KR102450417B1 - Method, Computing Device and Computer-readable Medium for Predicting the Effect of Exercise Prescription for High Blood Pressure and Diabetes Patients based on Artificial Intelligence - Google Patents

Abstract

The present invention relates to a method, a computing device, and a computer-readable medium for predicting an effect of exercise prescription for patients with hypertension and diabetes, based on artificial intelligence (AI) and, more specifically, to a method, a computing device, and a computer-readable medium for predicting an effect of exercise prescription for patients with hypertension and diabetes, based on AI, which comprises the steps of: collecting patient information on a plurality of patients with hypertension and patients with diabetes to build a database; sorting only patients who can do exercise among the patient information included in the built database to generate a data set for learning and verifying a machine learning model; and inputting physical information on a user into the machine learning model learned through the data set to output information on prediction of exercise effects of a user by one week.

Description

{Method, Computing Device and Computer-readable Medium for Predicting the Effect of Exercise Prescription for High Blood Pressure and Diabetes Patients based on Artificial Intelligence}

The present invention relates to an artificial intelligence-based method for predicting the effect of exercise prescription for hypertension and diabetes patients, a computing device, and a computer-readable medium, and more particularly, to a database by collecting patient information for a plurality of hypertension and diabetes patients. Among the patient information included in the established database, only patients who can exercise are selected to create a dataset for learning and verifying the machine learning model, and the user's body information is added to the machine learning model learned through the dataset. It relates to an artificial intelligence-based method of predicting exercise prescription effect for hypertension and diabetes patients, a computing device, and a computer-readable medium, for outputting the exercise effect prediction information of the user in units of a week by inputting the .

A chronic disease is a disease in which the symptoms persist for a long time and the intensity of the symptoms is moderate. However, in the case of chronic diseases, since the intensity of the symptoms according to the disease is weak, the need to manage from the point of view of a chronic disease patient is not greatly felt. In most cases, continuous management of chronic diseases cannot be achieved due to weakening of the will.

On the other hand, in order to suppress the intensity of symptoms according to chronic diseases and to reduce the possibility of complications due to chronic diseases, it is important to continuously exercise from the stage of chronic diseases with mild symptoms. In particular, high blood pressure and diabetes are among chronic diseases that can be effectively managed through regular exercise. Currently, it is estimated that there are about 12 million people with high blood pressure in Korea, and in the case of diabetes, including fasting blood sugar disorder, which is the pre-diabetes stage, about 9.48 million people have these diseases. However, looking at chronic disease health statistics operated by the Korea Centers for Disease Control and Prevention, the proportion of those who practice exercise among chronic disease patients is less than 50%.

Conventional methods for hypertensive and diabetic patients to exercise exercise merely prescribe a specific exercise routine to the hypertensive and diabetic patients in a medical institution, or merely provide an exercise performance log. There are problems that depend on the will of the person. Therefore, in the case of the prior art, it is difficult to consistently perform the exercise when the interest of the surrounding or the patient's will is insufficient.
Specifically, among the prior art in the above technical field, in Korea Patent No. 10-2188766 ("Artificial intelligence-based health care service providing device", registered on December 02, 2020), the user's health information And based on exercise information, predicts a disease that is likely to develop to the user, and describes a configuration that provides an appropriate exercise habit according to the predicted disease, but this is simply to provide an exercise method that the user should do, exercise As a result, it does not provide the expected effect. Therefore, as described above, in the case of the prior art, there is a problem that cannot exert any effect at all in terms of motivating the user to exercise.
After all, in order for hypertension and diabetes patients to exercise consistently, a motivating factor is required, and therefore, the development of a new method for providing a motivating factor is required.

delete

delete

Korean Patent Registration No. 10-2188766 (Artificial intelligence-based healthcare service provision device, registered on December 02, 2020)

The present invention relates to an artificial intelligence-based method for predicting the effect of exercise prescription for hypertension and diabetes patients, a computing device, and a computer-readable medium, and more particularly, to a database by collecting patient information for a plurality of hypertension and diabetes patients. Among the patient information included in the established database, only patients who can exercise are selected to create a dataset for learning and verifying the machine learning model, and the user's body information is added to the machine learning model learned through the dataset. An object of the present invention is to provide an artificial intelligence-based method for predicting the exercise prescription effect of hypertension and diabetes patients, a computing device, and a computer-readable medium, which outputs the exercise effect prediction information of the user on a weekly basis by inputting .

In order to solve the above problems, in one embodiment of the present invention, there is provided an artificial intelligence-based method for predicting the effect of exercise prescription for hypertension and diabetes patients performed by a computing device including one or more processors and one or more memories. Collect patient information including body information, exercise prescription information, and exercise effect information for hypertensive patients and diabetic patients, and based on the collected body information, exercise prescription information, and exercise effect information included in the plurality of collected patient information, the patient database construction step of constructing a body database for the body and disease of the patient and an exercise prescription database for the exercise and exercise results performed by the patient; an exercise impossibility determination step of determining one or more patients unable to exercise based on the information on each patient's body and disease included in the body database; Age information on each of the remaining patients except for one or more patients determined to be unable to exercise in the body database, blood pressure information of the patient if the patient is a hypertensive patient, and information on the patient's blood pressure if the patient is a diabetic patient Initial data including extraction data for each patient by extracting blood glucose information and extracting exercise prescription rate information and exercise performance history information for each of the remaining patients except for one or more patients determined to be unable to exercise from the exercise prescription database Initial data set construction step of building a set; A training set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of a plurality of extracted data included in the initial dataset by a predetermined first ratio value is generated, and a plurality of extractions included in the initial dataset are generated. a learning set generating step of generating a verification set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of data by a predetermined second ratio value; A supervised learning-based machine learning model is trained using the training set, the verification set is input to the machine learning model that has learned the training set, and the output result is used as performance information of the machine learning model that learned the training set. Thus, a machine learning model learning step of storing the performance information; And when receiving service request information including body information about the user who wants to predict the exercise prescription effect, input the body information into the learned machine learning model and the corresponding user for each of one or more preset weeks after the current time An exercise effect prediction step of outputting exercise effect prediction information of

In an embodiment of the present invention, the computing device for performing the exercise prescription effect prediction method may communicate with a plurality of user terminals, and the database building step is input by each user through the plurality of user terminals. The patient information can be collected by receiving one body information, exercise prescription information, and exercise effect information.

In one embodiment of the present invention, in the exercise impossibility determination step, the systolic blood pressure value is the first a blood pressure determination step of determining the patient as a patient unable to exercise when the determination value is greater than or equal to the diastolic blood pressure value greater than or equal to the second determination value; and with respect to a heartbeat value included in the information on the body and disease for each patient that is not determined to be unable to exercise in the blood pressure determination step, the heartbeat value is equal to or greater than the third determination value, or the heartbeat value is a fourth and a heartbeat determination step of determining the patient as a patient unable to exercise when the determination value is less than or equal to the determination value.

In one embodiment of the present invention, in the exercise impossibility determining step, with respect to the glucose level included in the information on the body and disease for each patient who is not determined to be unable to exercise in the heartbeat determination step, the glucose level is the first a glucose level determination step of determining the patient as a patient unable to exercise when the 5th determination value or more or the glucose level is below the 6th determination value; With respect to the cholesterol level included in the information on the body and disease of each patient who is not determined to be unable to exercise in the glucose level determination step, if the cholesterol level is greater than or equal to the 7th determination value, the patient is classified as a patient unable to exercise. Cholesterol level determination step to determine; With respect to the low-density cholesterol level included in the information on the body and disease of each patient who is not determined to be unable to exercise in the cholesterol level determination step, if the low-density cholesterol level is equal to or greater than the 8th determination value, the patient cannot exercise Low-density cholesterol level determination step to determine the patient; With respect to the high-density cholesterol level included in the body and disease information for each patient who is not determined to be unable to exercise in the low-density cholesterol level determination step, if the high-density cholesterol level is below the ninth determination value, the patient cannot exercise High-density cholesterol level determination step to determine a patient; and with respect to the glycated hemoglobin level included in the information on the body and disease of each patient for which exercise is not determined to be impossible in the high-density cholesterol level determination step, if the glycated hemoglobin level is equal to or greater than the 10th determination value, the patient is not exercised The method may further include a glycated hemoglobin level determination step of determining an impossible patient.

In an embodiment of the present invention, the supervised learning-based machine learning model includes a first detailed machine learning model to a tenth detailed machine learning model, and the machine learning model learning step includes a plurality of machine learning models included in the training set. The extracted data is divided into a plurality of preset age groups and each patient with hypertension or diabetes, and each of the plurality of extracted data divided into a plurality of numbers corresponds to any one of the first detailed machine learning model to the tenth detailed machine learning model. It can be used to train a detailed machine learning model.

In one embodiment of the present invention, the exercise effect prediction step, when receiving service request information including body information about the user who wants to predict the exercise prescription effect, the age of the user included in the body information Enter the user's body information into the learned detailed machine learning model corresponding to the information and the user's high blood pressure or diabetes mellitus. The systolic blood pressure predicted value and the diastolic blood pressure predicted value of the user are output as exercise effect prediction information, and when the user has diabetes, the predicted value of the user's glucose level for each preset one or more weeks after the current time is output as exercise effect prediction information can do.

