KR101940807B1 - Apparatus and method for predicting future health - Google Patents

Abstract

The method of predicting future health of a healthcare system according to the present invention includes the steps of receiving a user's health information, pre-processing the health information, searching for similar cases from the health database using the pre- Analyzing the health pattern of the user and predicting future health measurements based on similar cases, designing a health healing plan corresponding to the analyzed health pattern or the predicted health measurement, And outputting the corresponding health healing information to the display device.

Description

[0001] APPARATUS AND METHOD FOR PREDICTING FUTURE HEALTH [0002]

The present invention relates to a health care device and its future health prediction method.

Most future health predictions predicted future disease outbreaks based on historical statistics. And the prediction was a prediction of the disease level. For example, "If you smoke more than 20 cigarettes a day for 20 years or more, the probability of getting lung cancer increases by more than 200% compared to a person who does not smoke at all." On the other hand, modern people make various measurements in regular health examinations and medical examinations in order to know their current health condition. Systolic / diastolic blood pressure, fasting blood glucose, total cholesterol, HDL, LDL, triglyceride, and the like may be examples of such measurements. These can be called health metrics, but no technology has been developed to predict the future trends of health metrics themselves.
Health measurement is of significance in the sense that it can be used as an indicator of disease. For example, a systolic blood pressure of 140 or more can be diagnosed as a disease called hypertension. Predictions of disease are generally long-term predictions, such as after 10 years in general. However, people are also interested in what will happen next year. In this respect, providing a model suitable for short-term prediction is also a necessary technology for developing a new service market.

It is an object of the present invention to provide a device for future health measurement and a method for its long-term / short-term prediction.

A method for predicting future health of a healthcare apparatus according to an embodiment of the present invention includes: receiving health information of a user; pre-processing the health information; searching for similar cases from a health database using the pre- Analyzing the health pattern of the user and predicting future health metrics based on the similar cases, designing a health healing plan corresponding to the analyzed health pattern or the predicted health metric, And outputting health healing information corresponding to the healing plan to the display device.
In an embodiment, the health information includes health information such as height, weight, waist circumference, systolic blood pressure, diastolic blood pressure, pre-eclampsia, total cholesterol, triglyceride, HDL cholesterol, LDL cholesterol, hemoglobin, urinary glucose, And may include at least one of urinary PH, urinary protein, serum creatinine, (serum geothi) AST, (serum lipid) ALT, and gamma lipoprotein, and the health of the user includes sex, age, / High blood pressure / diabetes / hyperlipidemia / pulmonary tuberculosis past disease status, family stroke / heart disease / hypertension / diabetes / other (including cancer) disease patients, smoking status, past smoking period, past day smoking amount, current smoking period, current smoking amount, Habit, amount of one drink, intense exercise days more than 1 week, median exercise days more than 30 minutes, walking days more than 30 minutes a week, and body mass.
In one embodiment of the present invention, the step of searching for similar cases may include calculating a weight for each feature of the health information, grouping the users into groups having similar health states using the weights, Calculating the similarity between the health status of the at least one user and the health status of the user, and retrieving the similar cases from the health database using the weight and the similarity.
In one embodiment, the step of predicting the health measurement includes analyzing and predicting a health information change pattern into a healing knowledge base that stores the similar cases, and analyzing and predicting the health information of the user with the similar cases And performing the recognition.
In one embodiment, the step of predicting the future health measurements includes proceeding with a similar case grouping corresponding to the user's health information, analyzing health factors associated with the group, and using the health factors And estimating the health measurement value.
In an embodiment, the health measurement includes systolic blood pressure, and the health parameters may include information relating to at least one of the presence or absence of drinking, smoking, exercise, BMI (body mass index), and total cholesterol.
The step of performing the similar case grouping may include performing a regression analysis operation on similar cases in a time series, calculating a distance between the regression value and the actual blood pressure value based on the result of the regression analysis, And a step of performing similar case grouping using the set of the actual values of the similar cases and the set of weights.
In an embodiment, the regression analysis operation is a linear regression analysis operation.
In an embodiment, the similar case grouping may be any one of the rising, maintaining, and falling groups.
In an embodiment, the similar case grouping may be accomplished by further combining weight values inversely proportional to the difference between the linear regression analysis value and the actual value for the time series health measurement.
In the embodiment, the similar case grouping may use a K-means method of calculating a representative value of a cluster and repeating a process of dividing data based on the representative value.
In one embodiment, the step of predicting the health measurements further comprises calculating the importance of each of the health factors, and by combining the predicted health measurements with the importance, a final health measurement predicted value is calculated .
In an embodiment, the health database may be a personal health information case database including a personal health checkup, a medical record, and a life log for many years similar to the sample cohort database of the National Health Insurance Corporation.
The health management apparatus according to the embodiment of the present invention includes a similar case retrieval unit for retrieving similar cases corresponding to the user's health information from the health group database, analyzing the health pattern of the health information using the similar cases, A health pattern analysis and future health prediction unit for grouping the similar cases and analyzing health factors of each of the grouped similar cases and predicting future health measurement values corresponding to the health information using the analyzed health factors; A health healing planning unit for designing a health healing plan corresponding to the analyzed health pattern or the predicted health metric, and a display device for displaying the health healing plan.
In an exemplary embodiment, the health information input unit may further include a health information input unit receiving the health information, and a health information preprocessing unit filtering the input health information, normalizing the filtered health information, and grouping the normalized health information have.
In an embodiment, the similar case searching unit can search for the similar cases using the weight and the similarity corresponding to the health information.
In one embodiment, the health pattern analysis and future health prediction unit performs a regression analysis operation of each of the similar cases and calculates a weight based on the distance between the actual value corresponding to the health information and the regression analysis value And perform grouping using the actual value and the weight.
In an embodiment, the regression analysis operation is a linear regression analysis operation, and the grouping of the similar cases may be determined as any of an ascending group, a maintenance group, and a descending group.
In an embodiment, the health pattern analysis and future health prediction unit may use the cluster k grouping for grouping the similar cases.
In an embodiment, the health care apparatus is an electronic apparatus, and the electronic apparatus may be a wearable apparatus.
A computer program product according to an embodiment of the present invention includes a step of receiving a user's health information, a step of preprocessing the health information, a step of retrieving similar cases from a health database using the preprocessed health information, Analyzing the user's health pattern and predicting future health measurements; designing a health healing plan corresponding to the analyzed health pattern or the predicted health measurement; Outputting the health healing information to the display device may be encoded into a storage medium.
In one embodiment, the step of predicting the health pattern analysis and the future health measurement comprises: grouping the similar cases of the personal-public health database into a plurality of clusters, , And using the center value of each of the aligned clusters as a prediction candidate value.

