 0.276 for the detailed function 'protein metabolism', and 0.213 for the detailed function 'thyroid function'.

[0043] In addition, the weights can be selected such that the total sum of the selected weights is 1, i.e., '0.266 + 0.245 + 0.276 + 0.213 = 1'.

[0044] Accordingly, the biological function calculation unit (124) can calculate 95.627 as the biological function index of the biological function 'metabolic function', by adding 0.266×79 (detailed function index of carbohydrate metabolism) + 0.245×96 (detailed function index of lipid metabolism) + 0.276×101 (detailed function index of protein metabolism) + 0.213×109 (detailed function index of thyroid function).

[0045] The output unit (130) outputs the bio-function index and detailed function index to the user. That is, the output unit (130) can serve to visualize and display the calculated bio-function index and detailed function index.

[0046] In the output of the bio-function index and detailed function index, the output unit (130) can output the detailed function index calculated for each detailed function corresponding to the gender and age of the user, along with the difference from the median value of 100, to the user.

[0047] For example, the output unit (130) can output the detailed function index 79 to the user, along with the difference between the detailed function index 79 and the median value 100, '-21', corresponding to the user's gender 'male' and age '55 years old' for the detailed function index 79 of carbohydrate metabolism.

[0048] In addition, the output unit (130) can output the calculated bio-function index in parallel with the detailed function index.

[0049] For example, the output unit (130) can output a biofunctional index of 95.627 of the biofunctional 'metabolic function' in parallel with the detailed functional index of carbohydrate metabolism 79 (-21), etc.

[0050] According to an embodiment, the bio-function indexing system (100) can detect abnormal findings from health data and quantify the severity of the disease for the detected abnormal findings and show it to the user.

[0051] To this end, the computation unit (120) of the bio-function indexing system (100) may be configured to optionally include a processing unit (140).

[0052] The processing unit (140) can extract abnormal findings from the user's text-based health data, convert the extracted abnormal findings into standardized abnormal findings, and determine a severity score for the standardized abnormal findings.

[0053] Additionally, the processing unit (140) maps the predetermined severity score to the health data, and if multiple abnormal findings are extracted from the health data, it may map the highest severity score among the predetermined severity scores to the health data.

[0054] For example, in conjunction with the input of user health data, the processing unit (140) can analyze the input health data to extract abnormal findings and standardize them into three texts: 'calcified nodule', 'cystic nodule', and 'calcified lesion'.

[0055] Additionally, the processing unit (140) can assign a severity score to each of the three standardized abnormal findings, thereby determining a severity score of 6 for 'calcified nodules', a severity score of 5 for 'cystic nodules', and a severity score of 4 for 'calcified lesions'.

[0056] Additionally, the processing unit (140) can map the 6 points with the highest severity score to the health data.

[0057] The final severity score mapped through the above process can be used for detailed function calculations by undergoing the same process in the detailed function calculation unit (122).

[0058] According to the present invention, a method and system for quantifying bio-functions based on health data can be provided, which is applied to healthcare services to help users intuitively understand their current health status and degree of aging, and accordingly provide customized solutions needed by the user.

[0059] The bio-function indexing system (100) of the present invention designs a multidimensional structure to reflect the interaction of complex bio-functions, and thereby can provide an interpretation and a solution.

[0060] In addition, the bio-function indexing system (100) can reflect the importance of variables that appear differently by age by setting the average degree of aging of the same age and gender.

[0061] A bio-function indexing system (100) can be implemented by including a first step of receiving health data from a user, a second step of converting the received health data into necessary variables, a third step of generating a target index using a density estimation method based on the constituent variables of a detailed function using the received health data, a fourth step of creating a detailed function prediction model using the necessary variables and the target index, and a fifth step of calculating the user's final bio-function index through the detailed function prediction model.

[0062] FIG. 2 is a figure showing an example of the configuration of the bio-function indexing system of the present invention.

[0063] The bio-function indexing system (100) may be configured to include an input unit (210), a detailed function calculation unit (220), a bio-function calculation unit (230), and an output unit (240).

[0064] The input unit (210) can receive various health data from the user, such as genetic test results, skin type test results, and lifestyle data, based on the user's health checkup result data.

[0065] The detailed function calculation unit (220) can calculate a detailed function index for each detailed function based on the input health data. The detailed function calculation unit (220) can calculate a detailed function index by quantifying the defined detailed function for each biological function.

[0066] For example, with respect to the biological function 'musculoskeletal function', the detailed function calculation unit (220) can calculate a detailed function index for each of the detailed functions of the biological function 'musculoskeletal function', such as 'bone function', 'muscle maintenance function', and 'body fat control function', respectively, according to the 'utilization data'.

[0067] In calculating the detailed function index, the detailed function calculation unit (220) can implement the functions of generating a variable utilizing health data (221), inputting age and variable within the detailed function prediction model (222), and calculating the detailed function prediction index (223).

[0068] The detailed function operation unit (220) can convert health data into necessary variables by performing ratio calculation, loging, LLM text analysis, etc. on the health data in the generation of health data utilization variables (221).