In an embodiment of the present invention, the exercise prescription effect prediction method further comprises a machine learning model updating step performed after the time when the machine learning model is learned through the training set through the machine learning model learning step; and , the machine learning model updating step, when updating the body database and the exercise prescription database by collecting a predetermined number of additional patient information in the database building step, the updated body database and the updated exercise prescription database can be used to perform additional learning on the learned machine learning model.

In order to solve the above problems, in one embodiment of the present invention, a computing device including one or more processors and one or more memories to perform an artificial intelligence-based exercise prescription effect prediction method for a patient with hypertension and diabetes, collects patient information including body information, exercise prescription information, and exercise effect information for hypertensive patients and diabetic patients of database construction step of constructing a body database for the patient's body and disease and an exercise prescription database for the exercise and exercise results performed by the patient; an exercise impossibility determination step of determining one or more patients unable to exercise based on the information on each patient's body and disease included in the body database; Age information on each of the remaining patients except for one or more patients determined to be unable to exercise in the body database, blood pressure information of the patient if the patient is a hypertensive patient, and information on the patient's blood pressure if the patient is a diabetic patient Initial data including extraction data for each patient by extracting blood glucose information and extracting exercise prescription rate information and exercise performance history information for each of the remaining patients except for one or more patients determined to be unable to exercise from the exercise prescription database Initial data set construction step of building a set; A training set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of a plurality of extracted data included in the initial dataset by a predetermined first ratio value is generated, and a plurality of extractions included in the initial dataset are generated. a learning set generating step of generating a verification set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of data by a predetermined second ratio value; A supervised learning-based machine learning model is trained using the training set, the verification set is input to the machine learning model that has learned the training set, and the output result is used as performance information of the machine learning model that learned the training set. Thus, a machine learning model learning step of storing the performance information; And when receiving service request information including body information about the user who wants to predict the exercise prescription effect, input the body information into the learned machine learning model and the corresponding user for each of one or more preset weeks after the current time It provides a computing device that performs the exercise effect prediction step of outputting the exercise effect prediction information.

In order to solve the above problems, in one embodiment of the present invention, by executing a computer program in a computing device including one or more processors and one or more memories, artificial intelligence-based hypertension and diabetes performed by the computing device A computer-readable medium for implementing a method for predicting a patient's exercise prescription effect, the computer-readable medium storing instructions for causing the computing device to perform the following steps, the following steps comprising: collects patient information including body information, exercise prescription information, and exercise effect information for hypertensive patients and diabetic patients of database construction step of constructing a body database for the patient's body and disease and an exercise prescription database for the exercise and exercise results performed by the patient; an exercise impossibility determination step of determining one or more patients unable to exercise based on the information on each patient's body and disease included in the body database; Age information on each of the remaining patients except for one or more patients determined to be unable to exercise in the body database, blood pressure information of the patient if the patient is a hypertensive patient, and information on the patient's blood pressure if the patient is a diabetic patient Initial data including extraction data for each patient by extracting blood glucose information and extracting exercise prescription rate information and exercise performance history information for each of the remaining patients except for one or more patients determined to be unable to exercise from the exercise prescription database Initial data set construction step of building a set; A training set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of a plurality of extracted data included in the initial dataset by a predetermined first ratio value is generated, and a plurality of extractions included in the initial dataset are generated. a learning set generating step of generating a verification set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of data by a predetermined second ratio value; A supervised learning-based machine learning model is trained using the training set, the verification set is input to the machine learning model that has learned the training set, and the output result is used as performance information of the machine learning model that learned the training set. Thus, a machine learning model learning step of storing the performance information; And when receiving service request information including body information about the user who wants to predict the exercise prescription effect, input the body information into the learned machine learning model and the corresponding user for each of one or more preset weeks after the current time It provides a computer-readable medium, including a; exercise effect prediction step of outputting the exercise effect prediction information.

According to an embodiment of the present invention, the machine learning model is trained according to the collected patient information, and the user's body information is input to the learned machine learning model, thereby predicting the exercise effect when the user performs weekly exercise. By outputting the information, it is possible to exert an effect of providing motivation for exercise performance to a user with high blood pressure or diabetes.

According to an embodiment of the present invention, in the exercise impossibility determination step, exercise using determination values for blood pressure, heart rate, glucose, cholesterol, and glycated hemoglobin in a database constructed using a plurality of collected patient information. By determining this impossible patient, the machine learning model can learn information about the remaining patients except for the patient unable to exercise from the database, thereby exhibiting the effect of more accurately predicting the exercise effect.

According to an embodiment of the present invention, the machine learning model includes a plurality of detailed machine learning models according to age groups and hypertension or diabetes, and each detailed machine learning model includes a training set for a corresponding age group and hypertension or diabetes. Because it is learned through the user's age and disease, it is possible to exert the effect of more accurately predicting the exercise effect through a detailed machine learning model corresponding to the user's age and disease.

According to an embodiment of the present invention, in the machine learning model updating step, when patient information corresponding to a predetermined number is additionally collected in the database building step, the machine learning model learned based on the additionally collected patient information is additionally learned , so that the performance of the machine learning model can be improved, and thus, it is possible to exert the effect of more accurately predicting the exercise effect for the user.

1 schematically shows components for performing an artificial intelligence-based exercise prescription effect prediction method for a patient with hypertension and diabetes according to an embodiment of the present invention.
FIG. 2 schematically illustrates detailed steps of a method for predicting an exercise prescription effect for a patient with hypertension and diabetes based on artificial intelligence according to an embodiment of the present invention.
3 schematically illustrates a process of building a database through patient information collected from a user terminal according to an embodiment of the present invention.
4 schematically illustrates a process of establishing an initial data set by determining a patient unable to exercise by the exercise impossibility determination step according to an embodiment of the present invention.
5 schematically shows detailed steps of the exercise impossible determination step according to an embodiment of the present invention.
6 schematically illustrates a process of learning a machine learning model using a database constructed according to an embodiment of the present invention.
7 schematically illustrates a plurality of detailed machine learning models included in the machine learning model according to an embodiment of the present invention.
8 schematically shows a process of performing the exercise effect prediction step according to an embodiment of the present invention.
9 schematically shows a process of performing a machine learning model update step according to an embodiment of the present invention.
10 schematically shows the overall flow of a method for predicting an exercise prescription effect for a patient with high blood pressure and diabetes based on artificial intelligence according to an embodiment of the present invention.
11 schematically illustrates an internal configuration of a computing device according to an embodiment of the present invention.

In the following, various embodiments and/or aspects are now disclosed with reference to the drawings. In the following description, for purposes of explanation, numerous specific details are set forth to provide a thorough understanding of one or more aspects. However, it will also be recognized by one of ordinary skill in the art that such aspect(s) may be practiced without these specific details. The following description and accompanying drawings set forth in detail certain illustrative aspects of one or more aspects. These aspects are illustrative, however, and some of the various methods in principles of various aspects may be employed, and the descriptions set forth are intended to include all such aspects and their equivalents.

Further, various aspects and features will be presented by a system that may include a number of devices, components and/or modules, and the like. It is also noted that various systems may include additional apparatuses, components, and/or modules, etc. and/or may not include all of the apparatuses, components, modules, etc. discussed with respect to the drawings. must be understood and recognized.

As used herein, “embodiment”, “example”, “aspect”, “exemplary”, etc. may not be construed as an advantage or an advantage in any aspect or design described above over other aspects or designs. . The terms '~part', 'component', 'module', 'system', 'interface', etc. used below generally mean a computer-related entity, for example, hardware, hardware A combination of and software may mean software.

Also, the terms "comprises" and/or "comprising" mean that the feature and/or element is present, but excludes the presence or addition of one or more other features, elements and/or groups thereof. should be understood as not

Also, terms including an ordinal number such as first, second, etc. may be used to describe various elements, but the elements are not limited by the terms. The above terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, a first component may be referred to as a second component, and similarly, a second component may also be referred to as a first component. and/or includes a combination of a plurality of related listed items or any of a plurality of related listed items.

In addition, in the embodiments of the present invention, unless otherwise defined, all terms used herein, including technical or scientific terms, are generally understood by those of ordinary skill in the art to which the present invention belongs. have the same meaning as Terms such as those defined in a commonly used dictionary should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in an embodiment of the present invention, an ideal or excessively formal meaning is not interpreted as

1 schematically shows components for performing an artificial intelligence-based exercise prescription effect prediction method for a patient with hypertension and diabetes according to an embodiment of the present invention.

As shown in FIG. 1 , the computing device 1000 performing the method of predicting the effect of exercise prescription for hypertension and diabetes patients based on artificial intelligence of the present invention communicates with a plurality of user terminals 2000 . On the other hand, a separate application for performing communication with the computing device 1000 is installed in each of the plurality of user terminals 2000 , and each user of the plurality of user terminals 2000 can use the application You can input patient information including body information, information on the exercise you have been prescribed, and information on the performance and effect (exercise performance result) of the prescribed exercise, and the application uses the patient information input from the user. It is transmitted to the computing device 1000 .

On the other hand, as shown in FIG. 1 , the computing device 1000 includes a database building unit 1100 , an exercise impossibility determining unit 1200 , an initial data set building unit 1300 , a learning set generating unit 1400 , and a machine. It includes a learning model learning unit 1500 , an exercise effect prediction unit 1600 , and a machine learning model update unit 1700 .