The apparatus and the method for predicting future health metrics according to embodiments of the present invention can predict future health metrics based on pseudo time series cases based on time series data.

1 is a diagram illustrating an exemplary health service providing system according to an embodiment of the present invention.
2 is a view showing an embodiment of a health information preprocessing unit according to an embodiment of the present invention.
FIG. 3 is a diagram illustrating an exemplary operation of the health information preprocessing unit shown in FIG. 2. FIG.
4 is a diagram illustrating an example of a similar case search unit according to an embodiment of the present invention.
FIG. 5 is a diagram illustrating an exemplary method of calculating the similarity of the similar case search unit shown in FIG.
FIG. 6 is a diagram illustrating an exemplary operation of the health pattern analysis and future health prediction unit according to the embodiment of the present invention.
FIG. 7 is a flow chart illustrating an exemplary case grouping operation shown in FIG. 6; FIG.
8 is a diagram illustrating an exemplary linear regression analysis according to an embodiment of the present invention.
9 is a diagram illustrating an exemplary similar grouping according to an exemplary embodiment of the present invention.
FIG. 10 is a diagram illustrating an exemplary embodiment of individual health factor based prediction based on group-by-group health factor analysis according to an embodiment of the present invention.
11 is a diagram illustrating an embodiment of future health measurement prediction according to an embodiment of the present invention.
FIG. 12 is a diagram illustrating an example of a final future health measurement prediction that reflects the importance of health factors according to an embodiment of the present invention.
13 is an exemplary diagram illustrating data variability according to an embodiment of the present invention.
FIG. 14 is a view showing an embodiment of the operation of the health healing planning unit according to the embodiment of the present invention.
15 is a flowchart illustrating an exemplary health management method of a health information service system according to an embodiment of the present invention.
16 is an exemplary illustration of a network environment including an electronic device according to various embodiments.
17 is a view illustrating an exemplary health service providing system according to an embodiment of the present invention.

여기서 dmax는 f와 f'의 차이에 대한 절대값의 최대값이다.
한편, (t+ti)에 존재하는 유사 사례들의 실제값 집합{f}와 가중치 집합{w}를 이용하여 그룹핑이 실시될 수 있다. 유사 사례들의 그룹은 크게 상승/유지/하강 그룹의 3가지로 구분될 수 있다. 상승 그룹은 기존 값을 기준으로 값의 상승이 이루어진 경우이다. 유사하게, 유지/하강 그룹은 값이 유지되거나 상승한 그룹을 의미한다.
도 9는 본 발명의 실시 예에 따른 유사 그룹 그룹핑을 예시적으로 보여주는 도면이다. 도 9를 참조하면, 그룹핑을 할 때 w 값의 사용 방법은 다음과 같다. 데이터의 그룹핑은 클러스터링이라고도 불린다.
실시 예에 있어서, 클러스터링은 K-means 방법을 이용할 수 있다. K-means 방법은 데이터들을 k개의 클러스터(군집)들로 분할한다. 좀 더 상세하게 K-means 방법은 군집의 대표값을 계산하고, 대표값들을 기준으로 데이터들을 분할하는 과정을 반복한다. 자신의 군집을 결정할 때 자신과 군집의 대표값 사이의 거리가 가장 짧은 군집으로 자신의 소속 군집을 결정함으로써, 각 데이터가 분할된다.
여기서, 군집의 대표값을 계산할 때, 가중치(w) 값이 이용된다. 가중치(w)가 큰 값을 가질수록 군집의 대표값 계산에 더 많은 기여를 하는 것이다. 왜냐하면 가중치(w)가 큰 값을 가진다는 것은, 데이터의 변동성이 작다는 것을 의미이다. 반대로 가중치(w)가 작은 값을 가진다는 것은 데이터의 변동성이 크다는 의미이다. 군집의 대표값(Ck)은 아래의 수학식에 의해 계산될 수 있다.