[0069] The detailed function calculation unit (220) can input the converted required variable and the user's age into the defined detailed function prediction model in the variable input (222) within the detailed function prediction model.

[0070] The detailed function calculation unit (220) can output the output value from the detailed function prediction model as the detailed function index in the detailed function prediction index calculation (223).

[0071] The bio-function calculation unit (230) can derive a final bio-function index through calculations between detailed functions. The bio-function calculation unit (230) can perform calculation methods such as simple sum, weighted sum, machine learning, etc., and can represent the score, age, index, etc. as numerical values ​​as calculation results.

[0072] The output unit (240) can provide analysis results to the user based on the calculation results. As analysis results, the output unit (240) can provide all information that can provide value to the user, such as detailed explanations of biological functions, lifestyle recommendations, dietary recommendations, exercise therapy recommendations, and nutritional supplement recommendations.

[0073] According to the present invention, difficult-to-understand medical health conditions can be understood in an easy and intuitive manner, and through functional improvement of the body, it can contribute to extending a healthy life rather than merely extending lifespan.

[0074] Tables 1 to 8 list the detailed functions for each higher-level function (biological function) specified in the present invention.

[0075]

[0076] Table 1 lists the utilization data and variable names assigned to each of the three detailed functions, 'bone function, muscle maintenance function, and body fat control function,' for the higher-level function 'musculoskeletal function.'

[0077] The bio-function indexing system (100) can extract the result value of the utilization data 'ALP' belonging to the upper function 'musculoskeletal function' - detailed function 'bone function' from the health data received from the user, and convert the extracted result value of 'ALP' into a variable name 'ALP' based on Table 1.

[0078]

[0079] Table 2 lists the utilization data and variable names assigned to each of the two detailed functions, 'thyroid function and blood sugar control function,' for the upper function 'endocrine function.'

[0080] The bio-function indexing system (100) can extract result values ​​of utilization data 'fasting blood glucose (FBS)' and 'triglycerides (TB)' belonging to the upper function 'endocrine function' - detailed function 'blood glucose control function' from health data received from a user, and apply log analysis to the extracted result values ​​to convert them into variable names 'TyG index' based on Table 2.

[0081]

[0082] Table 3 lists the utilization data and variable names assigned to each of the two detailed functions, 'liver metabolic function' and 'bile excretion function,' for the higher-level function 'detoxification function.'

[0083] The bio-function indexing system (100) can extract the result value of the utilization data 'AST(SGOT)' belonging to the upper function 'detoxification function' - detailed function 'liver function' from the health data received from the user, and convert the extracted result value of 'AST(SGOT)' into the variable name 'AST(SGOT)' based on Table 3.

[0084]

[0085] Table 4 lists the utilization data and variable names assigned to each of the two detailed functions, 'cardiac function' and 'vascular function', for the higher-level function 'cardiovascular function'.

[0086] The bio-function indexing system (100) can extract result data of utilization data 'electrocardiogram' belonging to the upper function 'cardiovascular function' - detailed function 'heart function' from health data received from a user, and extract severity according to the processing method of the processing unit (140) that selectively includes the extracted text-form result data, and convert it into a variable name 'severity' based on Table 4.

[0087]

[0088] Table 5 lists the utilization data and variable names assigned to each of the two detailed functions, 'immune cell generation function' and 'antibody generation function', for the higher-level function 'immune function'.

[0089] The bio-function indexing system (100) can extract the result value of the utilization data 'white blood cell (WBC)' belonging to the upper function 'immune function' - detailed function 'immune cell generation function' from the health data received from the user, and convert the extracted result value of 'white blood cell (WBC)' into the variable name 'white blood cell (WBC)' based on Table 5.

[0090]

[0091] Table 6 lists the utilization data and variable names assigned to each of the two detailed functions, 'pancreatic digestive function' and 'gastrointestinal digestive function', for the higher-level function 'digestive function'.

[0092] The bio-function indexing system (100) can extract the result value of the utilization data 'Amylase' belonging to the upper function 'Digestive function' - detailed function 'Pancreatic digestive function' from the health data received from the user, and convert the extracted result value of 'Amylase' into the variable name 'Amylase' based on Table 6.

[0093]

[0094] Table 7 lists the utilization data and variable names assigned to each of the four detailed functions, 'carbohydrate metabolism, lipid metabolism, protein metabolism, and thyroid function,' for the higher-level function 'metabolic function.'

[0095] The bio-function indexing system (100) can extract the result value of the utilization data 'fasting blood glucose (FBS)' belonging to the upper function 'metabolic function' - detailed function 'carbohydrate metabolism' from the health data received from the user, and convert the extracted result value of 'fasting blood glucose (FBS)' into the variable name 'fasting blood glucose (FBS)' based on Table 7.

[0096]

[0097] Table 8 lists the utilization data and variable names assigned to one detailed function, 'waste excretion function', for the higher-level function 'excretion function'.

[0098] The bio-function indexing system (100) can extract the result value of the utilization data 'glomerular filtration rate (GFR)' belonging to the upper function 'excretion function' - detailed function 'waste excretion' from the health data received from the user, and convert the extracted result value of the 'glomerular filtration rate (GFR)' into the variable name 'glomerular filtration rate (GFR)' based on Table 8.