The database construction unit 1100 performs the database construction step (S10) to build a database using the patient information received from the plurality of user terminals 2000 described above. Specifically, the database building unit 1100 builds two databases, processes the body information included in the plurality of patient information, and builds a body database for a plurality of users (patients) and their diseases. , information on the prescribed exercise included in the plurality of patient information and information on the performance and effect of the prescribed exercise are processed to construct an exercise prescription database for exercise and exercise results for a plurality of users. The body database and the exercise prescription database may be used as data for learning a machine learning model used to predict the effect of exercise prescription for hypertension and diabetes patients.

The exercise impossibility determining unit 1200 performs an exercise impossibility determining step (S11). In the present invention, rather than allowing the machine learning model to learn all the information constructed in the body database and the exercise prescription database, the present invention predicts the effect of exercise prescription for hypertension and diabetes patients, and predicts these effects. For ease of use, the machine learning model is trained using only the information about the remaining patients except for the information about the patient determined to be unable to exercise in each database. For this configuration, in the exercise impossibility determination step (S11), it determines the patients unable to exercise in the built-up database.

The initial data set building unit 1300 performs the initial data set building step (S12). In the initial data set construction step (S12), the exercise impossibility determination step (S11) is performed, and information about the remaining patients is used except for one or more patients determined to be unable to exercise in the body database and the exercise prescription database. Build an initial dataset for training the machine learning model. The initial data set can generate extracted data by simply combining information from the body database for each patient capable of exercise and information from the exercise prescription database to build an initial data set including a plurality of extracted data for each patient, but it is preferable An initial dataset can be constructed by extracting only the information necessary to learn from the machine learning model and generating extracted data.

The training set generation unit 1400 generates a training set and a verification set by dividing the extracted data for each patient included in the initial dataset at a predetermined ratio by performing the training set generation step S13. The training set corresponds to a dataset for learning the machine learning model, and the verification set corresponds to a dataset for evaluating the performance of the machine learning model learned through the training set.

The machine learning model learning unit 1500 performs a machine learning model learning step ( S14 ) to learn a machine learning model using the generated learning set. Specifically, the machine learning model includes a multi-layer perceptron (MLP), which is one of the deep learning algorithms, and may perform learning based on supervised learning. Accordingly, each extracted data included in the learning set may be given a labeling value corresponding to the correct answer so that the machine learning model can learn based on supervised learning.

The exercise effect prediction unit 1600 performs the exercise effect prediction step (S15), the exercise effect prediction step (S15) through the application installed in the user terminal 2000, the user has high blood pressure and / or diabetes, When the user requests a service for predicting the effect of exercise prescription for hypertension and/or diabetes, according to the service request, the user's exercise effect prediction information is output using the learned machine learning model, and It is provided to the user terminal 2000 of the user. In more detail, the exercise effect prediction information may correspond to information on physical changes (such as a blood pressure predictive value and a diabetes value predictive value) when the user performs an exercise for each preset week from the current time point.

The machine learning model update unit 1700 performs a machine learning model update step. In the present invention, the exercise prescription effect of hypertension and diabetes patients is predicted using a machine learning model learned using information contained in the initially constructed database, and in the present invention, patient information is additionally updated in the initially constructed database. In this case, the machine learning model update step additionally trains the initially learned machine learning model by using additionally updated information in the database. As described above, in the present invention, it is possible to predict the effect of exercise prescription for hypertension and diabetes patients by using the additionally learned machine learning model, and it is possible to provide a more accurate prediction result to the user.

On the other hand, although not shown in FIG. 1, the computing device 1000 further includes a DB, and the DB includes a body database and an exercise prescription database built by the database building unit 1100, and an initial data set construction step (S12) The training set and the verification set generated by the initial dataset and learning set generating unit 1400 constructed in Each of the initially learned machine learning model, each of the machine learning models that are additionally learned and updated through the machine learning model update step can be stored, and each of the verification results output by inputting a verification set to each version of the machine learning model is stored. can

FIG. 2 schematically illustrates detailed steps of a method for predicting an exercise prescription effect for a patient with hypertension and diabetes based on artificial intelligence according to an embodiment of the present invention.

As shown in FIG. 2 , an artificial intelligence-based exercise prescription effect prediction method for hypertensive and diabetic patients performed by a computing device 1000 including one or more processors and one or more memories. Collects patient information including body information, exercise prescription information, and exercise effect information for the patient, and provides information on the patient's body and disease based on body information, exercise prescription information, and exercise effect information included in the collected plurality of patient information. Database building step (S10) of building a body database and an exercise prescription database for the exercise and exercise results performed by the patient; an exercise impossibility determination step (S11) of determining one or more patients unable to exercise based on the information on each patient's body and disease included in the body database; Age information on each of the remaining patients except for one or more patients determined to be unable to exercise in the body database, blood pressure information of the patient if the patient is a hypertensive patient, and information on the patient's blood pressure if the patient is a diabetic patient Initial data including extraction data for each patient by extracting blood glucose information and extracting exercise prescription rate information and exercise performance history information for each of the remaining patients except for one or more patients determined to be unable to exercise from the exercise prescription database Initial data set construction step (S12) of building a set; A training set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of a plurality of extracted data included in the initial dataset by a predetermined first ratio value is generated, and a plurality of extractions included in the initial dataset are generated. a learning set generating step (S13) of generating a verification set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of data by a predetermined second ratio value (S13); A supervised learning-based machine learning model is trained using the training set, the verification set is input to the machine learning model that has learned the training set, and the output result is used as performance information of the machine learning model that learned the training set. Thus, a machine learning model learning step of storing the performance information (S14); And when receiving service request information including body information about the user who wants to predict the exercise prescription effect, input the body information into the learned machine learning model and the corresponding user for each of one or more preset weeks after the current time It may include a; exercise effect prediction step (S15) of outputting the exercise effect prediction information.

Specifically, the database building step (S10) collects patient information for each user from a plurality of user terminals 2000, and builds a body database and an exercise prescription database using the collected patient information.

More specifically, the patient information collected from the user terminal 2000 includes basic body information of a user with high blood pressure or diabetes, exercise prescription information prescribed by the user through a medical institution, and performance history and performance of the exercise prescribed by the user. It may include exercise effect information including exercise results according to the

The body information may be directly generated by a user input through an application installed in the user terminal 2000, but in another embodiment of the present invention, the body information may correspond to an Electronic Medical Record (EMR). Therefore, in the database building step (S10), the electronic medical record for each patient can be received from a separate external server in which the electronic medical record is stored, and preferably, the user can receive the electronic medical record through the application. It is possible to input whether or not to consent to the provision of , and in the database construction step ( S10 ), only electronic medical records for the user who have accepted the consent to the provision may be collected.

In the exercise impossibility determination step (S11), exercise is impossible for each patient according to the information on the body and disease for each patient included in the body database built through the database building step (S10) and one or more preset determination criteria. Determine whether the patient is Specifically, the exercise impossibility determination step (S11) is performed in consideration of a plurality of physical numerical values related to age, hypertension, and diabetes for each patient included in the body database, or when exercise is performed, hypertension and diabetes For patients whose symptoms worsen or do not relieve

Through this configuration, since the machine learning model is learned by considering only the database for the remaining patients except for the patients determined to be unable to exercise through the exercise impossibility determination step (S11), data on the hypertensive and diabetic patients who cannot exercise Compared with learning including including

In the initial data set construction step (S12), a machine learning model using data of the remaining patients, except for one or more patients determined to be unable to exercise, among the data for each patient included in the body database and exercise prescription database built in advance. It learns and builds an initial dataset corresponding to a dataset that can verify the performance of the trained machine learning model.

Specifically, in the initial data set construction step (S12), age information about the remaining patients except for one or more patients determined to be unable to exercise in the body database, and blood pressure information of the patient when the patient is a hypertensive patient, When the patient is a diabetic patient, the blood glucose information of the patient is extracted, and the exercise prescription rate information and exercise performance history information for the remaining patients except for one or more patients determined to be unable to exercise from the exercise prescription database are extracted, and each Extracted data including information extracted for each patient is generated, and finally, the initial data set may include a plurality of extracted data for each patient.

In the learning set generating step (S13), a training set and a verification set are generated by dividing the extracted data for all the plurality of patients included in the constructed initial data set at a predetermined ratio. Specifically, the number of extracted data included in the training set generated in the learning set creation step S13 corresponds to the number obtained by multiplying the number of extracted data for all patients included in the initial dataset by the first ratio value. And, the number of extracted data included in the verification set may correspond to the number obtained by multiplying the number of extracted data for all the plurality of patients included in the initial data set by the second ratio value. For example, the first ratio value and the second ratio value may correspond to 7 and 3, respectively. Meanwhile, a plurality of extracted data included in the training set and a plurality of extracted data included in the verification set may be randomly selected.