군집화 알고리즘은, 군집을 초기화시키는 과정도 대표값을 계산하는 방법 못지 않게 중요하다. 본 발명의 유사 사례 그룹핑 방법은 상승/유지/하강이라는 명확히 정의된 군집으로 정의되기 때문에, 이것들에 적합한 초기화 과정이 필요하다.
우선적으로, 유지 그룹(군집)은 대표값이 0으로 고정이다. 상승과 하강 그룹은 유지 그룹의 대표값(0)과의 거리의 절대값을 기준으로 오름차순으로 순서를 정할 때(예를 들어), 1/3 변위치(quantile) 내에 들지 않은 데이터(비 유지 그룹(군집) 데이터)를 이용하여 계산될 수 있다.
비 유지 그룹(군집) 데이터들 중에서 양수의 거리값을 가지는 데이터는 상승 그룹(군집)에 할당되고, 음수의 거리값을 가지는 데이터는 하강 그룹(군집)에 할당될 수 있다. 이렇게 초기화된 군집들을 이용하여 대표값 계산과 데이터 군집 분할이 반복될 수 있다. 단, 유지 그룹(군집)의 대표값은 반복 계산에서도 항상 0으로 고정될 수 있다.
도 10은 본 발명의 실시 예에 따른 그룹별 연관 건강 인자 분석을 통한 개별 건강 인자 기반 예측에 대한 실시 예를 예시적으로 보여주는 도면이다. 도 10에서 건강 feature = 건강 계측치, 연관 건강 feature = 건강 인자이다.
수축기 혈압의 연관 건강 인자가 음주와 담배(흡연) 유무(여부)라고 했을 때, 각 그룹에 대한 인자가 도 10에 도시된 바와 같이 분석될 수 있다.
예를 들어, G1 그룹은 총 85명의 유사 사례가 존재하며, 담배 유는 80이고 담배 무는 5이다. 그리고, 음주 유는 45이고, 음주 무는 40이다. 이와 같이 각 그룹의 건강 인자의 분포를 분석하는 것이 그룹별 연관 건강 인자 분석이다.
마지막으로 미래 건강 계측치 예측 동작은 다음과 같이 요약할 수 있다, t+t_i 일 때의 각 그룹의 대표값(일종의 centroid)를 이용하여 나의 중요 건강 인자와 동일한 데이터를 반영하여 건강 계측치의 (인자 반영) 미래 추이를 예측할 수 있다. 미래 추이 예측, 즉, 미래 건강 계측치 예측은 두 단계로 이루어 진다. 첫째로 개별 인자의 영향을 반영한 미래 건강 계측치를 예측하는 단계이고, 두번째로 건강 인자들의 중요도를 반영한 최종 미래 건강 계측치를 예측하는 단계이다.
도 11은 본 발명의 실시 예에 따른 미래 건강 계측치 예측에 대한 실시 예를 보여주는 도면이다. 도 11에서, 예측 쿼리에서, 즉, 나의 건강 인자를 담배(흡연)는 무이고 음주는 유라고 가정하겠다. 그리고 11에 도시된 바와 같이, G1 그룹은 총 85명이 속해 있고 그 대표값을 140이라 하고, G2 그룹에는 소속 사례가 없고, G3그룹은 총 15명이 소속되어 있고 그 대표 값이 100이라 가정하겠다.
우선, 나의 건강 인자 중 하나인 담배 인자를 반영하면, G1 그룹에는 담배 무 인 사람이 5명, G3 그룹에는 담배 무가 10명이다. G1과 G3 그룹의 대표값과 이 분포를 반영하면, 담배 인자를 반영한 미래 건강 계측치(예를 들어 혈압)의 예측 값은 도 10에 도시된 바와 같이 113이 된다.
유사한 방법으로 음주 인자를 반영한 미래 건강 계측치(여기서는 혈압)의 예측 값은 136이 된다. 여기까지 개별 건강 인자의 영향을 반영한 미래 건강 계측치 예측이다.
건강 인자들의 중요도를 반영한 최종 미래 건강 계측치 예측의 실시 예는 다음과 같다. 우선 건강 인자들의 중요도는 미리 계산되어 있다고 가정한다.
도 12는 본 발명의 실시 예에 따른 건강 인자들의 중요도를 반영한 최종 미래 건강 계측치 예측에 대한 실시 예를 보여주는 도면이다. 도 12를 참조하면, 음주의 중요도가 0.8이고 담배의 중요도가 0.2라고 가정하면, 개별 인자의 영향을 반영한 미래 건강 계측치 예측값과 건강 인자의 중요도를 결합함으로써, 최종 미래 건강 계측치 예측 값은 117.6이 된다.
한편, 데이터 변동성을 반영한 예측 방법은, 데이터 자체의 변동성을 반영함으로써 군집의 대표값을 계산하는 것으로부터 출발하였다. 회귀분석 동작을 수행할 때, 실제값(f)과 예측값(f') 사이의 차이(d)는 데이터의 변동성을 의미한다. 이러한 차이 값(d)이 크면 클수록, 데이터의 변동성이 크다고 할 수 있다. 본 발명의 예측 방법은 예측 값 계산에서 변동성이 큰 데이터가 변동성이 작은 데이터와 동일한 영향력이 반영되는 것이 아닌 차등적인 영향력을 반영시킬 수 있다. 좀 더 구체적으로 변동성이 큰 데이터는 예측값 계산에서 작은 영향력을 갖도록 하고, 변동성이 작은 데이터는 큰 영향력을 갖도록, 예측 값 계산에 반영된다. 이러한 영향력은 수학식 1과 유사한 형태의 가중치로 반영될 수 있다.
도 13은 본 발명의 실시 예에 따른 데이터 변동성을 예시적으로 보여주는 도면이다. 도 13을 참조하면, 개별 인자의 영향을 반영한 미래 건강 계측치 예측에 있어서, 데이터 자체의 변동성을 반영한 예측 방법은 다음과 같다
G1 그룹의 담배를 피우지 않는(담배 무) 유사 사례 기반의 예측값은 수학식 3과 같이 계산될 수 있다.

여기서,

G3 그룹의 담배 무 유사 사례 기반의 혈압 예측값도 수학식 3과 유사한 방법으로 계산할 수 있다.
G1 그룹과 G3 그룹의 유사 사례를 모두 반영한 담배 무의 혈압 예측 값은 수학식 4와 같이 계산될 수 있다.

음주 유인 유사 사례를 반영한 혈압 예측 값도 수학식 4와 유사한 방법으로 계산할 수 있다. 마지막으로 담배(흡연)와 음주 습관을 반영한 최종 혈압 예측 값은 수학식 5에 의해 계산될 수 있다.