[0099] The bio-function indexing system (100) can classify the upper functions (bio-functions) into a total of 8 and the detailed functions into a total of 18, as shown in Tables 1 to 8.

[0100] Figure 3 is a diagram illustrating the calculation of the bio-functional index and the detailed functional index.

[0101] Figure 3 illustrates the calculation of the bio-functional index and detailed functional index according to the present invention by receiving health data generated according to the 2024 health checkup of a user who is a 55-year-old male.

[0102] The bio-function indexing system (100) can receive 2024 health data from the user.

[0103] The bio-function indexing system (100) can extract the result values ​​of the utilization data designated for each of the four sub-functions (detailed functions) 'carbohydrate metabolism', 'lipid metabolism', 'protein metabolism', and 'thyroid function' for the upper function (biological function) 'metabolic function' from the input 2024 health data based on the preceding Table 7, and convert the extracted result values ​​into each necessary variable.

[0104] In addition, the bio-function indexing system (100) inputs the converted required variables and the user's gender (male) / age (55 years old) into a detailed function prediction model, and can quantify the function index of the upper function 'metabolic function' output from the detailed function prediction model as '93(-7)' and display it on the user terminal.

[0105] Here, '(-7)' can represent the size of the value (93) calculated from the median value of the overall function index of the upper function 'metabolic function' of 100 in the group to which the user belongs (male 55 years old).

[0106] In addition, the bio-function indexing system (100) can quantify the function indices for each of the sub-functions 'carbohydrate metabolism', 'lipid metabolism', 'protein metabolism', and 'thyroid function' output from the detailed function prediction model into numerical values ​​such as '87(-13)', '102(+2)', '90(-10)', and '93(-7)' and display them on a user terminal.

[0107] The bio-function indexing system (100) can inform the user of the overall bio-function index as a higher-level function and optionally provide a detailed sub-function index as a lower-level function.

[0108] The bio-function indexing system (100) can perform preprocessing on variables based on numerical data.

[0109] As a preprocessing step, the bio-function indexing system (100) can perform outlier removal and processing according to the directionality of each variable.

[0110] In removing outliers, the bio-function indexing system (100) can remove outliers from the input health data as unnecessary factors that degrade the performance of the model.

[0111] The bio-function indexing system (100) can remove outliers from health data using the IQR (Inter Quantile Range) method.

[0112] Figure 4 is a figure illustrating an example of outlier removal using the IQR method.

[0113] As shown in FIG. 4, the bio-function indexing system (100) can remove outliers by using an IQR method, which divides the data interval into four quartiles and uses the difference between the first quartile and the third quartile.

[0114] At this time, IQR is 'Q3 - Q1', and the bio-function indexing system (100) can determine the IQR outlier range using Q1 and Q3 in the data range by multiplying IQR by 1.5.

[0115] In processing according to the directionality of each variable, the bio-function indexing system (100) can distinguish between examination items such as AST (SGOT) and ALT (SGPT) whose values ​​increase with aging, and body muscle mass and albumin whose values ​​decrease with aging, and set the directionality so that a state of less aging can be indexed with a higher score.

[0116] For screening items such as systolic and diastolic blood pressure, which are dangerous whether lower or higher than the normal range, the absolute difference from the average value for the corresponding age and gender can be calculated and standardized so that a smaller absolute value is indexed as a higher score.

[0117] The bio-function indexing system (100) can reflect the user's actual health condition by indexing both the high blood pressure range of 140 and the low blood pressure range of 80 as low scores when, for example, the systolic blood pressure deviates from the appropriate range of 90 to 120.

[0118] The bio-function indexing system (100) can extract abnormal findings by preprocessing text data-based variables.

[0119] Preprocessing of text data-based variables can be a process of efficiently identifying and analyzing abnormal findings based on health data by selecting the optimal model for powerful natural language processing (NLP) of a large language model (LLM) and training it based on real data using fine-tuning of the model.

[0120] Fine-tuning is essentially the process of further training an LLM model to suit a specific domain (medicine and health checkups). Through fine-tuning, the model can understand medical terminology, interpretation of test results, and medical document formats, and learn the complex linguistic contexts and patterns used by medical professionals.

[0121] Data-driven model training methods can perform training data preparation, training data labeling, and fine-tuning processes.

[0122] In preparing training data, health checkup results and abnormal finding case data can be collected.

[0123] In training data labeling, a training dataset can be generated by labeling abnormal text findings for each health checkup item.

[0124] In the fine-tuning process, the model can be trained to understand the structure and medical significance of health data by additionally training it with labeled data based on the fundamentally trained LLM model.

[0125] The bio-function indexing system (100) can extract abnormal findings using LLM.

[0126] The bio-function indexing system (100) can precisely analyze input text data using a fine-tuned model and effectively identify and extract abnormal findings.

[0127] The bio-function indexing system (100) can standardize abnormal findings.