In the machine learning model learning step (S14), the machine learning model in the stored initial state, preferably the machine learning model implemented based on supervised learning, is trained using the generated training set as learning data, and the verification set is used. Evaluate the performance of the trained machine learning model using

Specifically, in the machine learning model learning step (S14), the machine learning model learned through the training set is separately stored in the computing device 1000, and the performance of the learned machine learning model evaluated through the verification set Information is separately stored in the computing device 1000 . Through such a configuration, when the machine learning model learned based on the database additionally updated in the above-described machine learning model update step is additionally trained, machine learning with high performance among the stored machine learning model and the additionally learned machine learning model to use the model. In general, the performance of the additionally trained machine learning model appears to be high, but when overfitting due to excessive learning, the performance deteriorates compared to the previous version of the machine learning model. Each version of the machine learning model and performance information on each version of the machine learning model may be stored in the computing device 1000 .

In the exercise effect prediction step (S15), the exercise prescription effect for the user who requested the service is predicted using the learned machine learning model, and the result is provided to the user terminal 2000 of the user.

Specifically, the exercise effect prediction step (S15) receives the service request information from the user terminal (2000). The service request information includes body information about a user requesting a service, and the body information is input to the learned machine learning model, and exercise prescription effect prediction information for the user may be output. Meanwhile, the user requesting the service may preferably correspond to a user (patient) with hypertension and/or diabetes. Then, the exercise effect prediction step (S15) outputs the exercise prescription effect prediction information for the user by inputting the body information included in the received service request information to the learned machine learning model.

On the other hand, as described above, when the computing device 1000 includes a plurality of versions of the machine learning model, exercise prescription for the user by using the machine learning model with the highest performance in consideration of the performance information of each version Effect prediction information can be output.

In addition, the output exercise prescription effect prediction information does not simply predict the effect when a user with high blood pressure or diabetes performs an exercise, but a plurality of weeks preset based on the time when the service is specifically requested. ) can include predicted values for different physical values. For example, if the user who requested the service is a hypertensive patient and the user requested the service in the 1st week of October, the exercise prescription effect prediction information for the user is calculated for each of the 5 weeks from the 1st week of October. It may include a predicted value for the blood pressure level.

As described above, in the present invention, when the user specifically performs the prescribed exercise, the user is provided with a predicted value for a specific physical value for hypertension or diabetes to the user, thereby providing a substantial motivation for the user to consistently perform the exercise. can perform

3 schematically illustrates a process of building a database through patient information collected from a user terminal 2000 according to an embodiment of the present invention.

As shown in FIG. 3 , the computing device 1000 performing the exercise prescription effect prediction method may communicate with a plurality of user terminals 2000 , and the database building step ( S10 ) includes the plurality of user terminals 2000 . The patient information may be collected by receiving body information, exercise prescription information, and exercise effect information input by each user through the user terminal 2000 .

Specifically, the computing device 1000 may communicate with the user terminals 2000 , and in the database building step S10 , the patient information is received from each user terminal 2000 , and as described above, a plurality of It is possible to build a body database and exercise prescription database for patients of

More specifically, the patient information received from the user terminal 2000 includes body information, exercise prescription information, and exercise effect information about the patient (user). The body information includes basic information such as an age of the corresponding patient and a physical numerical value of the corresponding patient. For example, the physical numerical value may include values for each of a blood pressure level, a heart rate level, a glucose level, a cholesterol level, a low-density cholesterol level, a high-density cholesterol level, and a glycated hemoglobin level. The exercise prescription information may include information on the exercise prescribed by the patient at the medical institution, and information related to the routine (exercise event, repetition number, performance schedule, etc.) of the specifically prescribed exercise. The exercise effect information may include information on the history of performing the exercise prescribed by the patient and information on the exercise result according to the performance history, and the information on the exercise result indicates that the patient performed the prescribed exercise. It may include physical numerical values measured when

Meanwhile, in the database building step (S10), as shown in FIG. 3 , a body database is built based on body information among the received patient information for a plurality of patients, and exercise is performed based on exercise prescription information and exercise effect information. Build a prescription database.

In another embodiment of the present invention, the database building step (S10) is not performed only in the initial stage, but is always performed whenever patient information is received from the user terminal 2000, or within each period according to a preset period. It is possible to update the body database and exercise prescription database built based on the received one or more patient information.

4 schematically shows a process of establishing an initial data set by determining a patient unable to exercise by the exercise impossibility determination step (S11) according to an embodiment of the present invention.

As shown in FIG. 4 , data for a plurality of patients may be established in the body database and the exercise prescription database constructed through the database construction step ( S10 ). Specifically, the body database may store information on the body and diseases of each patient, and the exercise prescription database may store information on the exercise and exercise results performed by each patient.

Meanwhile, in the exercise impossibility determination step (S11), it is determined whether each patient is a patient unable to exercise based on information about a plurality of patient-specific bodies and diseases stored in the body database. The process of determining whether the patient is a patient incapable of exercise based on the information on the patient's body and the disease possessed will be described in detail with reference to FIG. 5 to be described later.

As described above, in the exercise impossibility determination step (S11), at least one patient among a plurality of patients included in the body database is determined as a patient unable to exercise (patient #2 and patient #4 in FIG. 4), and the initial In the data set construction step (S12), the information from the body database and the exercise prescription database for the remaining patients (Patient #1, Patient #3, and Patient #5 in Fig. 4) that are not determined to be incapable of exercise are used to initially Create a dataset.

The initial dataset includes extracted data for each of the remaining patients who are not determined to be unable to exercise, and each of the plurality of extracted data includes age information of the patient, blood pressure information of the patient when the patient is a hypertensive patient, If the patient is a diabetic patient, the patient's blood sugar information, exercise prescription rate information of the patient, and exercise performance history information of the patient are included. The age information, blood pressure information, or blood sugar information may be extracted from the body database, and the exercise prescription rate information and the exercise performance history information may be extracted from the exercise prescription database.

In this way, the initial dataset includes extracted data for each patient who is not determined to be an impossible patient, trains a machine learning model using the initial dataset, and verifies the learned machine learning model, so the machine The learning model can show higher performance in predicting the effect of the user's exercise prescription by learning only the extracted data about the patient who can exercise.

5 schematically shows detailed steps of the exercise impossible determination step (S11) according to an embodiment of the present invention.

As shown in FIG. 5 , in the exercise impossibility determination step ( S11 ), the systolic blood pressure value for the systolic blood pressure value and the diastolic blood pressure value included in the body and disease information for each patient included in the body database a blood pressure determination step (S20) of determining the patient as a patient unable to exercise when the first determination value or more or the diastolic blood pressure value is greater than or equal to the second determination value; and with respect to the heartbeat value included in the information on the body and disease for each patient that is not determined to be unable to exercise in the blood pressure determination step (S20), the heartbeat value is equal to or greater than the third determination value, or the heartbeat value and a heartbeat determination step (S21) of determining the patient as a patient unable to exercise when the fourth determination value is less than or equal to the fourth determination value.

As shown in FIG. 5 , in the exercise impossibility determination step ( S11 ), it is determined whether the corresponding patient is a patient unable to exercise by sequentially applying a plurality of determination criteria. To this end, the exercise impossibility determination step (S11) sequentially applies information on the body and disease of each of a plurality of patients stored in the body database to a plurality of standards, and when the information meets a specific standard, the corresponding patient may be determined as a patient unable to exercise, and when the corresponding information does not meet all of a plurality of criteria, the patient may be determined as a patient capable of exercise.

The patient who is unable to exercise as defined in the present invention may refer to any patient who is physically unable to exercise due to a decrease in physical function, or who is physically able to exercise but is concerned about side effects rather than the effects of exercise.

Specifically, in the exercise impossibility determination step (S11), the blood pressure determination step (S20) is first performed with respect to the information on the body and disease of a specific patient included in the body database. In the blood pressure determination step (S20), the systolic blood pressure value is greater than or equal to a predetermined first determination value with respect to the blood pressure value included in the information on the body and disease, more specifically, the systolic blood pressure value and the diastolic blood pressure value, When the diastolic blood pressure value is greater than or equal to the second predetermined value, it is determined that the specific patient cannot exercise. Preferably, the first determination value may correspond to 160 mmHg and the second determination value may correspond to 90 mmHg.

When it is not determined that the patient is unable to exercise through the blood pressure determination step S20, the heartbeat determination step S21 is performed with respect to information on the body and disease of the patient. In the heartbeat determination step S21, with respect to the heartbeat value included in the information on the body and disease, the heartbeat value is equal to or greater than a predetermined third determination value, or the heartbeat value is a predetermined fourth determination value. If the value is less than the value, it is determined that the specific patient is unable to exercise. Preferably, the third determination value may correspond to 100 bpm and the fourth determination value may correspond to 50 bpm.

Meanwhile, the exercise impossibility determination step S11 may further include criteria for determining whether the patient is unable to exercise in addition to the blood pressure determination step S20 and the heartbeat determination step S21 described above.

