JP6954568B2 - Health management support system, health management support method, and program - Google Patents

Description

The present invention is, health management support system, Kang management supporting method Takeshi, and a program.

In oriental medicine, "coldness" is considered to be the cause of all illnesses. If the cold is left as it is, blood circulation will continue to be poor, which may cause various illnesses such as deterioration of physical function and visceral diseases. In addition, in a constitution where peripheral blood flow is good and it is difficult to feel cold, even if you are not aware of poor circulation, cooling is accelerated and body temperature is lowered, which causes a decrease in immunity and causes illness. There is also.
In addition, oxygen deficiency due to decreased blood flow, associated with coldness, is also a risk factor for various diseases. Therefore, coldness and hypoxia are also considered to be the two major causes of illness. Not only illness, stiff shoulders, cold legs, cold hands and feet, numbness, dullness, sleeplessness, headache, fatigue, dizziness, menstrual pain, stomachache, blurred vision, easy to catch cold, rough skin, palpitation / shortness of breath, depression and autonomy Colds and diminished blood flow can be the root causes of a variety of disorders, such as palpitations. In the examination by oriental medicine, even if the symptoms are the same, the diagnosis is made by classifying (stratifying) according to whether the limbs and other parts of the body are cold or the central part of the body is cold. It is known that different treatments are performed depending on the condition.
On the other hand, in health examinations at medical institutions, biological information such as blood tests is regularly measured. Patent Document 1 discloses a technique for supporting health management based on health diagnosis data and receipt data. It is desired that presentations are made to maintain and manage health, effectively promote health, and prevent diseases based on biological information.

Japanese Unexamined Patent Publication No. 2014-182472

However, although many people undergo a health examination every year, most of them are aware of whether or not they have poor circulation, and there are few people who receive an objective examination by a specialist such as a Chinese medicine doctor. For this reason, even if one is aware of poor circulation, it is not possible to know what type of poor circulation it is, and there is a possibility that the wrong treatment may be selected.

The present invention has been made in view of such circumstances, the health management support system capable of classifying the poor circulation using biometric information, to provide a health management support method, and a program.

In order to solve the above-mentioned problems of the present invention, the present invention includes a first acquisition unit that acquires first information indicating whether or not the user has poor circulation, and measurement items as measurement results of the user's biological information. Based on the second acquisition unit that acquires the second information including the numerical value , the first information, and the characteristics extracted from the second information, the sensitivity to cold sensitivity is determined by the numerical value for the measurement item of the biological information. The first type that corresponds to the symptoms of coldness, the second type that has poor circulation and the numerical value for the measurement item of biometric information does not correspond to the symptoms of coldness, and the numerical value for the measurement item of biometric information is non-cold. Selection to select an index marker that is biometric information that serves as an index to classify into the third type that corresponds to the symptom and the fourth type that has non-cold sensitivity and the numerical value for the measurement item of biometric information does not correspond to the non-cold symptom. Based on the unit, information indicating whether or not the target user has poor circulation, and the numerical value of the index marker in the target user, the constitution of the target user is determined by the first type, the second type, and the third type. and a classification unit for classifying the one type of the fourth type, and an output unit for outputting a classified classification result by the classifying unit, the selection unit, poor circulation and two to the second information Correlation information that correlates with one cold symptom is extracted, and among the above correlation information, the correlation information and the one symptom are positive for one symptom and the other symptom of the two cold symptom. It is a health management support system that uses the correlation information showing a correlation and showing a negative correlation between the correlation information and the other symptom as the index marker.

In the present invention, the first acquisition unit acquires the first information indicating whether or not the user has poor circulation, and the second acquisition unit includes measurement items and numerical values as measurement results of the user's biological information. 2 Information is acquired, and the selection unit corresponds to poor circulation based on the first information and the characteristics extracted from the second information, and the numerical value for the measurement item of the biological information corresponds to the cold sensitivity. The first type, the second type, which has poor circulation and the numerical value for the measurement item of biological information does not correspond to the symptom of coldness, and the third type, which has poor circulation and the numerical value for the measurement item of biological information corresponds to the symptom of non-coldness. Select the index marker, which is the biometric information to be classified into the type, the fourth type, which is non-cold and the numerical value for the measurement item of the biometric information does not correspond to the non-cold symptom, and the classification unit is the target user. Based on the information indicating whether or not the subject has poor circulation and the numerical value of the index marker in the target user, the constitution of the target user is determined by the first type, the second type, the third type, and the fourth type. It is classified into one of the types, the output unit outputs the classification result classified by the classification unit , and the selection unit correlates with the poor circulation and the two cold symptoms from the second information. Information is extracted, and among the correlation information, the correlation information and the one symptom show a positive correlation with respect to one symptom and the other symptom of the two cold symptom, and the correlation information and the other symptom are positively correlated. It is a health management support method using the correlation information showing a negative correlation with the symptom of the above as the index marker.

The present invention is a first acquisition means for acquiring a first information indicating whether or not a user has poor circulation, and a second information including a measurement item and a numerical value as a measurement result of the user's biological information. Based on the second acquisition means, the first information, and the characteristics extracted from the second information, the first type of poor circulation is defined as the first type of poor circulation and the numerical value for the measurement item of the biological information corresponds to the poor circulation. The second type, which has poor circulation and the numerical value for the measurement item of biometric information does not correspond to the symptom of coldness, and the third type, which has non-cold sensitivity and the numerical value for the measurement item of biometric information corresponds to the symptom of non-coldness, and non-cold sensitivity. A selection means for selecting an index marker which is biometric information as an index for classifying into a fourth type in which the numerical value for the measurement item of biometric information does not correspond to the symptom of poor circulation.
Based on the information indicating whether or not the target user has poor circulation and the numerical value of the index marker in the target user, the constitution of the target user is determined by the first type, the second type, the third type, and the above. It is a program for executing a classification means for classifying into any of the fourth types and an output means for outputting the classification result classified by the classification means. And the correlation information that correlates with the two cold symptoms is extracted, and among the correlation information, the correlation information and the one symptom are obtained for one symptom and the other symptom of the two cold symptoms. This is a program in which the correlation information showing a positive correlation and showing a negative correlation between the correlation information and the other symptom is used as the index marker.

According to the present invention, poor circulation can be classified using biological information. Therefore, it is possible to stratify the type of "coldness" and understand the cause of "coldness" and the potential risk of illness or disorder. Therefore, even a user who is not aware of poor circulation can make the state of "coldness" manifest and grasp the potential risk. Furthermore, it is possible to make presentations aimed at effectively improving QOL (Quality Of Life) corresponding to the "cold" type, promoting health, and preventing risk-related diseases.

The outline of the health management support system 1 which concerns on embodiment is shown. It is a block diagram which shows the example of the system configuration of the health management support system 1 which concerns on embodiment. It is a sequence diagram which shows the flow of the process performed by the health management support system 1 which concerns on embodiment. It is a figure which shows the example of the data structure of the cold sensitivity information 320 which concerns on embodiment. It is a figure which shows the example of the data structure of the medical examination information 321 which concerns on embodiment. It is a figure explaining the process performed by the health management support system 1 which concerns on embodiment. It is a figure explaining the process performed by the health management support system 1 which concerns on embodiment. It is a figure which shows the example of the structure of the computer system 100 which concerns on embodiment. It is a block diagram which shows the example of the structure of the cold sensitivity learning server 20 which concerns on embodiment. It is a block diagram which shows the example of the structure of the constitution classification server 30 which concerns on embodiment.