[0128] The bio-function indexing system (100) can replace all abnormal findings with standardized terms so that data can be managed consistently and systematically by establishing a data standardization dictionary, as the methods of expressing abnormal findings used by each examination center differ.

[0129] The bio-function indexing system (100) can compare and analyze terms of abnormal findings provided by various screening centers and integrate expressions with similar meanings to establish a single standardized terminology system.

[0130] The bio-function indexing system (100) minimizes confusion that may occur during the analysis and management of abnormal findings by ensuring consistency between data, and can increase the reliability and efficiency of data analysis.

[0131] For example, the bio-function indexing system (100) can substitute terms such as ‘calcification within the prostate’, ‘local calcification within the prostate’, etc., with ‘prostate calcification’.

[0132] The bio-function indexing system (100) can map the severity of abnormal findings.

[0133] The bio-function indexing system (100) can map the extracted abnormal findings into a severity score through consultation with an MD (Medical Doctor).

[0134] Here, the mapping score can be set between 0 and 10 points, with normal being 0 and the score increasing according to severity.

[0135] Figures 5 and 6 are figures to illustrate an example of mapping severity scores to abnormal findings.

[0136] The bio-function indexing system (100) can analyze health data in text form, such as FIG. 5, and extract abnormal findings.

[0137] In FIG. 5, the bio-function indexing system (100) can extract abnormal findings including ‘calcified nodules, cystic nodules, calcified lesions’.

[0138] The bio-function indexing system (100) can perform mapping of severity scores from extracted abnormal findings.

[0139] The bio-function indexing system (100) can set the severity score to a value between 0 and 10 points.

[0140] The bio-function indexing system (100) can replace the severity score with a default value set in advance through consultation with an MD (Medical Doctor) for each test item in the case of a customer without data.

[0141] When abnormal findings such as ‘calcified nodules’, ‘cystic crystals’, and ‘calcified lesions’ are extracted in Fig. 5, the bio-function indexing system (100) can determine a severity score for each abnormal finding.

[0142] In FIG. 6, the bio-function indexing system (100) can determine a severity score of '6' for 'calcified nodules', a severity score of '5' for 'cystic crystals', and a severity score of '4' for 'calcified lesions'.

[0143] Afterwards, the bio-function indexing system (100) can reflect only the ‘calcified nodules’ with the highest severity score.

[0144] For example, the bio-function indexing system (100) can map an abnormal finding 'calcified nodule' - severity score '6' to the health data of FIG. 5, as shown in FIG. 6.

[0145] The bio-function indexing system (100) can perform data density estimation.

[0146] Densitization can be the estimation of the probability of all possible values ​​a variable can take from the distribution of data.

[0147] As an example of density measurement, Gaussian Density Estimation assumes that the data follows a single Gaussian distribution and uses a covariance matrix, which has the advantage of being insensitive to the range of data variables.

[0148] The bio-function indexing system (100) can perform Gaussian density estimation specialized in calculating the result value of the examination item.

[0149] The bio-function indexing system (100) can perform Gaussian density estimation specialized for the examination item to precisely analyze the distribution characteristics of the health data based on the examination item result value.

[0150] The bio-function indexing system (100) smoothly models the statistical distribution of continuous numerical data observed in each examination item through Gaussian density estimation, and enables efficient identification of the central tendency and variability of the examination item data.

[0151] In addition, the bio-function indexing system (100) allows for the unique identification of distributions by dividing them into gender and age.

[0152] Through this, the bio-function indexing system (100) can play an important role in supporting the interpretation of examination results and medical decision-making by providing more detailed and accurate data compared to existing simple average or range-based analysis.

[0153] Figure 7 is a figure illustrating the calculation of a density distribution-based function index.

[0154] Figure 7 illustrates the calculation of a function index based on gender / age density distribution for 'carbohydrate metabolism', one of the four detailed functions of the upper function (biological function) 'metabolic function' of a 55-year-old male user.

[0155] In Figure 7, the detailed function 'carbohydrate metabolism' includes the variable names of the health data items used as 'fasting blood glucose', 'TyG index', 'urinary glucose', and 'glycated hemoglobin'.

[0156] As shown in FIG. 7, the bio-function indexing system (100) can confirm from health data the result value of ‘fasting blood glucose’ ‘105’, the result value of ‘TyG index’ ‘9.34’, the result value of ‘urinary glucose’ ‘0.9’, and the result value of ‘glycated hemoglobin’ ‘5.80’.

[0157] Additionally, the bio-function indexing system (100) can convert the values ​​into an estimated value of ‘fasting blood glucose_DE’ of ‘65.93’, an estimated value of ‘TyG index_DE’ of ‘41.15’, an estimated value of ‘urinary glucose_DE’ of ‘58.03’, and an estimated value of ‘glycated hemoglobin_DE’ of ‘67.01’ by applying a density estimation specialized for each result value.

[0158] Afterwards, the bio-function indexing system (100) can calculate a density-based weight of '59.34' by applying density estimation to each converted estimate and then summing them up.

[0159] In addition, the bio-function indexing system (100) can calculate the final function index (detailed function index) of ‘carbohydrate metabolism’ as ‘79’ by converting the density-based weighting ‘59.34’ so that the median value becomes 100.