As shown in FIG. 5 , the exercise impossibility determination step (S11) is performed with respect to the glucose level included in the information on the body and disease of each patient that is not determined to be unable to exercise in the heartbeat determination step (S21). , a glucose level determination step (S22) of determining the patient as a patient unable to exercise when the glucose level is equal to or greater than the fifth determination value or the glucose level is less than or equal to the sixth determination value; With respect to the cholesterol level included in the information on the body and disease of each patient that is not determined to be unable to exercise in the glucose level determination step (S22), if the cholesterol level is equal to or greater than the 7th determination value, the patient cannot exercise Cholesterol level determination step of determining a patient (S23); With respect to the low-density cholesterol level included in the information on the body and disease of each patient that is not determined to be unable to exercise in the cholesterol level determination step (S23), if the low-density cholesterol level is equal to or greater than the 8th determination value, exercise the patient Low-density cholesterol level determination step (S24) for determining this impossible patient; With respect to the high-density cholesterol level included in the information on the body and disease for each patient who is not determined to be unable to exercise in the low-density cholesterol level determination step (S24), if the high-density cholesterol level is below the ninth determination value, the patient High-density cholesterol level determination step (S25) for determining the patient is unable to exercise; and with respect to the glycated hemoglobin level included in the information on the body and disease for each patient who is not determined to be unable to exercise in the high-density cholesterol level determination step (S25), when the glycated hemoglobin level is equal to or greater than the tenth determination value, the patient The method may further include a glycated hemoglobin level determination step (S26) of determining a patient unable to exercise.

Specifically, when it is not determined that the patient is unable to exercise through the heartbeat determination step S21, the glucose level determination step S22 is performed based on information about the body and disease of the patient. In the glucose level determination step (S22), with respect to the glucose level included in the information on the body and disease, the glucose level value is equal to or greater than a predetermined fifth determination value, or the glucose level value is a predetermined sixth determination value If the value is less than the value, it is determined that the specific patient is unable to exercise. Preferably, the fifth determination value may correspond to 300 mg/dL and the sixth determination value may correspond to 100 mg/dL.

When it is not determined that the patient is unable to exercise through the glucose level determination step S22, the cholesterol level determination step S23 is performed based on information on the body and disease of the patient. In the cholesterol level determination step (S23), with respect to the cholesterol level value included in the information on the body and disease, when the cholesterol level value is greater than or equal to a predetermined seventh determination value, it is determined that the specific patient cannot exercise. Preferably, the cholesterol level may correspond to a total cholesterol level, and the seventh determination value may correspond to 260 mg/dL.

If it is not determined that the patient is unable to exercise through the cholesterol level determination step (S23), the low-density cholesterol level determination step (S24) is performed based on information on the body and disease for the patient. In the low-density cholesterol level determination step (S24), when the low-density cholesterol level value included in the body and disease information is equal to or greater than the eighth determination value, it is determined that the specific patient cannot exercise. Preferably, the eighth determination value may correspond to 160 mg/dL.

If the patient is not determined to be unable to exercise through the low-density cholesterol level determination step (S24), the high-density cholesterol level determination step (S25) is performed based on information on the body and disease for the patient. In the high-density cholesterol level determination step (S25), when the high-density cholesterol level value included in the body and disease information is greater than or equal to the ninth determination value, it is determined that the specific patient cannot exercise. Preferably, the ninth determination value may correspond to 35 mg/dL.

Finally, when it is not determined that the patient is unable to exercise through the high-density cholesterol level determination step S25, the glycated hemoglobin level determination step S26 is performed based on information about the body and disease of the patient. do. In the glycated hemoglobin level determination step S26, when the glycated hemoglobin level value included in the body and disease information is equal to or greater than the tenth determination value, it is determined that the specific patient cannot exercise, and the glycated hemoglobin level value is the If it is less than the 10th determination value, it is determined that the specific patient is able to exercise. Preferably, the tenth determination value may correspond to 10.0%.

As such, the exercise impossibility determination step (S11) includes a plurality of steps (S20 to S26) corresponding to a plurality of criteria, and each step is sequentially connected, so that patients meeting all the steps can exercise The patient can finally be judged. Meanwhile, specific numerical values of the above-described first to tenth determination values may be changed to arbitrary values by a designer.

In addition, the sequence of each sequentially connected step may also be arbitrarily changed by the designer, but as shown in FIG. 5 , the blood pressure determination step (S20), the heartbeat determination step (S21), the glucose level determination step ( S22), the cholesterol level determination step (S23), the low-density cholesterol level determination step (S24), the high-density cholesterol level determination step (S25), and the glycated hemoglobin level determination step (S26) are preferably configured in this order. The above order corresponds to an order with a high probability that the patient is determined to be unable to exercise. Therefore, in applying the exercise impossibility determination step (S11) to each of a plurality of patients included in the body database, the patient is treated as a patient unable to exercise with the highest probability in the first step of the exercise impossibility determination step (S11). In the case of determination, since the following steps do not need to be additionally performed, it is possible to reduce the resources and time required to perform the exercise impossibility determination step ( S11 ) in the computing device 1000 of the present invention.

6 schematically illustrates a process of learning a machine learning model using a database constructed according to an embodiment of the present invention.

6A corresponds to a diagram schematically showing processes from the above-described database building step to the machine learning model learning step. Through step S30 included in the database building step, patient information corresponding to the CMS data shown in FIG. 6(A) is collected in conjunction with the application installed in the user terminal 2000. The CMS data corresponds to data managed through a commonly used content management system, and the data may correspond to patient information of the present invention. In addition, in step S31 included in the database building step, the file format of the patient information is converted according to the file format required by the database (body database and exercise prescription database) to be built.

In the exercise impossibility determination step (S32), it is determined whether the corresponding patient is a patient unable to exercise based on the information for each patient stored in the body database. As shown in (A) of FIG. 6 , when it is determined that the patient can exercise in the exercise impossibility determination step S32 (True), extract data for the patient is generated in the initial data set construction step. (S34), and when it is determined that the patient is unable to exercise in the movement impossible determination step (S32) (False), the movement impossible determination step (S32) may be performed for the next patient (Index+1).

On the other hand, the initial dataset constructed in the initial dataset construction step includes extracted data for one or more patients determined to be able to exercise, but as shown in FIG. 6(A), the initial dataset is diabetes data a set and a high blood pressure dataset, wherein the diabetic dataset includes extracted data for a patient determined to be able to exercise among patients with diabetes (S36), and the hypertension dataset includes data for exercise among patients with hypertension (S36). It may include extraction data for the patient determined to be (S37).

In this way, in the exercise impossibility determination step (S32), it is determined whether a specific patient is unable to exercise, and when it is determined that the specific patient is able to exercise, the process of generating and including the extracted data for the patient in the initial dataset is described above. It can be performed for all patients included in the body database. Therefore, as in step S38, the Index value starts from 0, and when it is determined that a specific patient cannot exercise, or when it is determined that a specific patient can exercise and extracted data is generated, the Index value increases sequentially, Finally, until the Index value is equal to or greater than the number of data of all patients included in the body database, the exercise impossibility determination step (S32) and the initial data set construction steps (S34 to S37) may be sequentially repeated.

As described above, when it is finally determined whether or not exercise is impossible for all patients included in the body database and an initial dataset is constructed, in the machine learning model learning step, the machine learning model is based on the constructed initial dataset. can be trained, and exercise effect prediction information for the user who requested the service can be output through the machine learning model learned in this way.

FIG. 6B schematically shows detailed processes for step S32 shown in FIG. 6A . As described above, in an embodiment of the present invention, an initial data set including extracted data for each of the remaining patients except for one or more patients determined to be unable to exercise through the exercise impossibility determination step may be generated.

On the other hand, as shown in (B) of Figure 6, in another embodiment of the present invention, in the initial data set construction step (S40), extract data for all patients included in the body database is generated, and all patients By deleting (S42) the extracted data for one or more patients determined to be unable to exercise through the exercise impossibility determination step among the extracted data for the exercise, an initial data set may be finally generated (S43). At this time, a separate normalization step (S41) is performed for the extracted data for the remaining patients who are not determined to be unable to exercise, and the initial data set may include one or more normalized extracted data.

FIG. 6(C) schematically shows detailed processes for step S39 shown in FIG. 6(A). The machine learning model of the present invention includes a plurality of detailed machine learning models, and each detailed machine learning model learns extracted data of different age groups. On the other hand, each detailed machine learning model before learning may be configured in the same form, but in another embodiment of the present invention, each detailed machine learning model is different from each other in order to properly learn the extracted data of the age group to be learned. It may be of the form.

On the other hand, as shown in (C) of Figure 6, the machine learning model learning step is an initial dataset, preferably one or more extractions corresponding to the first reference age in the training set used to learn the machine learning model. Data is loaded (S50). Here, the first reference age may correspond to the 20s. Next, it is determined whether the loaded first reference age is less than a preset critical age (S51), and if the first reference age is less than the critical age, a plurality of the first reference age in a specific detailed machine learning model The extracted data is learned (S52). Here, the critical age group may correspond to the 70s.

Thereafter, the machine learning model learning step gives a predetermined weight to the first reference age group (S53). The predetermined weight may correspond to 10, and the reference age group to which the predetermined weight is assigned may correspond to 30s. Then, step S51 is performed again, it is determined whether the reference age to which a predetermined weight is given is less than the threshold age, and if it is less than the extracted data of a plurality of patients in their 30s, specific detailed machine learning corresponding to their 30s. train the model. Such a process is repeatedly performed, and finally predetermined weights are accumulated, and may be repeatedly performed until the weighted reference age group corresponds to the 70s.

Accordingly, the number of the plurality of detailed machine learning models included in the machine learning model may correspond to the number of age groups from the initial reference age band to less than the critical age band. For example, when the first reference age is 20s and the critical age is 70s, the plurality of detailed machine learning models may be a total of 5 corresponding to the number of age groups in their 20s or more and less than 70s.