Hereinafter, embodiments of the invention will be described with reference to the drawings.

[Outline of health management support system 1]
The outline of the health management support system 1 in the present invention will be described. FIG. 1 is a diagram showing an outline of the health management support system 1 according to the present embodiment. The health management support system 1 is a system that predicts the constitution of the target user TU in consideration of the cold constitution and presents the health management according to the predicted constitution. For the prediction, the medical examination information 321 of the target user TU (see FIG. 5) is used. The medical examination information 321 here is biological information measured by a medical examination or the like. The biological information includes non-invasively measured biological information such as body weight, height, and blood pressure, and invasively measured biological information such as a blood test. In addition, the biological information may include the result of a medical inquiry by a doctor, the result of a pre-health questionnaire conducted in advance, and the like.

In the health management support system 1, a cold sensitivity prediction model 323 is created as a preliminary preparation. The cold sensitivity prediction model 323 is a model that predicts whether or not the target user TU has poor circulation based on the medical examination information 321 of the target user TU. The method of creating the poor circulation prediction model 323 will be described in detail later, but the cold sensitivity prediction model 323 is created by the cold sensitivity learning server 20. The cold disease learning server 20 creates a cold disease prediction model 323 by machine learning a learning data set in which medical examination information 321 by an unspecified user U and cold disease information 320 (see FIG. 4) are associated with each other. The sensitivity to cold information 320 here is information indicating whether or not the patient has poor circulation. The sensitivity to cold information 320 is obtained, for example, by conducting a questionnaire on the symptoms of coldness to the user U who has undergone a medical examination, or by hearing whether or not the user U has been diagnosed by a specialist in the past.

Further, in the health management support system 1, the index marker 322 is selected by the index marker selection unit 332 as a preliminary preparation. The index marker 322 is an index for classifying (stratifying) poor circulation, which is selected from the medical examination information 321. The index marker selection unit 332 selects the index marker 322 based on the cold sensitivity information 320 and the medical examination information 321 of the unspecified user U.

In the health management support system 1, it is predicted whether or not the target user TU has poor circulation by using the cold sensitivity prediction model 323. The cold sensitivity prediction model 323 outputs a cold sensitivity prediction (information for predicting whether or not the patient has poor circulation) based on the medical examination information 321 of the target user TU.

In the health management support system 1, the cold constitution classification unit 334 classifies the constitution of the target user TU into any type in which cold sensitivity is classified by an index marker. The cold constitution classification unit 334 classifies the constitution of the target user TU based on the value of the index marker in the medical examination information of the target user TU and the cold sensitivity prediction predicted by the cold sensitivity prediction model 323, and classifies the classification result. Output as a constitution prediction.

In the health management support system 1, based on the constitution prediction of the target user TU, preventive methods and countermeasures for susceptible diseases and the like, an improvement system for further improving the constitution, and the like are presented. As a result, not only disease prevention but also QOL can be improved and an independent life can be supported. The above is an explanation of the outline of the health management support system 1.

[Configuration of health management support system 1]
Next, the configuration of the health management support system 1 will be described. FIG. 2 is a block diagram showing an example of the configuration of the health management support system 1. The health management support system 1 includes, for example, a biological information storage server 10, a cold disease learning server 20, a constitution classification server 30, an information presentation server 40, and a plurality of user terminals 50 (user terminals 50-1, 50-2, ... 50-N). N is an arbitrary natural number.

The components of the health management support system 1 (biological information storage server 10, cold disease learning server 20, constitution classification server 30, information presentation server 40, and user terminal 50) are connected to a wireless or wired communication network NW and communicate with each other. It can be performed. In addition, each of the components of the health management support system 1 may be plural. For example, the health management support system 1 may include a plurality of biological information storage servers 10 and the like. In addition, some of the components of the health management support system 1 may be realized by one device. For example, the constitution classification server 30 and the information presentation server 40 may be provided in one server device.

The biological information storage server 10 is an electronic device including a computer system 100 (see FIG. 8). For example, the biometric information storage server 10 may be a cloud device, a server device, a personal computer, or the like. The biological information storage server 10 stores (stores) the cold sensitivity information 320 and the medical examination information 321 of the unspecified user U. The biological information storage server 10 acquires the medical examination result of the user U who has undergone the medical examination from, for example, a data server provided in a medical institution entrusted with the medical examination, and uses the acquired information as the medical examination information 321. Remember. In addition, the biometric information storage server 10 stores the user U's sensitivity to cold information 320 acquired by conducting a questionnaire or hearing with the user U together with or separately from the health examination.

The sensitivity to cold learning server 20 is an electronic device including a computer system 100. For example, the sensitivity to cold learning server 20 is a cloud device, a server device, a personal computer, or the like. The cold sensitivity learning server 20 creates a cold sensitivity prediction model 323 that predicts the cold sensitivity of the target user TU from the medical examination information of the target user TU based on the medical examination information 321 of the unspecified user U and the cold sensitivity information 320.

The constitution classification server 30 is an electronic device including a computer system 100. For example, the constitution classification server 30 is a cloud device, a server device, a personal computer, or the like. The constitution classification server 30 classifies which type of constitution of the target user TU is based on poor circulation.

The information presentation server 40 is an electronic device including a computer system 100. For example, the information presentation server 40 is a cloud device, a server device, a personal computer, or the like. The information presentation server 40 presents information on health management according to the constitution of the target user TU.

The user terminal 50 is an electronic device including a computer system 100. For example, the user terminal 50 may be a smartphone, a personal computer, a mobile phone, a tablet terminal, a wearable terminal, or the like. The user terminal 50 is, for example, a terminal device such as a smartphone of the target user TU. Application software (hereinafter referred to as a health management application) that provides a "health management service" is installed on the user terminal 50. In the "health management service", by inputting the result of one's own health examination, it is predicted which type of one's constitution is based on poor circulation. In addition, in the "health management service", disease prevention methods, countermeasures, improvement systems, etc. according to the predicted constitution are presented. The user terminal 50 may use the "health management service" by accessing the URL (Uniform Resource Locator) indicating the site of the "health management service" provided by the Web service.
The above is the description of the configuration of the health management support system 1.

[Operation of health management support system 1]
Next, the operation of the health management support system 1 will be described. FIG. 3 is a sequence diagram showing a flow of processing performed by the health management support system 1.

(Step S10)
The biological information storage server 10 stores (stores) the cold sensitivity information 320 and the medical examination information 321 of the unspecified user U.

(Step S11)
The biological information storage server 10 outputs the cold sensitivity information 320 and the medical examination information 321 of the unspecified user U to the cold sensitivity learning server 20.
(Step S12)
The poor circulation learning server 20 uses the cold sensitivity information 320 of the unspecified user U and the medical examination information 321 to cause the learning model to learn the correspondence between the diagnosis result of the health examination and the cold sensitivity, thereby providing the cold sensitivity prediction model 323. create. The method of creating the cold sensitivity prediction model 323 will be described in detail later.
(Step S13)
The poor circulation learning server 20 outputs the created cold sensitivity prediction model 323 to the constitution classification server 30.