[0160] The bio-function indexing system (100) can generate a machine learning-based prediction model using the calculated function index as a label.

[0161] The bio-function indexing system (100) can generate a machine learning-based detailed function index prediction model.

[0162] Figure 8 is a diagram summarizing the detailed functional indices of 'carbohydrate metabolism' calculated in Figure 7.

[0163] As shown in FIG. 8, the bio-function indexing system (100) can clearly organize and output the result value of ‘fasting blood glucose’, the result value of ‘TyG index’, the result value of ‘urinary glucose’, the result value of ‘5.80’, and the final function index (detailed function index) ‘79’ for the health data item ‘fasting blood glucose’ utilizing the detailed function ‘carbohydrate metabolism’.

[0164] The bio-function indexing system (100) can generate a machine learning prediction model using the calculated detailed function index of the detailed function 'carbohydrate metabolism'.

[0165] For example, the bio-function indexing system (100) can learn machine learning with about 1 million or more examination data by setting the detailed function index of ‘carbohydrate metabolism’ as the dependent variable and age, fasting blood glucose, TyG index, etc. as independent variables.

[0166] The bio-function indexing system (100) can generate two machine learning prediction models for each detailed function, for both men and women.

[0167] The bio-function indexing system (100) can perform prediction model training using known artificial intelligence algorithms such as Xgboost and Deep Neural Network.

[0168] The bio-function indexing system (100) can proceed with model training by selecting Xgboost through comparison of evaluation indicators between artificial intelligence algorithms.

[0169] The bio-function indexing system (100) can perform hyperparameter tuning.

[0170] The bio-function indexing system (100) can utilize a RandomSearch algorithm to find optimized hyperparameters (variables set in the model to implement the optimal training model) of a machine learning model.

[0171] The bio-function indexing system (100) can perform Xgboost model training by selecting an optimal combination of hyperparameters.

[0172] The bio-function indexing system (100) can utilize MAE (Mean Absolute Error) and R² (R-squared, coefficient of determination) as model evaluation indicators (performance indicators).

[0173] MAE is a regression evaluation metric that calculates the absolute difference between the model's predicted value and the actual value, and then finds the average. A smaller MAE value indicates that the prediction is closer to the actual value.

[0174] R² is a regression evaluation metric that indicates how well a model explains the data. R² represents the ratio of how well predicted values ​​explain actual values, and a value closer to 1 indicates better performance.

[0175] Table 9 lists the evaluation indicators.

[0176]

[0177] The bio-function indexing system (100) can calculate the final bio-function (higher-level function) index based on the function index of the auxiliary function.

[0178] Table 10 lists examples of upper and lower functions.

[0179]

[0180] The bio-function indexing system (100) can determine, for the upper function (biological function) 'metabolic function', 'carbohydrate metabolism', 'lipid metabolism', 'protein metabolism', and 'thyroid function' as lower functions (detailed functions).

[0181] The bio-function indexing system (100) can calculate the bio-function index for the higher-level function 'metabolic function' according to Formula 1.

[0182] [Formula 1]

[0183]

[0184] For example, the bio-function indexing system (100) can select a weight for each detailed function index by considering the importance of the detailed function under conditions where the detailed function index of carbohydrate metabolism 79, the detailed function index of lipid metabolism 96, the detailed function index of protein metabolism 101, and the detailed function index of thyroid function 109 are calculated.

[0185] The bio-function indexing system (100) can select 0.266 for the detailed function 'carbohydrate metabolism', 0.245 for the detailed function 'lipid metabolism', 0.276 for the detailed function 'protein metabolism', and 0.213 for the detailed function 'thyroid function'.

[0186] At this time, the bio-function indexing system (100) can adjust the size of the weights so that the total sum of the selected weights is 1 (= 0.266 + 0.245 + 0.276 + 0.213).

[0187] Subsequently, the bio-function indexing system (100) can calculate 95.627 ( = 0.266×79 + 0.245×96 + 0.276×101 + 0.213×109), which is the bio-function index for the upper function 'metabolic function', by applying the selected weight and detailed function index.

[0188] FIG. 9 is a flowchart illustrating the sequence of a bio-function indexing method based on health data according to the present invention.

[0189] The bio-function indexing method based on health data according to the present embodiment can be performed by the bio-function indexing system (100) described above.

[0190] First, the bio-function indexing system (100) receives multiple health data from a health information database (910). Step (910) may be a process of collecting multiple health data for deep learning from a health information database maintained at a contracted hospital, clinic, examination center, etc.

[0191] In collecting health data, the bio-function indexing system (100) can input health data after excluding key information (name, address, occupation, etc.) that can identify the examinee from the health data in order to prevent leakage of personal information.

[0192] The prediction model may receive multiple health data as input and be modeled by deep learning these health data. The prediction model in the present invention can be exemplified as a learning model that learns from multiple health data and outputs learned values ​​(e.g., density estimates) for detailed functions.