본 발명의 실시 예에 따른 미래 건강 계측치에 대한 예측 방법은, 시계열 데이터(예를 들어 5년치의 혈당 데이터 시계열 데이터 등)를 기반으로 찾은 유사 시계열 사례를 근거로 하여 미래 건강 계측치에 대한 예측을 할 수 있다. 예를 들어 '당신의 혈당이 미래에 어느 정도의 범위를 가질 것이다'를 예측할 수 있다. 또한 5년, 10년 뒤의 장기 예측뿐만 아닌 1년, 2년, 3년 후의 미래와 같은 단기 예측이 가능하다.
도 14는 본 발명의 실시 예에 따른 건강 힐링 플래닝부(150)의 동작에 대한 실시 예를 보여주는 도면이다. 도 14를 참조하면, 건강 힐링 플래닝부(150)는 다음과 같이 진행될 수 있다.
건강 힐링 플래닝부(150)는 건강 패턴 분석 및 미래 건강 예측부(140)로부터 결과값을 입력 받고, 힐링 지식베이스(400)를 근거로 하여 분석한 해당 개인의 건강 패턴을 개선시키는 정보를 생성할 수 있다. 즉, 건강 힐링 플래닝부(150)는 힐링 지식베이스(400)에 있는 각 질병별 위험도에 따른 개선 플래닝에 대한 지식을 이용함으로써 개인별 건강 상태 및 패턴에 맞는 건강 힐링 플랜을 설계할 수 있다.
또한 건강 힐링 플래닝부(150)는 건강 패턴 분석 및 미래 건강 예측부(140)에서 예측한 사용자의 건강 특징 변화 패턴 및 미래 건강 특징 수치를 근거로 하여 힐링 지식베이스(400)에 저장된 건강 특징 레벨맵, 질병별 플래닝맵과의 조합을 통해서 다양한 개인별 맞춤형 건강 상태 및 패턴에 맞는 건강 힐링 플랜을 설계할 수 있다.
건강 힐링 정보 출력부(160)는 설계한 개인별 맞춤형 건강 상태 및 패턴에 맞는 건강 힐링 플랜을 사용자 디스플레이 장치 등 외부로 출력할 수 있다.
힐링 지식베이스(400)는 건강 특징 레벨맵, 질병별 플래닝맵 등을 저장하고 있다. 이것들을 참조하면 사용자가 예상되는 건강 특징의 수치에 따른 맞춤형 플랜이 생성될 수 있다.
건강 특징 레벨맵은 중요 건강 특징들의 정상, 위험, 이상 등에 대한 기준 정보들을 저장할 수 있다. 예를 들어, 건강 특징맵은 혈압의 정상 범위, 위험군 범위, 이상 범위에 대한 정보, 혈당의 정상 범위와 이상 범위에 대한 정보를 포함할 수 있다.
질병별 플래닝맵은 특정 질병을 개선하는 데 필요한 식이, 운동, 생활 습관 등에 대한 정보를 저장할 수 있다. 예를 들어, 질병별 플래닝맵은, 당뇨 환자가 먹어서는 안되는 음식군과 체중과 신장에 따른 적정 칼로리 계산법 등을 저장할 수 있다.
개인별로 예측된 건강 특징의 변화 패턴은 힐링 지식베이스(400)의 건강 특징 레벨맵, 질병별 플래닝맵과의 조합을 통해 다양한 맞춤형 힐링 플랜을 생성할 수 있다(S610). 여러 가지 힐링 플랜 중에서 개인에게 적합한 힐링 플랜이 선택될 수 있다(S620). 이후 선택된 개인 맞춤 힐링 플랜이 출력될 수 있다(S630). 선택된 맞춤 힐링에 대한 실제 힐링 행동 및 사용자 건강 특징 변화가 피드백 되는 기능이 추가될 수 있다(S640).
건강 힐링 정보 출력부(160)는 건강 힐링 플래닝부(150)의 개인별 맞춤형 건강 힐링 플랜을 입력 받고, 이를 외부로 출력할 수 있다. 또한, 건강 힐링 정보 출력부(160)는 건강 정보 입력부(110)에서 입력 받은 개인 건강 정보와 유사한 사례들을 검색한 패턴의 결과를 분석하고, 예측된 결과값을 출력할 수 있다.
도 15는 본 발명의 실시 예에 따른 건강 정보 서비스 시스템의 건강 관리 방법을 예시적으로 보여주는 흐름도이다. 도 1 내지 도 15를 참조하면, 건강 관리 방법은 다음과 같다.
건강 정보 입력부(110)는 다양한 건강 정보 입력(또는 수집) 장치 또는 수집 경로를 통해 개인별 건강 정보(예컨대, 성별, 연령, 키, 체중, 혈압, 혈당, 체질량 등)를 입력 받을 수 있다(S710). 건강 정보 전처리부(120)는 건강 정보 입력부(110)에서 입력 받은 개인별 건강 정보를 전처리 할 수 있다(S720). 이때, 건강 정보 전처리부(120)는 입력된 건강 정보 중에 주요 특징의 건강 정보를 추출하여 정규화한다. 물론, 누락된 건강 정보가 있을 경우에 누락된 건강 정보가 재입력 될 수 있다.
이후, 유사 사례 검색부(130)는 전처리된 건강 정보의 특징별 가중치(w)를 계산하고, 이를 근거로 하여 건강 그룹 데이터베이스(300) 내의 다양한 개인 건강 정보 사례와의 특징별 유사도 계산을 수행하고, 하나 이상의 유사 사례들을 검색할 수 있다(S730).
이후, 건강 패턴 분석 및 미래 건강 예측부(140)는 힐링 지식베이스(400)에 저장된 건강 정보 변화 패턴에 대한 분석 및 예측에 관한 지식 맵을 이용함으로써, 유사 사례 검색부(130)로부터 유사 사례의 패턴을 분석하고, 대응하는 개인의 건강 패턴을 예측할 수 있다. 특히, 건강 패턴 분석 및 미래 건강 예측부(140)는, 도 6에설명된 바와 같이 유사 사례들을 근거로 하여 미래 건강 계측치에 대한 예측을 수행할 수 있다(S740).
이후, 건강 힐링 플래닝부(150)는 건강 패턴 분석 및 미래 건강 예측부(140)로부터 결과값을 입력 받고, 힐링 지식베이스(400)에 저장된 질병별 위험도에 따른 개선 플래닝에 대한 지식을 이용함으로써 개인별 건강 상태 및 패턴에 맞는 건강 힐링 플랜을 설계할 수 있다(S750).
이후, 건강 힐링 정보 출력부(160)는 건강 힐링 플래닝부(150)로부터 건강 힐링 정보(즉, 건강 힐링 플랜)을 입력 받고 출력할 수 있다(S760).
도 16은 다양한 실시 예에 따른 전자 장치(1100)를 포함하는 네트워크 환경(1000)을 예시적으로 보여주는 도면이다. 도 16을 참조하면, 전자 장치(1100)는 버스(1110), 프로세서(1120), 메모리(1130), 입출력 인터페이스(1140), 디스플레이(1150), 통신 인터페이스(1160) 및 추가 기능 모듈(1170)을 포함할 수 있다.
버스(1110)는 전술한 구성요소들을 서로 연결하고, 전술한 구성요소들 간의 통신(예: 제어 메시지 등)을 전달하는 회로일 수 있다. 프로세서(1120)는, 예를 들어, 버스(1110)를 통해 전술한 다른 구성 요소들(예: 메모리(1130), 입출력 인터페이스(1140), 디스플레이(1150), 통신 인터페이스(1160), 추가 기능 모듈(1170) 등의 적어도 하나)로부터 명령을 수신하여, 수신된 명령을 해독하고, 해독된 명령에 따른 연산이나 데이터 처리를 실행할 수 있다.
메모리(1130)는, 프로세서(1120) 혹은 다른 구성 요소들(예: 입출력 인터페이스(1140), 디스플레이(1150), 통신 인터페이스(1160), 추가 기능 모듈(1170) 등의 적어도 하나)로부터 수신되거나 프로세서(1120) 혹은 다른 구성 요소들에 의해 생성된 명령 혹은 데이터를 저장할 수 있다. 메모리(1130)는, 예를 들어, 커널(1131), 미들웨어(1132), 어플리케이션 프로그래밍 인터페이스(API: application programming interface, 133) 혹은 어플리케이션(1134) 등의 프로그래밍 모듈들을 포함할 수 있다. 전술한 각각의 프로그래밍 모듈들은 소프트웨어, 펌웨어, 하드웨어 혹은 이들 중 적어도 둘 이상의 조합으로 구성될 수 있다.
커널(1131)은 나머지 다른 프로그래밍 모듈들, 예를 들어, 미들웨어(1132), API(1133) 혹은 어플리케이션(1134)에 구현된 동작 혹은 기능을 실행하는 데 사용되는 시스템 리소스들(예: 버스(1110), 프로세서(1120), 메모리(1130) 등의 적어도 하나)을 제어 혹은 관리할 수 있다. 또한, 커널(1131)은 미들웨어(1132), API(1133) 혹은 어플리케이션(1134)에서 전자 장치(1100)의 개별 구성 요소에 접근하여 제어 혹은 관리할 수 있는 인터페이스를 제공할 수 있다.
미들웨어(1132)는 API(1133) 혹은 어플리케이션(1134)이 커널(1131)과 통신하여 데이터를 주고 받을 수 있도록 중개 역할을 수행할 수 있다. 또한, 미들웨어(1132)는 어플리케이션(1134)으로부터 수신된 작업 요청들과 관련하여, 예를 들어, 어플리케이션(1134) 중 적어도 하나의 어플리케이션에 전자 장치(1100)의 시스템 리소스(예: 버스(1110), 프로세서(1120), 메모리(1130) 등의 적어도 하나)를 사용할 수 있는 우선 순위를 배정하는 등의 방법을 이용하여 작업 요청에 대한 제어(예: 스케쥴링, 로드 밸런싱 등의 적어도 하나)을 수행할 수 있다.
API(1133)는 어플리케이션(1134)이 커널(1131) 혹은 미들웨어(1132)에서 제공되는 기능을 제어하기 위한 인터페이스로, 예를 들어, 파일 제어, 창 제어, 이미지 처리, 문자 제어 등의 적어도 하나를 위한 적어도 하나의 인터페이스 혹은 함수(예: 명령어 등)를 포함할 수 있다.
다양한 실시 예에 있어서, 어플리케이션(1134)은 SMS(Short Message Service)/MMS(Multimedia Messaging Service) 어플리케이션, 이메일 어플리케이션, 달력 어플리케이션, 알람 어플리케이션, 건강 관리(health care) 어플리케이션(예: 운동량, 혈당 등의 적어도 하나를 측정하는 어플리케이션), 환경 정보 어플리케이션(예: 기압, 습도, 온도 정보 등의 적어도 하나를 제공하는 어플리케이션) 등의 적어도 하나를 포함할 수 있다.