(Step S14)
The constitution classification server 30 acquires the cold sensitivity prediction model 323 and stores the acquired cold sensitivity prediction model 323.

(Step S15)
On the other hand, the biological information storage server 10 outputs the cold sensitivity information 320 of the unspecified user U and the medical examination information 321 to the constitution classification server 30.
(Step S16)
The constitution classification server 30 selects the index marker 322 based on the characteristics extracted from the poor circulation information 320 and the medical examination information 321. The method of selecting the index marker 322 will be described in detail later.

(Step S17)
On the other hand, on the user terminal 50, the "health management application" is started by a click operation of the target user TU. When the health management application is started, for example, an input form for inputting the diagnosis result of the health examination is displayed.
(Step S18)
When the diagnosis result of the health examination is input via the input form by the input operation of the target user TU or the like, the user terminal 50 transmits the health examination information 321 to the constitution classification server 30.
(Step S19)
The constitution classification server 30 predicts whether or not the target user TU has poor circulation by inputting the medical examination information 321 of the target user TU received from the user terminal 50 into the cold sensitivity prediction model 323.
(Step S20)
The constitution classification server 30 predicts whether or not the target user TU has poor circulation, and the constitution of the target user TU has poor circulation based on the value of the index marker extracted from the medical examination information 321 of the target user TU. Classify which type is based on.

(Step S21)
The constitution classification server 30 transmits the classification result of classifying the constitution of the target user TU to the information presentation server 40 and the user terminal 50 as constitution information.
(Step S22)
The information presentation server 40 selects information on disease prevention and constitution improvement proposals to be presented to the target user TU based on the result of classifying the constitution of the target user TU.
(Step S23)
The information presentation server 40 transmits the selected information as presentation information to the user terminal 50.
(Step S24)
The user terminal 50 displays the constitution information received from the constitution classification server 30 and the presentation information received from the information presentation server 40 in the output column of the "health application".
The above is the explanation of the operation of the health management support system 1.

[Data structure]
Next, the structure of various data used for the processing of the health management support system 1 will be described.

First, the sensitivity to cold sensitivity 320 will be described. The sensitivity to cold information 320 is data describing information regarding whether or not the user has poor circulation. FIG. 4 is a diagram showing an example of the data structure of the sensitivity to cold disease information 320. The cold sensitivity information 320 is shown as data in which information indicating a cold sensitivity flag, grounds, and the like is associated with a user ID (Identifier). The sensitivity to cold information 320 is created for each user U.

Information for uniquely identifying the user U is described in the user ID. Information indicating whether or not the user U has poor circulation is described in the poor circulation flag. The rationale shows the rationale for determining whether or not the patient has poor circulation. The rationale is shown in items such as self-diagnosis, judgment by questionnaire, and judgment by a specialist. For example, if whether or not the patient has poor circulation is based on the self-diagnosis of the user U, information indicating the self-diagnosis is described in the basis. If whether or not the patient has poor circulation is based on the results of a questionnaire conducted to the user U, information indicating the judgment by the questionnaire is described in the basis. If whether or not the person has poor circulation is diagnosed by a specialist, the basis includes information indicating the diagnosis by the specialist.

The sensitivity to cold flag may be binary information indicating whether or not the patient has poor circulation, and there are a plurality of levels such as poor circulation, slightly cold sensitivity, non-cold sensitivity, and neither. It may be information indicating the degree of poor circulation. In addition, the sensitivity to cold information 320 may include information indicating user attributes such as the name, gender, age, occupation, and family structure of the user U. The user attribute here may include arbitrary information of the user U, but it is preferable that the user attribute includes information that may be related to poor circulation, such as gender, age, and work environment. In addition, whether or not you have poor circulation can change depending on your age and season. Information or the like indicating whether there is a season to be performed or a season to be unaware may be shown.

Next, the medical examination information 321 will be described. The health examination information 321 is data describing the result of the health examination of the user U. FIG. 5 is a diagram showing an example of the data structure of the medical examination information 321. The health examination information 321 is described as data in which the ID of the user U is associated with the biometric information measured as a result of the health examination. The medical examination information 321 is created for each user U. Information for uniquely identifying the user U is described in the user ID. The medical examination items describe the diagnostic items performed in the medical examination. Health check items include, for example, blood tests, blood pressure, allergies, insomnia or depression. Blood tests include items such as bilirubin and AST (GOT), as well as ALT, γGT, total protein, creatinine, uric acid, cholesterol, TG, serum iron, blood glucose, HDL, ALB, white blood cells, and red blood cells, which are omitted in the figure. , Hematocrit, hemoglobin, neutrophils, platelets, ferritin, LDL, HbA1c, C peptide, C3, C4, glycoalbumin and the like. The blood pressure measured by the user U is described in the blood pressure. Information on allergies of user U is described in allergies. The information on allergies may be self-reported information from the user U, or may be information based on the results of a medical examination, blood test, or the like. Insomnia or depression describes information about user U's insomnia or depression. The information regarding insomnia and depression may be information self-reported by the user U, or information based on the results of interviews and hospital visits.
The above is the description of the data structure.

[Method of creating a cold sensitivity prediction model 323, method of selecting an index marker 322]
Next, a method of creating a cold sensitivity prediction model 323 and a method of selecting the index marker 322 will be described. FIG. 6 is a diagram illustrating a process performed by the health management support system 1 according to the embodiment. Hereinafter, a method of creating a poor circulation prediction model 323 and a method of selecting an index marker 322 will be described in the order of steps S100 to S104 shown in FIG.

(Step S100)
The health management support system 1 analyzes the factors of "coldness". Specifically, the health management support system 1 extracts factors strongly related to poor circulation from the medical examination information 321 by the cold sensitivity learning server 20.

(First example of analyzing the factors of "coldness")
For example, the sensitivity to cold learning server 20 analyzes the factors of "coldness" by binary logistic regression analysis. Specifically, the cold sensitivity learning server 20 applies binomial logistic regression analysis to the cold sensitivity information 320 and the medical examination information 321. This makes it possible to extract factors that have a significant effect on poor circulation. In addition, the cold sensitivity learning server 20 may extract factors that have a significant effect on poor circulation by applying binomial logistic regression analysis to the information indicating the symptoms of "coldness" and the medical examination information 321. The information indicating the symptom of "coldness" is information that can be used as a judgment material for determining whether or not the patient has poor circulation, and is, for example, information indicating low body temperature, low skin temperature, low blood flow, and the like. .. The blood flow rate here particularly indicates the blood flow rate of the skin. Table 1 shows an example of the results of binomial logistic regression analysis.

Table 1 shows the dependent variables on the left and the factors that have a significant effect on the right, along with their odds ratios. A body temperature of less than 36 means that the body temperature is less than 36 degrees, a skin temperature of less than 30 means that the skin temperature of the fingertip is less than 30 degrees, a blood flow rate of 10 or less means that the blood flow rate is 10 ml / min or less, and blood. A flow rate of 20 or less means that the blood flow rate is 20 ml / min or less. The inventors have stated that "coldness" including cold disease includes body composition (BMI, body fat ratio, muscle mass, visceral fat, bone density, aldosterone, vasopresin, etc.), blood (hemoglobin, hematocrit, MCV, MCH, MCHC, etc.) Finding that factors related to anemia), glucose metabolism (blood glucose level, HbA1C, C peptide, insulin, glycoalbumin, etc.), blood pressure, and renal function (urea nitrogen, creatinine, uric acid, etc.) have a significant effect. rice field.