[0193] Additionally, the bio-function indexing system (100) calculates a bio-function index and a detailed function index for a specific user through a prediction model that is modeled by deep learning the plurality of health data (920). Step (920) may be a process of numerically calculating the bio-function index / detailed function index of the user's bio-function and the degree of activation of the detailed functions of the bio-function.

[0194] Biological function refers to an important biological ability that reflects the deterioration of a user's health status due to aging, and in the present invention, it can be classified into eight functions: 'musculoskeletal function', 'endocrine function', 'detoxification function', 'cardiovascular function', 'immune function', 'digestive function', 'metabolic function', and 'excretion function'.

[0195] Detailed functions refer to detailed capabilities for each biological function, and in the present invention, they can be classified into 18 functions: 'bone function, muscle maintenance function, body fat control function' (detailed function of 'musculoskeletal function'), 'thyroid function, blood sugar control function' (detailed function of 'endocrine function'), 'liver metabolic function, bile excretion function' (detailed function of 'detoxification function'), 'heart function, blood vessel function' (detailed function of 'cardiovascular function'), 'immune cell generation function, antibody generation function' (detailed function of 'immune function'), 'pancreatic digestive function, gastrointestinal digestive function' (detailed function of 'digestive function'), 'carbohydrate metabolism, lipid metabolism, protein metabolism, thyroid function' (detailed function of 'metabolic function'), and 'waste excretion function' (detailed function of 'excretion function').

[0196] The bio-function indexing system (100) can identify the bio-functions defined by the prediction model and the utilization data for each detailed function of the bio-functions in the health data input by the user. That is, the bio-function indexing system (100) can identify the utilization data of the detailed functions by analyzing the health data collected from a user who wants to know their function index through the health data.

[0197] For example, the bio-function indexing system (100) can identify ‘carbohydrate metabolism,’ which is a detailed function of the bio-function ‘metabolic function,’ in conjunction with the input of health data, and identify ‘fasting blood glucose (FBS),’ ‘triglycerides (TG),’ ‘urinary glucose (Gluco),’ and ‘glycated hemoglobin (HbA1c)’ as utilization data for the detailed function ‘carbohydrate metabolism’ by referring to Table 7, etc.

[0198] Additionally, the bio-function indexing system (100) can input the result value measured by the above-mentioned utilization data into the prediction model. That is, the bio-function indexing system (100) can find the result value of the utilization data identified in the user's health data and input it as an input into the prediction model.

[0199] In the example described above where data on the utilization of 'carbohydrate metabolism' is confirmed, the bio-function indexing system (100) can find the result value of 'fasting blood glucose (FBS)' '105', the result value of 'triglycerides (TG)' '9.34', the result value of 'urinary glucose (Gluco)' '0.9', and the result value of 'glycated hemoglobin (HbA1c)' '5.80' from the health data and input them into a prediction model.

[0200] The bio-function indexing system (100) can derive a density estimate of the detailed function from the prediction model learned from the result value. That is, the bio-function indexing system (100) can derive a density estimate as an output from the prediction model learned from the result value.

[0201] The above prediction model can convert each of the above result values ​​into an estimate by applying density estimation according to the user's gender and age. That is, the prediction model can convert result values ​​obtained in different ranges into an estimate obtained within a normalized range between min 0 and max 200 by applying density estimation.

[0202] For example, regarding the result value '105' of 'Fasting Blood Sugar (FBS)' described above, the prediction model can apply density estimation to convert it into an estimate of '65.93' within a normalized range between min 0 and max 200. Similarly, the prediction model can convert the result value '9.34' of 'Triglycerides (TG)' into an estimate of '41.15', the result value '0.9' of 'Gluco' into an estimate of '58.03', and the result value '5.80' of 'Glycated Hemoglobin (HbA1c)' into an estimate of '67.01'.

[0203] In addition, the prediction model may provide the density estimate as a sum obtained by reflecting density-based weights according to the user's gender and age to the transformed estimate. That is, the prediction model can calculate the density estimate by considering the user's gender and age, reflecting density-based weights to each estimate, and summing all estimates after reflecting the density-based weights.

[0204] For example, regarding the above-mentioned estimates of ‘fasting blood glucose (FBS)’ (65.93), ‘triglycerides (TG)’ (41.15), ‘glucose in urine (Gluco)’ (58.03), and ‘glycated hemoglobin (HbA1c)’ (67.01), the prediction model can provide a value of ‘59.34’ as a density estimate, which is the sum of values ​​after reflecting density-based weights according to user gender ‘male’ and age ‘55 years’.

[0205] Subsequently, the bio-function indexing system (100) can calculate a detailed function index for the detailed function through the density estimate. That is, the bio-function indexing system (100) can calculate a detailed function index by interpreting the density estimate derived from the prediction model as a value within a Gaussian distribution where the median is 100.

[0206] In calculating the detailed function index, the bio-function indexing system (100) can create a Gaussian distribution such that the median of the distribution is 100, as a distribution regarding the total density estimate of the group of gender and age to which the user belongs. That is, the bio-function indexing system (100) can create a Gaussian distribution for each group corresponding to a specific gender and age, and create a Gaussian distribution with the median of the total density estimate being 100.