추가적으로 혹은 대체적으로, 어플리케이션(1134)은 전자 장치(1100)와 외부 전자 장치(예: 전자장치(1200) 등) 사이의 정보 교환과 관련된 어플리케이션일 수 있다. 정보 교환과 관련된 어플리케이션은, 예를 들어, 외부 전자 장치에 특정 정보를 전달하기 위한 알림 전달(notification relay) 어플리케이션, 혹은 외부 전자 장치를 관리하기 위한 장치 관리(device management) 어플리케이션을 포함할 수 있다.
예를 들어, 알림 전달 어플리케이션은 전자 장치(1100)의 다른 어플리케이션(예: SMS/MMS 어플리케이션, 이메일 어플리케이션, 건강 관리 어플리케이션 혹은 환경 정보 어플리케이션 등)에서 발생한 알림 정보를 외부 전자 장치(예: 전자장치(1200) 등)로 전달하는 기능을 포함할 수 있다.
추가적으로 혹은 대체적으로, 알림 전달 어플리케이션은, 예를 들어, 외부 전자 장치(예: 전자장치(1200) 등)로부터 알림 정보를 수신하여 사용자에게 제공할 수 있다. 장치 관리 어플리케이션은, 예를 들어, 전자 장치(1100)와 통신하는 외부 전자 장치(예: 전자장치(1200) 등)의 적어도 일부에 대한 기능(예: 외부 전자 장치 자체(혹은, 일부 구성 부품)의 턴온/턴오프, 디스플레이의 밝기(혹은, 해상도) 조절 등의 적어도 하나), 외부 전자 장치에서 동작하는 어플리케이션 혹은 외부 전자 장치에서 제공되는 서비스(예: 통화 서비스, 메시지 서비스 등)를 관리(예: 설치, 삭제, 업데이트 등의 적어도 하나)할 수 있다.
실시 예에 있어서, 어플리케이션(1134)은 외부 전자 장치(예: 전자장치(1200) 등)의 속성(예: 전자 장치의 종류 등)에 따라 지정된 어플리케이션을 포함할 수 있다. 예를 들어, 외부 전자 장치가 MP3 플레이어인 경우, 어플리케이션(1134)은 음악 재생과 관련된 어플리케이션을 포함할 수 있다.
유사하게, 외부 전자 장치가 모바일 의료기기인 경우, 어플리케이션(1134)은 건강 관리와 관련된 어플리케이션을 포함할 수 있다. 실시 예에 있어서, 어플리케이션(1134)은 전자 장치(1100)에 지정된 어플리케이션 혹은 외부 전자 장치(예: 서버(1300), 전자장치(1200) 등)로부터 수신된 어플리케이션 중 적어도 하나를 포함할 수 있다.
입출력 인터페이스(1140)는, 입출력 장치(예: 센서, 키보드, 터치 스크린 등의 적어도 하나)를 통하여 사용자로부터 입력된 명령 혹은 데이터를, 예를 들어, 버스(1110)를 통해 프로세서(1120), 메모리(1130), 통신 인터페이스(1160), 추가 기능 모듈(1170) 등의 적어도 하나에 전달할 수 있다.
예를 들어, 입출력 인터페이스(1140)는 터치 스크린을 통하여 입력된 사용자의 터치에 대한 데이터를 프로세서(1120)로 제공할 수 있다. 또한, 입출력 인터페이스(1140)는, 예를 들어, 버스(1110)를 통해 프로세서(1120), 메모리(1130), 통신 인터페이스(1160), 추가 기능 모듈(1170) 등의 적어도 하나로부터 수신된 명령 혹은 데이터를 입출력 장치(예: 스피커, 디스플레이 등의 적어도 하나)를 통하여 출력할 수 있다. 예를 들어, 입출력 인터페이스(1140)는 프로세서(1120)를 통하여 처리된 음성 데이터를 스피커를 통하여 사용자에게 출력할 수 있다.
디스플레이(1150)는 사용자에게 각종 정보(예: 멀티미디어 데이터, 텍스트 데이터 등의 적어도 하나)를 표시할 수 있다.
통신 인터페이스(1160)는 전자 장치(1100)와 외부 전자 장치(예: 전자장치(1200), 서버(1300) 등) 간의 통신을 연결할 수 있다. 예를 들어, 통신 인터페이스(1160)는 무선 통신 혹은 유선 통신을 통해서 네트워크(1001)에 연결되어 외부 전자 장치와 통신할 수 있다.
무선 통신은, 예를 들어, WiFi(wireless fidelity), WiFi 다이렉트, BT(Bluetooth), NFC(near field communication), GPS(global positioning system), 셀룰러(cellular) 통신(예: LTE, LTE-A, CDMA, WCDMA, UMTS, WiBro 혹은 GSM 등) 등의 적어도 하나를 포함할 수 있다. 유선 통신은, 예를 들어, USB(universal serial bus), HDMI(high definition multimedia interface), RS-232(recommended standard 232), POTS(plain old telephone service) 등의 적어도 하나를 포함할 수 있다.
실시 예에 있어서, 네트워크(1001)는 통신 네트워크(telecommunications network)일 수 있다. 통신 네트워크는 컴퓨터 네트워크(computer network), 인터넷(internet), 사물 인터넷(internet of things), 전화망(telephone network) 등의 적어도 하나를 포함할 수 있다. 실시 예에 있어서, 전자 장치(1100)와 외부 전자 장치 간의 통신을 위한 프로토콜(예: transport layer protocol, data link layer protocol, physical layer protocol 등)은 어플리케이션(1134), API(1133), 미들웨어(1132), 커널(1131), 통신 인터페이스(1160) 등의 적어도 하나에서 지원될 수 있다.
실시 예에 있어서, 추가 기능 모듈(1170)은 전자 장치(1100)에서 구현되는 동작(혹은, 기능)들 중 적어도 하나의 동작을 수행함으로써, 전자 장치(1100)의 구동을 지원할 수 있다. 예를 들어, 서버(1300)는 전자 장치(1100)에 구현된 추가 기능 모듈(1170)을 지원할 수 있는 추가 기능 서버 모듈(1320)을 포함할 수 있다. 예컨대, 추가 기능 서버 모듈(1320)은 추가 기능 모듈(1170)의 적어도 하나의 구성 요소를 포함하여, 추가 기능 모듈(1170)이 수행하는 동작들 중 적어도 하나의 동작을 수행(예: 대행 등)할 수 있다.
추가 기능 모듈(1170)은, 다른 구성요소들(예: 프로세서(1120), 메모리(1130), 입출력 인터페이스(1140), 통신 인터페이스(1160) 등의 적어도 하나)로부터 획득된 정보 중 적어도 일부를 처리하고, 이를 다양한 방법으로 이용할 수 있다. 예를 들어, 추가 기능 모듈(1170)은 프로세서(1120)를 이용하여 혹은 이와는 독립적으로, 전자 장치(1100)가 다른 전자 장치(예: 전자장치(1200), 서버(1300) 등)와 연동하도록 전자 장치(1100)의 적어도 일부 기능을 제어할 수 있다. 추가 기능 모듈(1170)은 프로세서(1120) 혹은 통신 인터페이스(1160)에 통합될 수 있다. 실시 예에 있어서, 추가 기능 모듈(1170)의 적어도 하나의 구성은 서버(1300)(예: 추가 기능 서버 모듈(1320) 등)에 포함될 수 있으며, 서버(1300)로부터 추가 기능 모듈(1170)에서 구현되는 적어도 하나의 동작을 지원받을 수 있다.
한편, 본 발명은 다양한 종류의 전자 장치들에 건강 서비스를 제공할 수 있다.
도 17은 본 발명의 실시 예에 따른 건강 서비스 제공 시스템을 예시적으로 보여주는 도면이다. 도 17을 참조하면, 건강 서비스 제공 시스템(2000)은 복수의 단말 장치들(2100, 2200) 및 건강 서비스 제공 서버(2300)를 포함할 수 있다.
복수의 단말 장치들(2100, 2200) 각각은 도 1 내지 도 16에서 설명된 전자 장치로 구현될 수 있다.
건강 서비스 제공 서버(2300)는 건강 서비스를 제공하기 위한 시계열 빅데이터를 저장하고, 단말 장치들(2100, 2200) 중 적어도 하나의 요청에 따라 건강 서비스를 제공하도록 구현될 수 있다.
본 발명의 실시 예에 따른 플로우챠트 설명들, 혹은 블록 다이어그램들, 그것들의 조합들은 컴퓨터 프로그램 명령들에 의해 구현될 수 있다. 이러한 명령들은 범용 컴퓨터, 특수 목적 컴퓨터, 혹은 기타 프로그램 가능한 데이터 프로세싱 장치의 프로세서에 제공될 수 있다. 프로세서에 의해 실행될 때, 명령들은 다이어그램들의 각 블록 혹은 그것들의 조합에서, 기능들, 활동들 혹은 특수한 이벤트들을 구현하기 위한 수단들을 발생할 수 있다.
이와 관련하여, 플로우챠트의 각 블록 혹은 블록 다이어그램들은 모듈, 세그먼트, 혹은 코드의 일부에 대응할 수 있다. 여기서 코드의 일부는 특수한 논리 기능을 구현하기 위한 하나 이상의 명령어들을 포함할 수 있다. 한편, 임의의 블록과 연관된 기능은 도면들에 언급된 순서를 벗어날 수도 있다. 예를 들어, 두 개의 블록들이 연속적으로 도시되어 있지만, 실제로는 동시에 수행되거나 때로는 반대로 수행될 수도 있다.
당업자는 본 발명의 기술 사상을 장치, 시스템, 방법 혹은 컴퓨터 프로그램 제품으로 구현할 수 있다고 이해되어야 할 것이다. 따라서, 본 발명에 따른, 회로들, 모듈들, 구성들 혹은 시스템들로 언급된 것들은 하드웨어로 구현되거나 소프트웨어(펌웨어, 상주 소프트웨어, 마이크로코드 포함)로 구현되거나, 혹은 하드웨어와 소프트웨어의 조합으로 구현될 수 있다. 하드웨어와 소프트웨어의 조합은, 컴퓨터 읽기 가능한 프로그램 코드를 갖는 컴퓨터 읽기 가능 매체에 구현된 컴퓨터 프로그램 제품을 포함할 수 있다.