7 schematically illustrates a plurality of detailed machine learning models included in the machine learning model according to an embodiment of the present invention.

7, the supervised learning-based machine learning model includes a first detailed machine learning model to a tenth detailed machine learning model, and the machine learning model learning step (S14) is included in the training set The plurality of extracted data is divided by a plurality of preset age groups and hypertension or diabetes patients, and each of the plurality of extracted data divided into a plurality of numbers is any of the first detailed machine learning model to the tenth detailed machine learning model. It can be used to train a corresponding detailed machine learning model.

Specifically, as described above in (C) of FIG. 6, the machine learning model includes a detailed machine learning model corresponding to each of the preset age groups (20s to 60s in FIG. 7), and another embodiment of the present invention 7, the machine learning model is a detailed machine learning model for deriving exercise effect prediction information of a user with high blood pressure for each age group, and a detailed machine learning model for deriving exercise effect prediction information of a user with diabetes It can include machine learning models.

That is, the machine learning model includes first to tenth detailed machine learning models, and the first detailed machine learning model includes a plurality of extracted data for patients in their twenties and high blood pressure among a plurality of extracted data included in the training set. , and the second detailed machine learning model learns a plurality of extracted data for patients in their 20s and diabetes among the plurality of extracted data included in the training set, and the third detailed machine learning model is included in the training set. Among the plurality of extracted data, the plurality of extracted data is learned for patients in their 30s and patients with hypertension, and the fourth detailed machine learning model is applied to patients in their 30s and diabetes among the plurality of extracted data included in the training set. The fifth detailed machine learning model learns a plurality of extracted data for patients in their 40s and high blood pressure among the plurality of extracted data included in the training set, and the sixth detailed machine learning model learns a plurality of extracted data for patients in their 40s and diabetes among the plurality of extracted data included in the training set, and the 7th detailed machine learning model learns a plurality of extracted data for patients in their 40s and hypertension among the plurality of extracted data included in the training set learns a plurality of extracted data for a patient with The 9th detailed machine learning model learns multiple extracted data for patients in their 60s and high blood pressure among the multiple extracted data included in the training set, and the 10th detailed machine learning model learns multiple extracted data included in the training set. It is possible to learn multiple extracted data for patients in their 60s and diabetes.

As such, in the present invention, rather than learning all the extracted data from one machine learning model, the machine learning model includes detailed machine learning models for each predetermined age group and each disease for hypertension and diabetes, and each detailed machine The learning model learns only a plurality of extracted data corresponding to the corresponding age group and disease for hypertension or diabetes, so that each detailed machine learning model learns only the characteristics of the extracted data for the age group and disease assigned to it. Using a specific detailed machine learning model corresponding to the age and disease (hypertension or diabetes) of the user requesting the service, it is possible to more accurately derive the exercise effect prediction information for the user.

8 schematically shows a process of performing the exercise effect prediction step (S15) according to an embodiment of the present invention.

As shown in Fig. 8, the exercise effect prediction step (S15), when receiving service request information including body information about the user who wants to predict the exercise prescription effect, the corresponding information included in the body information Input the user's age information and the user's body information to the learned detailed machine learning model corresponding to the user's possession of hypertension or diabetes, and if the user has hypertension, The predicted systolic blood pressure and diastolic blood pressure of the user for each week are output as exercise effect prediction information, and when the user has diabetes, the predicted exercise effect is predicted by using the predicted value of the user's glucose level for each week or more preset after the current time. information can be output.

Specifically, the user may input service request information for a service for which the user wants to receive the exercise effect prediction information through the application installed in the user terminal 2000, and the exercise effect prediction step (S15) is the service. Request information may be received from the user terminal 2000 . Detailed information included in the service request information may be different depending on a disease possessed by the user, more specifically, whether the user is a hypertensive or diabetic patient.

As shown in FIG. 8A , when the user is a hypertension patient, the service request information may include age information, blood pressure information, BMI information, exercise prescription rate information, and parking information for the user. The blood pressure information includes a systolic blood pressure value and a diastolic blood pressure value of the current user, the BMI information includes a body mass index (BMI) of the current user, and the exercise prescription rate information includes the exercise prescription rate information of the current user. Including information on the exercise prescribed by the user, the parking information may include a performance history and results for the exercise prescribed by the user for each parking.

On the other hand, as shown in FIG. 8B, when the user is a diabetic patient, the service request information includes age information, glucose level information, BMI information, exercise prescription rate information, and parking information for the user. can do. Compared with the above-described service request information for a hypertensive patient, the service request information for a diabetic patient does not include blood pressure information, but includes glucose level information. The glucose level information includes the current user's glucose level.

Meanwhile, in the exercise effect prediction step (S15), it is possible to determine whether the user is a hypertension patient or a diabetic patient by identifying whether the service request information includes blood pressure information or glucose level information. In another embodiment of the present invention, the service request information may additionally include separate unique information for identifying whether the user is a hypertension patient or a diabetic patient, and the unique information is identified in the exercise effect prediction step (S15). It is also possible to determine whether the user is a hypertensive patient or a diabetic patient.

The exercise effect prediction step (S15) of receiving the service request information is a specific detailed machine learning included in the above-described machine learning model in consideration of the age information included in the service request information and the determination of the user as a high blood pressure or diabetes patient. The service request information can be entered into the model. For example, when the age information included in the service request information is 30 years old and blood pressure information is included in the service request information to determine the user as a hypertensive patient, the exercise effect prediction step (S15) is shown in FIG. 7 The service request information may be input to the third detailed machine learning model that has learned the extracted data of patients in their 30s and hypertension described above.

Next, the exercise effect prediction step (S15) is to input service request information to a specific detailed machine learning model, and exercise effect prediction information for predicting how the physical figure changes when the user performs the prescribed exercise. can be printed out. The exercise effect prediction information may include a predicted value for a physical value for each predetermined week in the future from the time the user requests the service, and the physical value may be different depending on whether the user is a hypertensive patient or a diabetic patient . Specifically, as shown in FIG. 8(A), when the user is a hypertensive patient, the exercise effect prediction information includes a systolic blood pressure predicted value and a diastolic blood pressure predicted value for each week, and as shown in FIG. 8(B), the user is diabetic. In the case of a patient, the exercise effect prediction report includes a glucose level prediction value for each week.

On the other hand, the total number of parkings included in the exercise effect prediction information may be variously set, but preferably, the exercise effect prediction information may include prediction values for physical figures for each 52 weeks from the current time point. In this way, the derived exercise effect prediction information is provided to the user terminal 2000 of the user who requested the service, and the corresponding user terminal 2000 that has received the exercise effect prediction information numerically through the installed application is Predicted values for numerical values can be displayed, or predicted values for physical figures for each week can be displayed through graphic elements such as graphs.

As described above, in the present invention, since a quantitative prediction value for a physical level (blood pressure or glucose level) predicted when a hypertension or diabetes patient performs a prescribed exercise is provided, it motivates the user to perform the exercise can be effective.

9 schematically shows a process of performing a machine learning model update step according to an embodiment of the present invention.

As shown in Figure 9, the exercise prescription effect prediction method, the machine learning model through the machine learning model learning step (S14), a machine learning model update step performed after the point in time the machine learning model learned through the training set; Further comprising, in the updating of the machine learning model, when the body database and the exercise prescription database are updated by collecting a predetermined number of additional patient information in the database building step (S10), the updated body database and the Additional learning can be performed on the learned machine learning model using the updated exercise prescription database.

As described above, in one embodiment of the present invention, exercise effect prediction information for a user using a machine learning model learned based on the body database and exercise prescription database constructed by performing the database construction step (S10) once However, in another embodiment of the present invention, the database construction step (S10) is additionally received in the body database and exercise prescription database whenever patient information is received from the preset period or additional user terminal 2000. Patient information may be updated, and the machine learning model learning step ( S14 ) may additionally perform additional learning on the machine learning model learned based on the updated body database and exercise prescription database.

Specifically, as shown in FIG. 9 , Crontab corresponds to a scheduler for periodically repeating a specific task according to a preset time or date. When the period corresponds to the preset period as described above, the initial data set construction step (S12) determines whether the additionally received patient information in the database construction step (S10) is equal to or greater than a preset number, and the additionally received patient information When the number is greater than or equal to the set number, an initial data set according to steps S61 and S62 is constructed based on the additionally received patient information, and a detailed method of constructing the initial data set is omitted as described above. For example, the preset number may correspond to 50, and in the initial dataset building step (S12), when the additionally received patient information is 50 or more, an initial dataset is created based on 50 or more patient information. can be built

Preferably, for each of the patient information additionally received in the database building step (S10), one or more patients who cannot exercise are determined in the above-described motion impossible determination step (S11), and the initial data set building step (S12) determines whether the remaining patient information is greater than or equal to a preset number, excluding the patient information about one or more patients determined to be unable to exercise through the movement impossible determination step (S11) among the additionally received patient information.

On the other hand, although not shown in FIG. 9 , when the initial dataset is constructed according to the additionally received patient information, the learning set creation step S13 is performed, and a plurality of extracted data included in the additionally constructed initial dataset are predetermined. A training set and a validation set can be generated by dividing by the ratio of .

In the machine learning model update step, an initial data set is additionally constructed according to the received patient information, and a training set and a verification set generated by dividing a plurality of extracted data included in the initial data set at a predetermined ratio are generated, Additional training can be performed on previously trained machine learning models.