(Second example of analyzing the factors of "coldness")
For example, the poor circulation learning server 20 analyzes the factors of "coldness" by the Bayesian network. Specifically, the cold sensitivity learning server 20 applies a Bayesian network to the cold sensitivity information 320 and the medical examination information 321. The sensitivity to cold learning server 20 acquires factors that have a particularly strong causal relationship with sensitivity to cold, extracted by the Bayesian network.

The inventors found that when the Bayesian network was applied to non-invasive data, poor circulation had a very strong causal relationship with gender, and gender was an important factor for predicting poor circulation. The non-invasive data here is information obtained from user U non-invasively, for example, gender, BMI (Body Mass Index), skin temperature, blood flow, systolic blood pressure, CES-D score (depression). An example of a standard for self-evaluation). The inventors also found that BMI and blood flow in men and BMI, systolic blood pressure, blood flow, and CES-D score in women each have a strong causal relationship with poor circulation. Although the case where the Bayesian network is applied to the non-invasive data has been described here as an example, the present invention is not limited to this. It goes without saying that the poor circulation learning server 20 may apply the Bayesian network to arbitrary medical examination information 321 to extract important factors for predicting poor circulation.

(Step S101)
Next, the health management support system 1 extracts important factors for predicting poor circulation. In the health management support system 1, cold sensitivity is predicted by the cold sensitivity prediction model 323 generated by using a machine learning method. The machine learning method may be any existing learning method, for example, supervised learning, random tree, C5.0 decision tree, CHAID (CHi-squared Automatic Interaction Detector), linear support vector machine, etc. This is a method using a learning model such as a neural network. Which training model is trained in what training data set determines how the sensitivity prediction model 323 predicts sensitivity to cold.

The sensitivity to cold learning server 20 extracts factors important for the prediction of poor circulation based on the prediction accuracy of the cold sensitivity prediction model 323. Specifically, the cold disease learning server 20 generates trained models in which the same machine learning algorithm model is trained with learning data sets having different contents. The cold sensitivity learning server 20 causes each trained model to predict cold sensitivity. The cold sensitivity learning server 20 uses the content trained by the trained model with the highest prediction accuracy as an important factor in predicting cold sensitivity. Further, the cold sensitivity learning server 20 uses the learned model with the highest prediction accuracy as the cold sensitivity prediction model 323.

The cold sensitivity learning server 20 may determine factors important for predicting cold sensitivity for each of a plurality of machine learning algorithms. Specifically, the cold disease learning server 20 generates a trained model in which learning data sets having different contents are trained in a model corresponding to each of a plurality of machine learning algorithms. The cold sensitivity learning server 20 causes each trained model to predict cold sensitivity. The sensitivity to cold learning server 20 determines a trained model with the highest prediction accuracy for each of the plurality of machine learning algorithms. The cold sensitivity learning server 20 uses the content trained by the trained model having the highest prediction accuracy for each model of the machine learning algorithm as an important factor for the prediction of each cold sensitivity. Further, the cold sensitivity learning server 20 may use the medical examination information 321 which is a common important factor in each model as a particularly important factor for the prediction of poor circulation.

Table 2 shows an example of the results of associating the prediction accuracy of the trained model with the important factors.

Table 2 shows the target, the machine learning algorithm, the prediction accuracy (%), and the important factors in order from the left. Subjects show what the model predicts, and this table shows poor circulation, hypothermia, low skin temperature, and low blood flow. The machine learning algorithm is applied as a learning model to show the machine learning method (algorithm), and this table shows each of the random tree, C5.0 decision tree, CHAID, and linear support vector machine. .. The prediction accuracy shows the system predicted by the model as a percentage. The important factor indicates the training data (explanatory variable) trained by the model. For example, in this table, the prediction accuracy of a trained model in which a random tree trains the correspondence between cold disease as teacher data (objective variable) and BMI, glycoalbumin, and hemoglobin as explanatory variables is 86.8%. It is shown that it was. As a result of extracting the important factors of "coldness", the inventors predict "coldness" by body composition (BMI, body fat percentage, muscle mass, bone density, etc.), anemia, and depression (CES-D). It was found that factors related to glucose metabolism such as score), inflammation (high-sensitivity CRP), blood pressure, and glycoalbumin (for example, blood glucose level, HbA1C, C-peptide, insulin, etc.) are important factors. We also found that body composition, renal function, bone metabolism (TRACP-5b), and PWV are important factors in predicting body temperature. It was also found that body composition, glucose metabolism (blood glucose level, HbA1C, C peptide, insulin, glycoalbumin, etc.) and cortisol are important factors in predicting skin temperature. We found that body composition, glucose metabolism, and blood pressure are important factors in predicting blood flow.

(Step S102)
Next, the health management support system 1 evaluates the risk of "coldness" on health. Specifically, the constitution classification server 30 evaluates the risk of "coldness" on health by performing cross-tabulation analysis with medical examination information 321 which is considered to be strongly related to health and poor circulation.

For example, the constitution classification server 30 performs dummy variable conversion processing for each item of the health diagnosis, setting the normal range to 0 (zero) and setting the case outside the normal range to 1. In addition, the constitution classification server 30 sets the case of having a disease in the questionnaire to the user U as 1, and sets the chi-square test (independence test) of the contingency table (1 × m cross tabulation table) and the number of associations (phi coefficient). ), Risk estimation, and adjusted relative risk (odds ratio) by Mantel-Haenszel. Table 3 shows an example of items that are significantly related to "cold sensitivity".

In Table 3, the medical examination items (health examination items) are shown on the left side, and the adjusted relative risk is shown on the right side together with the significance probability. From this table, the medical examination items that are considered to be correlated (related) with "cold sensitivity" are those related to low BMI, anemia, and insulin resistance symptoms (in this table, high HbA1C, high C peptide, and insulin). Resistance), and has been shown to be depressed.

In addition, the constitution classification server 30 has an effect on health by performing cross-tabulation analysis with medical examination information 321 which is considered to be strongly related to health and cold symptoms (for example, hypothermia, hyposkin temperature, hypoblood flow). Assess the risk of "coldness". As a method for evaluating the risk of these factors on health, the same method as the above-mentioned method (method for cross-tabulation analysis with medical examination information 321 considered to be strongly related to health and poor circulation) can be applied.

Table 4 shows an example of items that are significantly related to “hypothermia”.

In Table 4, as in Table 3, the medical examination items (health examination items) are shown on the left side, and the adjusted relative risk is shown on the right side together with the significance probability. From this table, the medical examination items that are considered to have a strong correlation (association) with "decrease in body temperature" are those related to renal function (high creatinine level), bone resorption (high TRACP-5b), and insulin resistance symptoms. It is shown to be an item (high C-peptide in this table).

Table 5 shows an example of items that are significantly related to "low skin temperature".

In Table 5, as in Table 3, the medical examination items (health examination items) are shown on the left side, and the adjusted relative risk is shown on the right side together with the significance probability. From this table, the medical examination items that are considered to have a strong correlation (association) with "low skin temperature" are low BMI and insulin resistance symptoms (in this table, hyperglycemia, high HbA1C, etc.) High C-peptide, insulin resistance) has been shown.