한편, 상술 된 본 발명의 내용은 발명을 실시하기 위한 구체적인 실시 예들에 불과하다. 본 발명은 구체적이고 실제로 이용할 수 있는 수단 자체뿐 아니라, 장차 기술로 활용할 수 있는 추상적이고 개념적인 아이디어인 기술적 사상을 포함할 것이다.BRIEF DESCRIPTION OF THE DRAWINGS The above and other objects, features and advantages of the present invention will be more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which: FIG.
The embodiments according to the concept of the present invention can make various changes and can take various forms, so that the embodiments are illustrated in the drawings and described in detail herein. It is not intended to be exhaustive or to limit the invention to the precise form disclosed, but on the contrary, is intended to cover all modifications, equivalents, or alternatives falling within the spirit and scope of the invention. The terms first, second, etc. may be used to describe various elements, but the elements should not be limited by terms. The terms may be named for the purpose of distinguishing one element from another, for example, without departing from the scope of the right according to the concept of the present invention, the first element may be referred to as a second element, The element may also be referred to as a first element.
It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between. Other expressions that describe the relationship between components, such as "between" and "between" or "neighboring to" and "directly adjacent to" should be interpreted as well.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, the terms " comprises ", or " having ", and the like, specify that the presence of the features, numbers, steps, operations, elements, But do not preclude the presence or addition of one or more other features, integers, steps, operations, components, parts, or combinations thereof.
Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the meaning of the context in the relevant art and, unless explicitly defined herein, are to be interpreted as ideal or overly formal Do not.
The apparatus for predicting future health according to an embodiment of the present invention and the method for predicting the future health predict a future trend of health measurement based on similar cases derived from longitudinal data related to health examination have. In an embodiment, the health measurement is a value that predicts, for example, " your blood sugar will have some range in the future ".
Such a health prediction may be capable of short-term prediction as well as long-term prediction. That is, short-term forecasts such as 1 year, 2 years, and 3 years future can be made as well as long-term forecasts after 5 years and 10 years. That is, the future health prediction method according to an embodiment of the present invention can be performed based on a pseudo time-series case found based on time-series data (for example, 5-year blood glucose data time series data).
In an embodiment of the present invention, health measurements may be values measured in systolic / diastolic blood pressure, fasting blood glucose, total cholesterol, HDL, LDL, triglyceride, and other health screenings.
1 is a diagram illustrating an exemplary health service providing system 10 according to an embodiment of the present invention. Referring to FIG. 1, the health service providing system 10 may include a health care apparatus 100, a health big database 200, a health group database 300, and a healing knowledge base 400.
The health care apparatus 100 may be implemented with various types of electronic devices. For example, the electronic device may be a personal computer, a notebook, a smart phone, a tablet PC, a wearable device, or the like. The health management apparatus 100 is connected to the health big database 200, the health group database 300, and the healing knowledge base 400 through a wired / wireless network and includes a health big database 200, a health group database 300 ), And the healing knowledge base (400). In another embodiment, the healthcare device 100 may be implemented to store the collected health information in at least one of the health big database 200, the health group database 300, and the healing knowledge base 400.
1, the health care apparatus 100 includes a health information input unit 110, a health information preprocessing unit 120, a similar case searching unit 130, a health pattern analysis and future health prediction unit 140, A healing planning unit 150, and a health healing information output unit 160. [
The health information input unit 100 receives personal health information (for example, sex, age, height, height, etc.) through medical record data, health check data, wearable health measurement smart device, Weight, blood pressure, blood sugar, body mass, etc.).