Specifically, as shown in step S63, in the updating of the machine learning model, the most recent version of the learned machine learning model among the learned machine learning models of each version stored in the computing device 1000 is loaded. Subsequently, as in step S64, additional learning is performed using the additionally generated training set for the latest version of the learned machine learning model, and the corresponding machine learning model is used using the additionally generated validation set in the additionally trained machine learning model. performance results can be derived. In this way, the additionally trained machine learning model and the performance results for the additionally learned machine learning model may be stored in the computing device 1000 as in step S65.

As such, in the present invention, it is possible to continuously improve the prediction performance by continuously updating the machine learning model in consideration of additional patient information that is periodically collected, rather than learning only the initially constructed data for the machine learning model. Each version of the machine learning model is stored individually, so that the previous version of the machine learning model can be used in case the performance of the machine learning model deteriorates due to additional learning due to overfitting.

10 schematically shows the overall flow of a method for predicting an exercise prescription effect for a patient with high blood pressure and diabetes based on artificial intelligence according to an embodiment of the present invention.

As shown in FIG. 10 , 'patient information CMS' corresponds to the patient information transmitted from the user terminal 2000 described above, and the patient information includes body information (patient user data in FIG. 10) and exercise prescription information (FIG. It includes exercise prescription data of 10) and exercise effect information (exercise effect data of FIG. 10).

The 'dataset collection RESTful API' shown in FIG. 10 is based on a plurality of patient information collected by performing the above-described database construction step (S10), the above-described body database (patient diagnosis information database in FIG. 10) and exercise prescription database. (Exercise prescription result database of Fig. 10) is built. The body database and exercise prescription database constructed in this way may be stored in the computing device 1000 .

In the 'exercise prescription effect verification artificial intelligence model' shown in FIG. 10, the 'data set creation' may correspond to the above-described exercise impossibility determination step (S11), and the exercise impossibility determination step (S11) is performed in the built-up body database. As a pre-processing process, one or more patients who are unable to exercise can be judged.

The process from 'extracting main data' to 'creating and classifying a data set file' shown in FIG. 10 may correspond to the above-described initial dataset construction step S12 and learning set creation step S13. That is, in the initial data set construction step (S12), the initial data including the extracted data for each patient based on the body database and the exercise prescription database for the remaining patients except for one or more patients determined in the exercise impossibility determination step (S11). create a set In the training set creation step S13, a training set (Training Set in FIG. 10) and a verification set (Test Set in FIG. 10) are generated by dividing the plurality of extracted data included in the generated initial dataset by a predetermined ratio.

The 'artificial intelligence model creation' shown in FIG. 10 may correspond to the above-described machine learning model learning step S14. That is, the machine learning model learning step (S14) trains the machine learning model through the generated training set, and inputs the generated verification set for the learned machine learning model to obtain the performance result (exercise prescription verification result in FIG. 10). print out The performance result output as described above may be stored in the computing device 1000 . In addition, the training set and verification set for learning and verifying the machine learning model, and the learned machine learning model may also be stored in the computing device 1000 as shown in 'Artificial intelligence model management' of FIG. 10 .

Finally, 'exercise prescription effect verification data management' shown in FIG. 10 may correspond to the above-described machine learning model update step. That is, when a predetermined number of additional patient information (New data in FIG. 10) is received from the plurality of user terminals 2000, the machine learning model updating step is performed based on the additionally received predetermined number of patient information. Perform additional training on the machine learning model. As such, the additionally learned machine learning model and the performance results derived according to the verification set generated based on the additionally received patient information in the additionally learned machine learning model may also be individually stored in the computing device 1000 .

As described above, the present invention continuously learns a machine learning model based on a plurality of patient information received from the user terminal 2000, and uses the learned machine learning model according to the user's request to exercise effect according to hypertension or diabetes. Prediction information may be derived and provided to a corresponding user.

11 schematically illustrates an internal configuration of a computing device according to an embodiment of the present invention.

The above-described computing device 1000 illustrated in FIG. 1 may include components of the computing device 11000 illustrated in FIG. 11 .

11, the computing device 11000 includes at least one processor 11100, a memory 11200, a peripheral interface 11300, an input/output subsystem ( It may include at least an I/O subsystem) 11400 , a power circuit 11500 , and a communication circuit 11600 . In this case, the computing device 11000 may correspond to the computing device 1000 illustrated in FIG. 1 .

The memory 11200 may include, for example, high-speed random access memory, a magnetic disk, SRAM, DRAM, ROM, flash memory, or non-volatile memory. . The memory 11200 may include a software module, an instruction set, or other various data required for the operation of the computing device 11000 .

In this case, access to the memory 11200 from other components such as the processor 11100 or the peripheral device interface 11300 may be controlled by the processor 11100 .

Peripheral interface 11300 may couple input and/or output peripherals of computing device 11000 to processor 11100 and memory 11200 . The processor 11100 may execute a software module or an instruction set stored in the memory 11200 to perform various functions for the computing device 11000 and process data.

The input/output subsystem may couple various input/output peripherals to the peripheral interface 11300 . For example, the input/output subsystem may include a controller for coupling a peripheral device such as a monitor or keyboard, mouse, printer, or a touch screen or sensor as required to the peripheral interface 11300 . According to another aspect, input/output peripherals may be coupled to peripheral interface 11300 without going through an input/output subsystem.

The power circuit 11500 may supply power to all or some of the components of the terminal. For example, the power circuit 11500 may include a power management system, one or more power sources such as batteries or alternating current (AC), a charging system, a power failure detection circuit, a power converter or inverter, a power status indicator, or a power source. It may include any other components for creation, management, and distribution.

The communication circuit 11600 may enable communication with another computing device using at least one external port.

Alternatively, as described above, if necessary, the communication circuit 11600 may transmit and receive an RF signal, also known as an electromagnetic signal, including an RF circuit, thereby enabling communication with other computing devices.

This embodiment of FIG. 11 is only an example of the computing device 11000, and the computing device 11000 may omit some components shown in FIG. 11 or further include additional components not shown in FIG. 11, or 2 It may have a configuration or arrangement that combines two or more components. For example, a computing device for a communication terminal in a mobile environment may further include a touch screen or a sensor in addition to the components shown in FIG. 11 , and may include various communication methods (WiFi, 3G, LTE) in the communication circuit 11600 . , Bluetooth, NFC, Zigbee, etc.) may include a circuit for RF communication. Components that may be included in the computing device 11000 may be implemented in hardware, software, or a combination of both hardware and software including an integrated circuit specialized for one or more signal processing or applications.

Methods according to an embodiment of the present invention may be implemented in the form of program instructions that can be executed through various computing devices and recorded in a computer-readable medium. In particular, the program according to the present embodiment may be configured as a PC-based program or an application dedicated to a mobile terminal. An application to which the present invention is applied may be installed in the computing device 11000 through a file provided by the file distribution system. For example, the file distribution system may include a file transmission unit (not shown) that transmits the file according to the request of the computing device 11000 .

The device described above may be implemented as a hardware component, a software component, and/or a combination of the hardware component and the software component. For example, devices and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA). , a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions, may be implemented using one or more general purpose or special purpose computers. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that may include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

Software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or device, to be interpreted by or to provide instructions or data to the processing device. , or may be permanently or temporarily embody in a transmitted signal wave. The software may be distributed over networked computing devices, and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

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 in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

According to an embodiment of the present invention, the machine learning model is trained according to the collected patient information, and the user's body information is input to the learned machine learning model, thereby predicting the exercise effect when the user performs weekly exercise. By outputting the information, it is possible to exert an effect of providing motivation for exercise performance to a user with high blood pressure or diabetes.

According to an embodiment of the present invention, in the exercise impossibility determination step, exercise using determination values for blood pressure, heart rate, glucose, cholesterol, and glycated hemoglobin in a database constructed using a plurality of collected patient information. By determining this impossible patient, the machine learning model can learn information about the remaining patients except for the patient unable to exercise from the database, thereby exhibiting the effect of more accurately predicting the exercise effect.

According to an embodiment of the present invention, the machine learning model includes a plurality of detailed machine learning models according to age groups and hypertension or diabetes, and each detailed machine learning model includes a training set for a corresponding age group and hypertension or diabetes. Because it is learned through the user's age and disease, it is possible to exert the effect of more accurately predicting the exercise effect through a detailed machine learning model corresponding to the user's age and disease.

According to an embodiment of the present invention, in the machine learning model updating step, when patient information corresponding to a predetermined number is additionally collected in the database building step, the machine learning model learned based on the additionally collected patient information is additionally learned , so that the performance of the machine learning model can be improved, and thus, it is possible to exert the effect of more accurately predicting the exercise effect for the user.