Table 6 shows an example of items that are significantly related to “low blood flow”.

In Table 6, as in Table 3, the medical examination items (health examination items) are shown on the left side, and the adjusted relative risk is shown on the right side together with the significance probability. From this table, the medical examination items that are considered to have a strong correlation (association) with "decrease in blood flow" are those related to low BMI, anemia, and insulin resistance symptoms (in this table, hyperglycemia and high HbA1C). , High C-peptide, high glycoalbumin, insulin resistance), and bilirubin.

(Step S103)
Next, the health management support system 1 analyzes the causal relationship between the "cold" factors. Specifically, the constitution classification server 30 evaluates the risk of poor circulation and the symptoms of coldness (hypothermia, low skin temperature, low blood flow) as shown in Tables 3 to 6 and causes poor circulation. An index that can be classified (stratified), that is, an index marker 322 is determined.

For example, the constitution classification server 30 has a correlation with each of the symptoms of coldness (hypothermia, hyposkin temperature, low blood flow) among the medical examination items considered to correlate with "cold sensitivity", and the correlation coefficient. The medical examination items having a different relationship with each other are designated as index markers 322.

Specifically, the constitution classification server 30 extracts medical examination items that correlate with each of "cold sensitivity" and cold symptoms (hypothermia, hyposkin temperature, hypoblood flow). The constitution classification server 30 selects items common to each of "cold sensitivity" and cold symptoms (hypothermia, low skin temperature, low blood flow) from the extracted medical examination items. The constitution classification server 30 calculates the correlation coefficient with each of "cold sensitivity" and cold symptoms (hypothermia, low skin temperature, low blood flow) for the selected medical examination item. When the code of the calculated correlation coefficient is different for each of "cold sensitivity" and cold symptoms (hypothermia, low skin temperature, low blood flow), the constitution classification server 30 sets the medical examination item as an index marker 322. To decide.

For example, in the examples of Tables 3 to 6, it is shown that all the tables correlate with the medical examination items related to the symptoms of insulin resistance. That is, according to Table 3, the medical examination items related to insulin resistance (high HbA1C, high C peptide, insulin resistance) have a negative correlation with cold sensitivity. In other words, it can be said that the higher the value of the medical examination items related to insulin resistance symptoms, the less likely it is to have poor circulation. In addition, according to Table 4, the medical examination item (C-peptide) related to insulin resistance has a negative correlation with body temperature. In other words, it can be said that the higher the value of the medical examination items related to insulin resistance symptoms, the lower the body temperature tends to be. In addition, according to Table 5, the medical examination items (hyperglycemia, high HbA1C, high C peptide, insulin resistance) related to the symptoms of insulin resistance have a positive correlation with skin temperature. In other words, it can be said that the higher the value of the medical examination items related to insulin resistance symptoms, the higher the skin temperature tends to be. In addition, according to Table 6, the medical examination items (hyperglycemia, high HbA1C, high C peptide, high glycoalbumin, insulin resistance) related to insulin resistance symptoms have a positive correlation with blood flow (at the fingertips). Have. In other words, it can be said that the higher the numerical value of the medical examination item related to the symptoms of insulin resistance, the higher (more) the blood flow tends to be.

When the results shown in Tables 3 to 6 are obtained, the constitution classification server 30 has a medical examination that correlates in common with each of "cold sensitivity" and cold symptoms (hypothermia, hyposkin temperature, hypoblood flow). Among the items, a high C-peptide is selected as a medical examination item having a different sign of the correlation coefficient, and the selected high-C peptide is determined as an index marker 322.

The results in Tables 3 to 6 suggest that C-peptide is involved in heat dissipation due to hypothermia and increased skin blood flow. That is, it is considered that as the C-peptide increases, it causes hypothermia, while increasing the skin temperature and increasing the skin blood flow. In addition, as the amount of C-peptide increases, the awareness of poor circulation tends to disappear, so that it may act so as not to make the person feel "cold".

(Step S104)
Next, the health management support system 1 performs "cold" classification and feature extraction by Two Step cluster analysis and categorization. Specifically, the constitution classification server 30 classifies whether or not the constitution of the target user TU is based on coldness by using whether or not he / she has poor circulation and the index marker 322. In addition, the information presentation server 40 presents information on disease prevention and constitution improvement according to the constitution of the target user TU based on the results classified by the constitution classification server 30.

FIG. 7 is a diagram illustrating a process performed by the health management support system 1 according to the embodiment. FIG. 7 schematically shows the types classified by the constitution classification server 30. The constitution classification server 30 classifies the constitution of the user U as cold or non-cold based on the cold sensitivity information 320. The constitution classification server 30 further classifies the sensitivity to cold according to the value of the index marker 322 (here, C-peptide). Similarly, the constitution classification server 30 further classifies non-cold sensitivity according to the value of the index marker 322 (here, C-peptide). Here, an example of classification according to whether the value of C-peptide is within the normal range or the value of C-peptide is higher than the normal range is shown. As a result of examining index markers effective for classification of constitutional characteristics for poor circulation and cold index (body temperature, skin temperature, blood flow) by Two Step cluster analysis, this C-peptide has good cluster quality (Chosion and demarcated silhouette). It was found that the index can be classified into 4 categories (0.5 or more).

That is, in this figure, the constitution classification server 30 determines the constitution of the user U.
1) Type T1 with poor circulation and normal C-peptide range
2) Type T2 with poor circulation and high C-peptide
3) Type T3 with non-cold sensitivity and normal range of C-peptide
4) Type T4 with non-cold sensitivity and high C-peptide
An example of classifying into the categories of is shown.

The information presentation server 40 presents information on disease prevention and constitution improvement according to which type of type T1 to T4 the constitution of the target user TU is. The information presentation server 40 makes presentations by type based on, for example, the statistics of an unspecified user U. Specifically, the constitution classification server 30 classifies the type of each user U based on the cold sensitivity information 320 and the medical examination information 321 of the unspecified user U. The information presentation server 40 extracts the tendency of the disease for each type based on the medical examination information 321 of the user U for each type. For example, as a result of statistical processing, the information presentation server 40 tends to have a lean type with low BMI and few skeletal muscles, low water content and low heat production (heat produced in the living body), and capillaries due to saccharification. It is assumed that there is a tendency for blood vessels to be damaged, a tendency for food to be thin, and a low sense of health, a decrease in reproductive function, and a risk of frailty in old age. Type T2 has problems with water metabolism and tends to swell the lower limbs, leptin is high and insulin resistance tends to occur, food tends to be vigorous, lower limb pain and physical dysfunction and sarcopenia, and elderly people are concerned about renal dysfunction. It shall be. Type T3 is concerned about a tendency to be healthy rather than poor circulation, a tendency to maintain homeostasis, an increase in bone metabolism, and an overimmunity. Type T4 tends to be visceral fat obesity (metabo), high peripheral blood circulation, low basal metabolism, low temperature sensitivity, health anxiety, metabolic syndrome and decreased adrenal function. It shall be. The information presentation server 40 presents according to the type of the target user TU based on a tendency predetermined for each type.