In an embodiment, the health information comprises at least one of the following information: a user's height, weight, waist circumference, systolic blood pressure, diastolic blood pressure, pre-eclampsia, total cholesterol, triglyceride, HDL cholesterol, LDL cholesterol, hemoglobin, urinary glucose, , Serum creatinine, (serum geothi) AST, (serum lipid) ALT, and gamma-lipoprotein.
In an embodiment, the health information may further include a health factor of the user. Herein, the health factor is the health status of the user, such as the sex, age, the presence of the patient's cerebral palsy / cardiovascular disease / hypertension / diabetes / hyperlipemia / pulmonary tuberculosis past disease, family stroke / heart disease / hypertension / diabetes / Exercise duration in the past, the amount of smoking in the past, the current smoking period, current smoking, drinking habits, the amount of alcohol consumed, intense exercise days more than 20 minutes a week, more than 30 minutes of medium exercise days, , And a body mass.
The health information preprocessing unit 120 stores the individual health information output from the health information input unit 110 into the similar case search unit 130, the health pattern analysis and future health prediction unit 140, and the health healing planning unit 150 ), Or may convert the data into a data format for storage in the healing knowledge base 400. [
The similar case searching unit 130 calculates weights of the features of the health information output from the health information preprocessing unit 120, groups them into groups having similar health states using the weights, You can search for cases similar to your current health status. In an embodiment, the similar case searching unit 130 may search for cases similar to the health state of at least one or more users. For example, the similar-case searching unit 130 may calculate the weight and similarity of each feature of the health information, and search for a similar case from the health group database 300 using the similarity.
The health pattern analysis and future health predicting unit 140 may compare a case similar to the health state of the user searched in the similar case searching unit 130 to a knowledge about analysis and prediction of the health information change pattern stored in the healing knowledge base 400 By using the map, matching and recognition with the retrieved similar cases can be performed. The health pattern analysis and future health prediction unit 140 may be implemented to predict future trends using time series data and similar time series cases.
The health healing planning unit 150 receives the results of the health pattern analysis and future health prediction unit 140 and generates information for improving the health pattern of the individual based on the healing knowledge base 400 can do. In other words, the health healing planning unit 150 can design a health healing plan that meets individual health conditions and patterns, using knowledge of improvement planning according to each disease-specific risk stored in the healing knowledge base 400. [
In addition, the health healing planner 150 searches for a health feature level (health level) stored in the healing knowledge base 400 based on the health pattern analysis and the health characteristic change pattern of the user predicted by the future health predictor 140, Maps, and disease-specific planning maps, you can design a health-healing plan that fits your personalized health status and / or pattern.
The health healing information output unit 160 may output a health healing plan corresponding to a personalized health condition and / or pattern designed by the health healing plan unit 150 to the outside such as a user's display device.
The health big database 200 can be implemented to store health information collected in real time from public health information database, health care records collected from medical institutions, health examination data, or wearable health collection apparatus by case. The health big database 200 can be implemented to store the time series health information that can confirm the change of the numerical value for each long term health information by case. That is, the health big database 200 may be implemented to store various types of individual health information cases.
In the embodiment, the health big database 200 is a personal health information case database including a multi-year personal health examination, a medical record, and a life log similar to a sample cohort DB published by the National Health Insurance Corporation .