[0207] Additionally, the bio-function indexing system (100) can calculate a detailed function index for the detailed function by substituting the density estimate derived for the user with a value in the created Gaussian distribution. That is, the bio-function indexing system (100) can find a value corresponding to the density estimate derived from the prediction model in the Gaussian distribution created for the group to which the user belongs, and calculate the found value as the detailed function index.

[0208] For example, in the case where the density estimate '59.34' described above is derived, the bio-function indexing system (100) can calculate a detailed function index for the detailed function 'carbohydrate metabolism' by replacing the density estimate '59.34' with a value '79' that makes the median value 100.

[0209] The bio-function indexing system (100) can calculate the value obtained by assigning a selected weight to each of the detailed function indices and summing them as the bio-function index for the bio-function. That is, the bio-function indexing system (100) can apply a weight to the detailed function indices and sum the detailed function indices after application to calculate the bio-function index of the bio-function to which these detailed functions belong.

[0210] At this time, the above weights may be selected for each detailed function index by considering the contribution of each detailed function to the biological function, and selected so that the total sum is 1.

[0211] For example, the weights can be set to 0.266 for the detailed function 'carbohydrate metabolism', 0.245 for the detailed function 'lipid metabolism', 0.276 for the detailed function 'protein metabolism', and 0.213 for the detailed function 'thyroid function'.

[0212] In addition, the weights can be selected such that the total sum of the selected weights is 1, i.e., '0.266 + 0.245 + 0.276 + 0.213 = 1'.

[0213] Accordingly, the biofunction indexing system (100) can calculate 95.627 as the biofunction index of the biofunction 'metabolic function', by adding 0.266×79 (detailed function index of carbohydrate metabolism) + 0.245×96 (detailed function index of lipid metabolism) + 0.276×101 (detailed function index of protein metabolism) + 0.213×109 (detailed function index of thyroid function).

[0214] Continuing, the bio-function indexing system (100) outputs the bio-function index and detailed function index to the user (930). Step (930) may be a process of visualizing and displaying the calculated bio-function index and detailed function index.

[0215] In outputting the bio-function index and detailed function index, the bio-function indexing system (100) can output the detailed function index calculated for each detailed function corresponding to the gender and age of the user, along with the difference from the median value of 100, to the user.

[0216] For example, the bio-function indexing system (100) can output the detailed function index 79 to the user, along with the difference between the detailed function index 79 and the median value 100, '-21', corresponding to the user's gender 'male' and age '55 years old' for the detailed function index 79 of carbohydrate metabolism.

[0217] In addition, the bio-function indexing system (100) can output the calculated bio-function index in parallel with the detailed function index.

[0218] For example, the biofunction indexing system (100) can output a biofunction index of 95.627 for the biofunction 'metabolic function' in parallel with the detailed function index of carbohydrate metabolism 79 (-21), etc.

[0219] According to an embodiment, the bio-function indexing system (100) can detect abnormal findings from health data and quantify the severity of the disease for the detected abnormal findings and show it to the user.

[0220] To this end, the bio-function indexing system (100) can extract abnormal findings from the user's text-based health data, convert the extracted abnormal findings into standardized abnormal findings, and determine a severity score for the standardized abnormal findings.

[0221] Additionally, the bio-function indexing system (100) maps the above-determined severity score to the health data, and when multiple abnormal findings are extracted from the health data, it can map the highest severity score among the above-determined severity scores to the health data.

[0222] For example, in conjunction with the input of user health data, the bio-function indexing system (100) can analyze the input health data to extract abnormal findings and standardize them into three texts: 'calcified nodule', 'cystic nodule', and 'calcified lesion'.

[0223] In addition, the bio-function indexing system (100) can assign a severity score to each of the three standardized abnormal findings, thereby determining a severity score of 6 for 'calcified nodules', a severity score of 5 for 'cystic nodules', and a severity score of 4 for 'calcified lesions'.

[0224] Additionally, the bio-function indexing system (100) can map the highest severity score of 6 points to the health data.

[0225] The final severity score mapped through the above process can be used for detailed function calculations by undergoing the same process in the bio-function indexing system (100).

[0226] According to the present invention, a method and system for quantifying bio-functions based on health data can be provided, which is applied to healthcare services to help users intuitively understand their current health status and degree of aging, and accordingly provide customized solutions needed by the user.

[0227] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either alone or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the embodiment, or they may be those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware systems specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware system described above may be configured to operate as one or more software modules to perform the operation of the embodiment, and vice versa.

[0228] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing system to operate as desired or command the processing system independently or collectively. Software and / or data may be stored on any type of machine, component, physical system, virtual equipment, computer storage medium, or system so as to be interpreted by the processing system or to provide instructions or data to the processing system. Software may be distributed across networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

[0229] Although the embodiments have been described above with reference to the limited drawings, those skilled in the art can apply various technical modifications and variations based on the above. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or the components of the described system, structure, system, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.