The information presentation server 40 may further subdivide the cause of type T1 by using the medical examination information 321. For example, the information presentation server 40 sets the target user TU classified as type T1 into a type having a low BMI and a large amount of subcutaneous fat in women (type T11), a type having a low BMI in men (type T12), and BNP (brain natriuresis). It is classified into a type with high diuretic polypeptide (type T13), a type with low blood pressure (type T14), a type with high glycoalbumin (type T15), and the like. Thereby, in the health management support system 1, it is possible to show the presentation according to each type (for example, types T11 to T15) in which a specific type (for example, type T1) in cold sensitivity is further subdivided.

By performing such constitution classification, the health management support system 1 can not only grasp the cause of "coldness" and the potential risk of illness or disorder, but also a precise method of health promotion and disease prevention for each cause. May be presented. For example, the health management support system 1 may present treatment with Chinese herbs according to the type. For example, the health management support system 1 has Ginseng Yoei-to and Juzen-taiho-to for type T1, Hachimijio-maru for type T2, Rokumiji-komaru for type T4, and for types T11 to T15, respectively. Tokishakuyakusan, Juzentaihoto, Hochuekkito, Kamikihito, Kamiyoyosan, etc. may be presented.

In addition, the health management support system 1 may present dietary improvement and functional foods according to the type. For example, the health management support system 1 is seeded with high-quality protein and iron supplementation for type T1, restriction of electrolytes and sugars for type T2, dietary fiber and glucosamine, carotenoids and broccoli for type T3, and reduced salt for type T4. Invite them to consume species, seaweeds and cruciferous vegetables. Further, the health management support system 1 may present dietary improvement and functional foods corresponding to the user's medical examination information even for the types (types T11 to T15) in which the type T1 is subdivided.

Furthermore, it is speculated that C-peptide acts on the hypothalamic receptor VACM-1 to reduce homeostatic function. Therefore, it may be presented to inhibit the binding between C-peptide and VACM-1. Thereby, it is possible to suppress the deterioration of the homeostatic function.

In addition, C-peptide may be presented to act on the oxytocin receptor of another hypothalamic receptor. In addition, it may be presented to inhibit the binding between C-peptide and vasopressin receptors V1a and V1b. This improves the autonomic nervous system and can suppress the decline in homeostatic function. By making such a presentation, it is possible to present a means that is more useful for improving QOL and improving health, or preventing diseases.

[Configuration of computer system 100]
Next, the configuration of the computer system 100 of each device included in the health management support system 1 will be described. FIG. 8 is a block diagram showing the configuration of the computer system 100. The computer system 100 includes a communication unit 101, an input unit 102, an output unit 103, a storage unit 104, and a CPU (Central Processing Unit) 105. These components are communicably connected to each other via bus B.

The communication unit 101 is a communication module such as a communication IC (Integrated Circuit). The input unit 102 is for input including, for example, a pointing device such as a mouse and a touch pad, a keyboard, a microphone, a camera, and various other sensors (for example, an acceleration sensor, a gyro sensor, a GPS (Global Positioning System) sensor, and a proximity sensor). It is a module. The input unit 102 may be integrally configured with the display as a touch panel. The output unit 103 is an output module for a display panel, a speaker, a woofer, or the like.

The storage unit 104 includes, for example, an HDD (Hard Disk Drive), an SSD (Solid State Drive), an EEPROM (Electrically Erasable Program Memory), a ROM (Read-Ony Memory), a ROM (Read-Ony Memory), and a ROM (Read-Only Memory) Memory Memory Memory Memory It is a module for. The storage unit 104 is not limited to the built-in one, and may be an external storage module connected by a digital input / output port or the like such as USB (Universal Serial Bus).
The CPU 105 executes various programs stored in the storage unit 104 and controls each configuration included in the computer system 100.
The above is the description of the configuration of the computer system 100.

[Configuration of cold sensitivity learning server 20]
Next, the configuration of the poor circulation learning server 20 will be described. FIG. 9 is a block diagram showing an example of the configuration of the cold sensitivity learning server 20 according to the embodiment. The sensitivity to cold learning server 20 includes a communication unit 21, a storage unit 22, and a control unit 23. The communication unit 21 communicates with another device connected to the communication network NW. Other devices here are a biometric information storage server 10, a constitution classification server 30, an information presentation server 40, and a user terminal 50.

The storage unit 22 stores variables, various data, programs, and the like necessary for realizing the functions of the poor circulation learning server 20. The storage unit 22 stores the learning data candidate 220, the definite learning data 221 and the cold disease prediction model 222. The learning data candidate 220 is information indicating a learning data candidate. The learning data candidate 220 is, for example, information in the medical examination information 321 that has a correlation with poor circulation and can be an index for predicting poor circulation. The definite learning data 221 is information indicating a medical examination item confirmed as learning data among the learning data candidates 220. The cold sensitivity prediction model 222 is information indicating a learning model (cold sensitivity prediction model) created by the cold sensitivity learning server 20.

The control unit 23 controls each configuration of the poor circulation learning server 20. The control unit 23 is realized, for example, by the CPU of the poor circulation learning server 20 executing a program stored in advance in the storage unit 22. Further, the control unit 23 may be realized as an integrated circuit such as an ASIC (Application Specific Integrated Circuit).

The control unit 23 includes, for example, a cold sensitivity information acquisition unit 230, a medical examination information acquisition unit 231, a preprocessing unit 232, a model creation unit 233, and a learning data determination unit 234. The poor circulation information acquisition unit 230 acquires the cold sensitivity information of the unspecified user U from the biological information storage server 10 via the communication unit 21. The medical examination information acquisition unit 231 acquires the medical examination information of the unspecified user U from the biological information storage server 10 via the communication unit 21.

The pretreatment unit 232 performs pretreatment for creating a model (cold sensitivity prediction model 222). The preprocessing unit 232 creates, for example, a training data set to be trained by the model. The training data set is a set of training data (explanatory variable) and teacher data (objective variable). The preprocessing unit 232 analyzes the cause of coldness and extracts learning data candidates from the medical examination information. The preprocessing unit 232 generates a learning data set by associating the training data candidates with the information on poor circulation as teacher data.

The model creation unit 233 creates a cold sensitivity prediction model 222. The model creation unit 233 trains the learning model by adjusting the parameters of the learning model so that the output obtained when the learning data is input to the learning model approaches the teacher data. The model creation unit 233 trains the learning model until a predetermined end condition is satisfied, and sets the learning model that satisfies the predetermined end condition as a trained model. The model creation unit 233 creates a plurality of trained models that have learned different contents by changing the contents of the training data. Alternatively, the model creation unit 233 creates a plurality of trained models having different algorithms.

The learning data determination unit 234 causes each of the trained models to predict the sensitivity to cold, and based on the prediction accuracy, the one with the highest prediction accuracy is referred to as the sensitivity prediction model 222. The learning data determination unit 234 stores the created cold sensitivity prediction model 222 in the storage unit 22. The learning data confirmation unit 234 confirms the learning data candidates trained by the cold sensitivity prediction model 222 as training data.
The above is the description of the configuration of the poor circulation learning server 20.

[Configuration of constitution classification server 30]
Next, the configuration of the constitution classification server 30 will be described. FIG. 10 is a block diagram showing the configuration of the constitution classification server 30. The constitution classification server 30 includes a communication unit 31, a storage unit 32, and a control unit 33.

The communication unit 31 communicates with another device connected to the communication network NW. Other devices here are a biological information storage server 10, a cold sensitivity learning server 20, an information presentation server 40, and a user terminal 50.