As described above, although the embodiments have been described with reference to the limited embodiments and drawings, various modifications and variations are possible by those skilled in the art from the above description. For example, the described techniques are performed in a different order than the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (9)

An artificial intelligence-based exercise prescription effect prediction method for hypertensive and diabetic patients performed in a computing device including one or more processors and one or more memories,
Collect patient information including physical information, exercise prescription information, and exercise effect information for a plurality of hypertensive patients and diabetic patients, and based on the collected body information, exercise prescription information, and exercise effect information included in the collected patient information database construction step of constructing a physical database for the patient's body and disease and an exercise prescription database for the exercise and exercise results performed by the patient;
an exercise impossibility determination step of determining one or more patients unable to exercise based on the information on each patient's body and disease included in the body database;
Age information on each of the remaining patients except for one or more patients determined to be unable to exercise in the body database, blood pressure information of the patient if the patient is a hypertensive patient, and information on the patient's blood pressure if the patient is a diabetic patient Initial data including extraction data for each patient by extracting blood glucose information and extracting exercise prescription rate information and exercise performance history information for each of the remaining patients except for one or more patients determined to be unable to exercise from the exercise prescription database Initial data set construction step of building a set;
A training set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of a plurality of extracted data included in the initial dataset by a predetermined first ratio value is generated, and a plurality of extractions included in the initial dataset are generated. a learning set generating step of generating a verification set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of data by a predetermined second ratio value;
A supervised learning-based machine learning model is trained using the training set, the verification set is input to the machine learning model that has learned the training set, and the output result is used as performance information of the machine learning model that learned the training set. Thus, a machine learning model learning step of storing the performance information; and
When receiving service request information including body information about the user who wants to predict the exercise prescription effect, the body information is input to the learned machine learning model and the user's information for each predetermined one or more weeks after the current time is received. Exercise effect prediction step of outputting exercise effect prediction information; Artificial intelligence-based exercise prescription effect prediction method for hypertension and diabetes patients, including.
The method according to claim 1,
The computing device for performing the exercise prescription effect prediction method may communicate with a plurality of user terminals,
The database building step is
An artificial intelligence-based exercise prescription effect prediction method for hypertension and diabetes patients, which collects the patient information by receiving body information, exercise prescription information, and exercise effect information input by each user through the plurality of user terminals.
The method according to claim 1,
The exercise impossibility determination step is,
With respect to the systolic blood pressure value and the diastolic blood pressure value included in the body and disease information for each patient included in the body database, the systolic blood pressure value is equal to or greater than the first determination value, or the diastolic blood pressure value is equal to or greater than the second determination value a blood pressure determination step of determining the patient as a patient unable to exercise; and
With respect to the heartbeat value included in the information on the body and disease for each patient that is not determined to be unable to exercise in the blood pressure determination step, the heartbeat value is greater than or equal to the third determination value, or the heartbeat value is the fourth determination A heartbeat determination step of determining the patient as a patient unable to exercise when the value is less than or equal to the value; Artificial intelligence-based exercise prescription effect prediction method for hypertension and diabetes patients, including.
4. The method of claim 3,
The exercise impossibility determination step is,
With respect to the glucose level included in the information on the body and disease for each patient who is not determined to be unable to exercise in the heartbeat determination step, the glucose level is greater than or equal to the fifth determination value, or the glucose level is less than or equal to the sixth determination value a glucose level determination step of determining the patient as a patient unable to exercise;
With respect to the cholesterol level included in the information on the body and disease of each patient who is not determined to be unable to exercise in the glucose level determination step, if the cholesterol level is greater than or equal to the 7th determination value, the patient is classified as a patient unable to exercise. Cholesterol level determination step to determine;
With respect to the low-density cholesterol level included in the information on the body and disease of each patient who is not determined to be unable to exercise in the cholesterol level determination step, if the low-density cholesterol level is equal to or greater than the 8th determination value, the patient cannot exercise Low-density cholesterol level determination step to determine the patient;
With respect to the high-density cholesterol level included in the body and disease information for each patient who is not determined to be unable to exercise in the low-density cholesterol level determination step, if the high-density cholesterol level is below the ninth determination value, the patient cannot exercise High-density cholesterol level determination step to determine a patient; and
With respect to the glycated hemoglobin level included in the body and disease information for each patient who is not determined to be unable to exercise in the high-density cholesterol level determination step, if the glycated hemoglobin level is higher than the 10th determination value, the patient cannot exercise A glycated hemoglobin level determination step of judging a patient as a patient; Artificial intelligence-based exercise prescription effect prediction method for hypertension and diabetes patients, further comprising.
The method according to claim 1,
The supervised learning-based machine learning model includes a first detailed machine learning model to a tenth detailed machine learning model,
The machine learning model learning step is,
A plurality of extracted data included in the training set is divided into a plurality of preset age groups and for each hypertension or diabetes patient, and each of the plurality of extracted data divided into a plurality of numbers is the first detailed machine learning model to the tenth detail. An artificial intelligence-based exercise prescription effect prediction method for high blood pressure and diabetes patients, which is used to train a detailed machine learning model corresponding to any one of the machine learning models.
6. The method of claim 5,
The exercise effect prediction step is,
When receiving service request information including body information about the user who wants to predict the exercise prescription effect, the age information of the user included in the body information and learning corresponding to the user's possession of high blood pressure or diabetic disease The user's body information is input into the detailed machine learning model, and when the user has hypertension, the predicted systolic blood pressure and the diastolic blood pressure of the user for each of 1 or more preset weeks after the current time are used as exercise effect prediction information. Prediction of exercise prescription effect for hypertensive and diabetic patients based on artificial intelligence, which outputs the predicted value of the user's glucose level for each week of one or more preset weeks after the current point in time as exercise effect prediction information when the user has diabetes Way.
The method according to claim 1,
The exercise prescription effect prediction method,
Further comprising; a machine learning model updating step performed after the point in time when the machine learning model learns through the training set through the machine learning model learning step;
The machine learning model update step is,
In the case of updating the body database and the exercise prescription database by collecting a predetermined number of additional patient information in the database building step, the learned machine learning model using the updated body database and the updated exercise prescription database An artificial intelligence-based method of predicting the effect of exercise prescription in hypertensive and diabetic patients.
A computing device comprising one or more processors and one or more memories to perform an artificial intelligence-based exercise prescription effect prediction method for patients with hypertension and diabetes,
Collect patient information including physical information, exercise prescription information, and exercise effect information for a plurality of hypertensive patients and diabetic patients, and based on the collected body information, exercise prescription information, and exercise effect information included in the collected patient information a database construction unit for constructing a body database for the patient's body and disease and an exercise prescription database for the exercise and exercise results performed by the patient;
an exercise impossibility determination unit for determining at least one patient unable to exercise based on the information on each patient's body and disease included in the body database;
Age information on each of the remaining patients except for one or more patients determined to be unable to exercise in the body database, blood pressure information of the patient if the patient is a hypertensive patient, and information on the patient's blood pressure if the patient is a diabetic patient Initial data including extraction data for each patient by extracting blood glucose information and extracting exercise prescription rate information and exercise performance history information for each of the remaining patients except for one or more patients determined to be unable to exercise from the exercise prescription database Initial data set construction unit to build a set;
A training set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of a plurality of extracted data included in the initial dataset by a predetermined first ratio value is generated, and a plurality of extractions included in the initial dataset are generated. a learning set generating unit for generating a verification set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of data by a predetermined second ratio value;
A supervised learning-based machine learning model is trained using the training set, the verification set is input to the machine learning model that has learned the training set, and the output result is used as performance information of the machine learning model that learned the training set. Thus, a machine learning model learning unit for storing the performance information; and
When receiving service request information including body information about the user who wants to predict the exercise prescription effect, the body information is input to the learned machine learning model and the user's information for each predetermined one or more weeks after the current time is received. Computing device comprising a; exercise effect prediction unit for outputting the exercise effect prediction information.
By executing a computer program in a computing device including one or more processors and one or more memories, as a computer-readable medium for implementing an artificial intelligence-based exercise prescription effect prediction method for hypertension and diabetes patients performed by the computing device. , the computer-readable medium stores instructions for causing the computing device to perform the following steps,
The following steps are
Collect patient information including physical information, exercise prescription information, and exercise effect information for a plurality of hypertensive patients and diabetic patients, and based on the collected body information, exercise prescription information, and exercise effect information included in the collected patient information database construction step of constructing a physical database for the patient's body and disease and an exercise prescription database for the exercise and exercise results performed by the patient;
an exercise impossibility determination step of determining one or more patients unable to exercise based on the information on each patient's body and disease included in the body database;
Age information on each of the remaining patients except for one or more patients determined to be unable to exercise in the body database, blood pressure information of the patient if the patient is a hypertensive patient, and information on the patient's blood pressure if the patient is a diabetic patient Initial data including extraction data for each patient by extracting blood glucose information and extracting exercise prescription rate information and exercise performance history information for each of the remaining patients except for one or more patients determined to be unable to exercise from the exercise prescription database Initial data set construction step of building a set;
A training set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of a plurality of extracted data included in the initial dataset by a predetermined first ratio value is generated, and a plurality of extractions included in the initial dataset are generated. a learning set generating step of generating a verification set including a plurality of extracted data corresponding to a size obtained by multiplying the total number of data by a predetermined second ratio value;
A supervised learning-based machine learning model is trained using the training set, the verification set is input to the machine learning model that has learned the training set, and the output result is used as performance information of the machine learning model that learned the training set. Thus, a machine learning model learning step of storing the performance information; and
When receiving service request information including body information about the user who wants to predict the exercise prescription effect, the body information is input to the learned machine learning model and the user's information for each predetermined one or more weeks after the current time is received. A computer-readable medium comprising a; exercise effect prediction step of outputting the exercise effect prediction information.

Images