[0230] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims set forth below. Explanation of the symbols

[0231] 100: Biofunctional Index System 110 : Input section 120 : Operation section 122: Detailed function calculation unit 124: Biological function calculation unit 130 : Output section 140 : Processing section

Claims

Claim 1 A method for quantifying biofunctionality based on health data, comprising: a step of receiving a plurality of health data from a health information database; a step of calculating a biofunctionality index and a detailed functionality index for a specific user through a prediction model modeled by deep learning the plurality of health data; and a step of outputting the biofunctionality index and the detailed functionality index to the user; further comprising: a step of extracting abnormal findings from the user's text-based health data; a step of converting the extracted abnormal findings into standardized abnormal findings; a step of determining a severity score for the standardized abnormal findings; and a step of mapping the severity score to the health data, wherein the user's text-based health data includes health checkup result data. Claim 2 A method for quantifying bio-functions based on health data, wherein the step of calculating comprises: a step of confirming, in health data input by the user, bio-functions defined by the prediction model and utilization data for each detailed function of the bio-functions; a step of inputting a result value measured from the utilization data into the prediction model; a step of deriving a density estimate of the detailed function from the prediction model learned from the result value; and a step of calculating a detailed function index for the detailed function through the density estimate. Claim 3 A method for quantifying biofunctionality based on health data, wherein, in paragraph 2, the prediction model converts each of the result values ​​into an estimate by applying density estimation according to the user's gender and age, and provides a value obtained by summing the converted estimates with density-based weights according to the user's gender and age as the density estimate. Claim 4 A method for quantifying biofunctionality based on health data, wherein the step of calculating the detailed functional index comprises: a step of creating a Gaussian distribution such that the median of the distribution is 100, which is a distribution regarding the total density estimate of the gender and age group to which the user belongs; and a step of calculating a detailed functional index for the detailed function by substituting the density estimate derived for the user with the value in the created Gaussian distribution. Claim 5 A method for quantifying biofunctionality based on health data, wherein the calculation step further includes the step of calculating the sum of the values ​​obtained by assigning selected weights to each of the detailed function indices as a biofunctionality index for the biofunctionality, and wherein the weights are selected for each of the detailed function indices considering the contribution of each detailed function to the biofunctionality, such that the total sum is 1. Claim 6 A method for quantifying biofunctionality based on health data, wherein the step of outputting to the user further comprises: the step of outputting to the user, with the difference from the median value of 100, the detailed function index calculated for each detailed function corresponding to the user's gender and age; and the step of outputting the biofunctionality index calculated in parallel with the detailed function index. Claim 7 A method for quantifying bio-function based on health data, wherein, in claim 1, the step of mapping the severity score to the health data further comprises the step of mapping the highest severity score among the determined severity scores to the health data when a plurality of abnormal findings are extracted from the health data. Claim 8 A biofunctional indexing system based on health data, comprising: an input unit that receives a plurality of health data from a health information database; a calculation unit that calculates a biofunctional index and a detailed functional index for a specific user through a prediction model modeled by deep learning the plurality of health data; and an output unit that outputs the biofunctional index and the detailed functional index to the user; further comprising a processing unit that extracts abnormal findings from the user's text-based health data, converts the extracted abnormal findings into standardized abnormal findings, determines a severity score for the standardized abnormal findings, and maps the severity score to the health data, wherein the user's text-based health data includes health checkup result data. Claim 9 In claim 8, the above-mentioned computational unit identifies the bio-function and utilization data for each detailed function of the bio-function specified in the prediction model from the health data input by the user, inputs the result value measured from the utilization data into the prediction model, derives the density estimate of the detailed function from the prediction model learned from the result value, and calculates the detailed function index for the detailed function through the density estimate, thereby comprising a detailed function computational unit, a bio-function indexing system based on health data. Claim 10 A bio-functional indexing system based on health data, wherein the prediction model converts each of the result values ​​into an estimate by applying density estimation according to the user's gender and age, and provides a value obtained by summing the converted estimates with density-based weights according to the user's gender and age as the density estimate. Claim 11 In claim 9, the detailed function calculation unit creates a Gaussian distribution such that the median of the distribution is 100, as a distribution regarding the total density estimate of the gender and age group to which the user belongs, and calculates a detailed function index for the detailed function by substituting the density estimate derived for the user with the value in the created Gaussian distribution. This is a bio-function indexing system based on health data. Claim 12 In claim 8, the above calculation unit further includes a biofunction calculation unit that calculates the sum of the values ​​obtained by assigning selected weights to each of the above detailed function indices as a biofunction index for the above biofunction, wherein the weights are selected for each of the above detailed function indices considering the contribution of each detailed function to the biofunction, and are selected such that the total sum is 1. A biofunction indexing system based on health data. Claim 13 In claim 12, the output unit outputs the detailed function index calculated for each detailed function, corresponding to the user's gender and age, along with the difference from the median value of 100, to the user, and outputs the bio-function index calculated in parallel with the detailed function index, a bio-function index system based on health data. Claim 14 In claim 8, the processing unit maps the highest severity score among the determined severity scores to the health data when a plurality of abnormal findings are extracted from the health data, a bio-function indexing system based on health data. Claim 15 A computer-readable recording medium having a program for executing any one of the methods of paragraphs 1 through 7.

Images