The storage unit 32 stores variables, various data, programs, and the like necessary for realizing the functions of the constitution classification server 30. The storage unit 32 stores the cold sensitivity information 320, the medical examination information 321, the index marker 322, the cold sensitivity prediction model 323, and the classification determination threshold value 324. The sensitivity to coldness information 320 is information on the sensitivity to coldness of the target user TU. The medical examination information 321 is the medical examination information of the target user TU. The index marker 322 is information about the index marker of the target user TU. The poor circulation prediction model 323 is information indicating the cold sensitivity prediction model created by the cold sensitivity learning server 20. The classification determination threshold value 324 is information indicating a threshold value for classifying poor circulation by an index marker.

The control unit 33 controls each configuration of the constitution classification server 30. The control unit 33 is realized, for example, by the CPU of the constitution classification server 30 executing a program stored in advance in the storage unit 32. The control unit 33 includes, for example, a cold sensitivity information acquisition unit 330, a medical examination information acquisition unit 331, an index marker selection unit 332, a cold sensitivity prediction unit 333, a cold constitution classification unit 334, and a classification result output unit 335. ..

The constitution classification server 30 has a pre-preparation phase and a classification execution phase, and performs different processing in each phase. Hereinafter, the preparation phase and the classification execution phase will be described in this order.

(Preparation phase)
The poor circulation information acquisition unit 330 acquires the cold sensitivity information of the unspecified user U from the biological information storage server 10 via the communication unit 31. The medical examination information acquisition unit 331 acquires the medical examination information of the unspecified user U from the biological information storage server 10 via the communication unit 31.

The index marker selection unit 332 selects the index marker 322 based on the relationship between the cold sensitivity information and the medical examination information of the unspecified user U. Since the method for selecting the index marker 322 by the index marker selection unit 332 has already been described, the description thereof will be omitted. The index marker selection unit 332 stores the selected index marker 322 in the storage unit 32.

(Classification execution phase)
The cold sensitivity information acquisition unit 330 acquires the cold sensitivity information of the target user TU from the user terminal 50 via the communication unit 31. The medical examination information acquisition unit 331 acquires the medical examination information of the target user TU from the user terminal 50 via the communication unit 31.

The cold sensitivity prediction unit 333 predicts whether or not the target user TU has poor circulation by inputting the medical examination information of the target user TU into the cold sensitivity prediction model 323. The cold sensitivity prediction unit 333 outputs the prediction result to the cold constitution classification unit 334.

The cold constitution classification unit 334 determines the constitution of the target user TU to be any of the types based on the cold sensitivity (for example, types T1 to T4) based on the prediction result by the cold sensitivity prediction unit 333 and the numerical value of the index marker in the target user TU. Classify into crabs. The classification result output unit 335 outputs the classification result by the cold constitution classification unit 334 to the information presentation server 40 and the user terminal 50.
The above is the description of the configuration of the constitution classification server 30.

[Summary of this embodiment]
As described above, the health management support system 1 according to the embodiment includes a cold sensitivity information acquisition unit 330, a medical examination information acquisition unit 331, an index marker selection unit 332, a cold constitution classification unit 334, and a classification result output. A unit 335 is provided. The cold sensitivity information acquisition unit 330 acquires the cold sensitivity information 320 of the unspecified user U. The medical examination information acquisition unit 331 acquires the medical examination information 321 of the unspecified user U. The medical examination information acquisition unit 331 selects the index marker 322 based on the characteristics extracted from the cold sensitivity information 320 and the medical examination information 321 of the unspecified user U. The cold constitution classification unit 334 classifies the constitution of the target user TU into any type based on the cold constitution based on the cold sensitivity information 320 of the target user TU and the numerical value of the index marker 322 in the target user TU. As a result, the health management support system 1 according to the embodiment can select an index marker from the biological information, and can classify the sensitivity to cold using the index marker. Therefore, it is possible to provide appropriate information depending on what type of sensitivity to cold is.

Here, the sensitivity to cold sensitivity 320 is an example of "first information".
The cold sensitivity information acquisition unit 330 is an example of the “first acquisition unit”.
The medical examination information 321 is an example of "second information".
The medical examination information acquisition unit 331 is an example of the "second acquisition unit".
The index marker selection unit 332 is an example of the “selection unit”.
The cold constitution classification unit 334 is an example of the "classification unit".
The classification result output unit 335 is an example of the “output unit”.
The sensitivity to cold learning server 20 is an example of a “learning device”.
The constitution classification server 30 is an example of a “classification device”.
The model creation unit 233 is an example of the “learning unit”.

Further, in the above-described embodiment, the case where the cold constitution classification unit 334 acquires the cold sensitivity information 320 of the target user TU predicted by the cold sensitivity prediction unit 333 has been illustrated and described. However, it is not limited to this. The cold constitution classification unit 334 may acquire information indicating whether or not the target user TU has poor circulation from the user terminal 50. in this case. In the input form of the "health management application" displayed on the user terminal 50, an input field for inputting information indicating whether or not the patient has poor circulation is displayed. The target user TU performs an operation of inputting whether or not the person has poor circulation in the input form, and then, for example, touch-operates a button indicating transmission. As a result, the constitution classification server 30 receives the cold sensitivity information of the target user TU. The cold constitution classification unit 334 classifies the type of the target user TU by using the cold sensitivity information of the target user TU received by the constitution classification server 30.

Further, a program for realizing the functions of the above-mentioned health management support system 1, cold disease learning server 20, and constitution classification server 30 is recorded on a computer-readable recording medium, and the program recorded on the recording medium is recorded on the computer system. You may perform the process by reading it into the server and executing it. Here, "loading and executing a program recorded on a recording medium into a computer system" includes installing the program in the computer system. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. Further, the "computer system" may include a plurality of computer devices connected via a network including a communication line such as the Internet, WAN, LAN, and a dedicated line. Further, the "computer-readable recording medium" refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, or a CD-ROM, or a storage device such as a hard disk built in a computer system. As described above, the recording medium in which the program is stored may be a non-transient recording medium such as a CD-ROM. The recording medium also includes an internal or external recording medium that can be accessed from the distribution server to distribute the program. The code of the program stored in the recording medium of the distribution server may be different from the code of the program in a format that can be executed by the terminal device. That is, the format stored in the distribution server does not matter as long as it can be downloaded from the distribution server and installed in a form that can be executed by the terminal device. The program may be divided into a plurality of parts, downloaded at different timings, and then combined by the terminal device, or the distribution server for distributing each of the divided programs may be different. Furthermore, a "computer-readable recording medium" is a volatile memory (RAM) inside a computer system that serves as a server or client when a program is transmitted via a network, and holds the program for a certain period of time. It shall also include things. Further, the above program may be for realizing a part of the above-mentioned functions. Further, it may be a so-called difference file (difference program) that can realize the above-mentioned function in combination with a program already recorded in the computer system.

1 ... Health management support system 10 ... Biological information storage server 20 ... Cold disease learning server 21 ... Communication unit 22 ... Storage unit 23 ... Control unit 230 ... Cold disease information acquisition unit 231 ... Medical examination information acquisition unit 232 ... Preprocessing unit 233 ... Model Creation unit 234 ... Learning data confirmation unit 30 ... Constitution classification server 31 ... Communication unit 32 ... Storage unit 320 ... Cold disease information 321 ... Medical examination information 33 ... Control unit 330 ... Cold disease information acquisition unit 331 ... Medical examination information acquisition unit 332 ... Index Marker selection unit 333 ... Cold disease prediction unit 334 ... Cold constitution classification unit 335 ... Classification result output unit 40 ... Information presentation server 50 ... User terminal

Claims (8)

The first acquisition unit that acquires the first information indicating whether or not the user has poor circulation,
A second acquisition unit that acquires a second information including a measurement item and a numerical value as a measurement result of the user's biological information, and a second acquisition unit.
Based on the first information and the characteristics extracted from the second information, the first type of poor circulation and the numerical value for the measurement item of the biological information corresponds to the cold sensitivity, and the poor circulation and the living body. The second type in which the numerical value for the measurement item of information does not correspond to the symptom of coldness, the third type in which the numerical value for the measurement item of biometric information corresponds to the symptom of non-coldness, and the biometric information The fourth type in which the numerical value for the measurement item of is not corresponding to the symptom of poor circulation, the selection unit for selecting the index marker which is the biological information as the index to be classified into, and
Based on the information indicating whether or not the target user has poor circulation and the numerical value of the index marker in the target user, the constitution of the target user is determined by the first type, the second type, the third type, and the above. A classification unit that classifies into any of the 4th types,
An output unit that outputs the classification results classified by the classification unit, and
Equipped with a,
The selection unit extracts correlation information that correlates with cold disease and two cold symptoms from the second information, and of the correlation information, one symptom of the two cold symptoms and the other symptom. The correlation information in which the correlation information and the one symptom show a positive correlation and the correlation information and the other symptom show a negative correlation is used as the index marker.
Health management support system. The symptom of coldness is when the body temperature corresponds to hypothermia, the blood flow corresponds to the low blood flow value, or the skin temperature corresponds to the low skin temperature value.
The non-cold symptom is when the body temperature does not correspond to hypothermia, the blood flow does not correspond to the low blood flow value, or the skin temperature does not correspond to the low skin temperature value.
The health management support system according to claim 1. The classification unit
The type in which the first information indicates that the patient has poor circulation and the numerical value of the index marker is in the normal range is classified into the first type.
A type in which the first information indicates that the patient has poor circulation and the numerical value of the index marker is not in the normal range is classified into the second type.
The type in which the first information indicates that the person does not have poor circulation and the numerical value of the index marker is in the normal range is classified into the third type.
A type in which the first information indicates that the person does not have poor circulation and the numerical value of the index marker is not in the normal range is classified into the fourth type.
The health management support system according to claim 1 or 2. Learning to generate a cold sensitivity prediction model that predicts whether or not the target user has poor circulation from the second information of the target user by letting the learning model learn the correspondence between the first information and the second information. With more parts,
The classification unit, the first information predicted the target user by using the poor circulation prediction model, and classify based on the index markers,
The health management support system according to claim 1. The learning unit creates a plurality of trained models having different learning contents, and uses the trained model selected based on the prediction accuracy of each of the created trained models as the cold sensitivity prediction model.
The health management support system according to claim 4. The first acquisition unit that acquires the first information indicating whether or not the user has poor circulation,
A second acquisition unit that acquires a second information including a measurement item and a numerical value as a measurement result of the user's biological information, and a second acquisition unit.
Learning to generate a cold sensitivity prediction model that predicts whether or not the target user has poor circulation from the second information of the target user by letting the learning model learn the correspondence between the first information and the second information. Department and
With a learning device that has
Based on the first information and the characteristics extracted from the second information, the first type of poor circulation and the numerical value for the measurement item of the biological information corresponds to the cold sensitivity, and the poor circulation and the living body. The second type in which the numerical value for the measurement item of information does not correspond to the symptom of coldness, the third type in which the numerical value for the measurement item of biometric information corresponds to the symptom of non-coldness, and the biometric information The fourth type in which the numerical value for the measurement item of is not corresponding to the symptom of poor circulation, the selection unit for selecting the index marker which is the biological information as the index to be classified into, and
Based on the information indicating whether or not the target user has poor circulation and the numerical value of the index marker in the target user, the constitution of the target user is determined by the first type, the second type, the third type, and the above. A classification unit that classifies into any of the 4th types,
An output unit that outputs the classification results classified by the classification unit, and
Have a,
The selection unit extracts correlation information that correlates with cold disease and two cold symptoms from the second information, and of the correlation information, one symptom of the two cold symptoms and the other symptom. The correlation information in which the correlation information and the one symptom show a positive correlation and the correlation information and the other symptom show a negative correlation is used as the index marker.
Sorting device and
Health management support system equipped with. The first acquisition unit acquires the first information indicating whether or not the user has poor circulation, and obtains the first information.
The second acquisition unit acquires the second information including the measurement item and the numerical value as the measurement result of the biometric information of the user.
Based on the characteristics extracted from the first information and the second information, the selection unit determines the sensitivity to cold, the first type in which the numerical value for the measurement item of the biological information corresponds to the sensitivity to cold, and the sensitivity to cold. The second type, in which the numerical value for the measurement item of biometric information does not correspond to the symptom of coldness, and the third type, which is non-cold and the numerical value for the measurement item of biometric information corresponds to the symptom of non-coldness, and non-cold sensitivity. Select the index marker, which is biometric information that serves as an index to classify into the fourth type, which is present and the numerical value for the measurement item of biometric information does not correspond to the symptom of poor circulation.
Based on the information indicating whether or not the target user has poor circulation and the numerical value of the index marker in the target user, the classification unit determines the constitution of the target user as the first type, the second type, and the third type. Classify into type and any of the above 4 types,
The output unit outputs the classification result classified by the classification unit .
The selection unit extracts correlation information that correlates with cold disease and two cold symptoms from the second information, and of the correlation information, one symptom of the two cold symptoms and the other symptom. The correlation information showing a positive correlation between the correlation information and the one symptom and a negative correlation between the correlation information and the other symptom is used as the index marker .
Health management support method. On the computer
First acquisition means for acquiring first information indicating whether or not the user has poor circulation,
A second acquisition means for acquiring second information including measurement items and numerical values as a measurement result of the user's biological information,
Based on the first information and the characteristics extracted from the second information, the first type of poor circulation and the numerical value for the measurement item of the biological information corresponds to the cold sensitivity, and the poor circulation and the living body. The second type in which the numerical value for the measurement item of information does not correspond to the symptom of coldness, the third type in which the numerical value for the measurement item of biometric information corresponds to the symptom of non-coldness, and the biometric information A selection means for selecting an index marker, which is biometric information that serves as an index for classifying into the fourth type in which the numerical value for the measurement item of is not a symptom of poor circulation.
Based on the information indicating whether or not the target user has poor circulation and the numerical value of the index marker in the target user, the constitution of the target user is determined by the first type, the second type, the third type, and the above. Classification means for classifying into any of the 4th types,
An output means that outputs the classification result classified by the classification means,
It is a program to execute
In the selection means, correlation information that correlates with cold disease and two cold symptoms is extracted from the second information, and among the correlation information, one symptom of the two cold symptoms and the other symptom are obtained. The correlation information in which the correlation information and the one symptom show a positive correlation and the correlation information and the other symptom show a negative correlation is used as the index marker.
program.

Images