On the other hand, the health big database 200 may include sets of health data that are not processed.
The health group database 300 can group the raw health data of the health big database 200 into classification elements such as age, sex, disease, health status, and store the grouped data for real-time similar case search . In an embodiment, the health group database 300 may be grouped by unique classification elements. In another embodiment, the health group database 300 may be grouped by combining one or more classification elements.
For example, the criteria for similar groups of health status may be specific diseases, such as sex (male, female), age (teenage, twenties, ...., over 80), or hypertension . In addition, groups may be classified according to their health status, such as healthy group or non-healthy group.
In an embodiment, the health data may subdivide the group into one or more conditions, e.g., one of a plurality of conditions, such as sex = "F" and age = "50" and disease = "hypertension" May be generated.
In an embodiment, any health case may be categorized and stored in one or more groups.
In an embodiment, health group database 300 may be implemented using parallel storage for real-time similar case retrieval.
Meanwhile, in FIG. 1, the health big database 200 and the health group database 300 are shown as separate configurations. However, the present invention is not limited thereto. The health big database 200 and the health group database 300 may be implemented in a single configuration. It is also possible to manage the items of the database in the health big database by a separate table for each health group, or to search by grouping directly from the health big database. The health big database 200 and the health group database 300 may be collectively referred to as a health database.
The healing knowledge base 400 may include a vector relating to health characteristic vector weights and feature-specific associations for each major disease. In addition, the healing knowledge base 400 has N recognizers for recognizing whether or not a major disease N (where N represents an arbitrary number of diseases) is disease or not. In addition, the healing knowledge base 400 has a knowledge map for analysis and prediction of the change pattern of time series health information. In addition, the healing knowledge base 400 has knowledge of the improvement plan according to each disease risk.
In addition, the healing knowledge base 400 analyzes the health big database 200 that has been subjected to the health data filtering to generate a disease-related key health characteristic-related map, a disease-related map, a health characteristic level map, Feature maps, disease-specific planning maps, and the like, which are obtained by mining the health database 200.
In an embodiment, data is input from a public healing record, and data pre-processing is performed to map (link table or network form) the association between disease and disease (e.g., the relationship between stroke and hypertension) , And a healing knowledge base 400 can be constructed by indexing and associating the relationship (for example, age-related, sex-related relationship for each disease) with major health characteristics of each disease.
For example, the results can be stored in the healing knowledge base 400 by mining the association between the disease and the characteristic and mining the association between the disease and the disease. In this way, it is possible to construct a map by setting levels according to the health characteristics, and to construct a related feature map for each major health feature or to construct a disease-specific planning map.
The healing knowledge base 400 stores a health feature level map, a disease-specific plan map, and the like, and can generate a customized plan according to a numerical value of an expected health feature by a user. The health feature level map stores baseline information about the normal, risk, and abnormality of important health features.
For example, it contains information on the normal range of blood pressure, range of risk, and abnormal range, and contains information on normal range and abnormal range of blood sugar. Disease-specific planning maps contain information on dietary, exercise and lifestyle habits needed to improve specific diseases. Examples of disease-specific planning maps include foods that diabetics should not eat, and how to calculate calorie counts according to body weight and height.
The variation pattern of the health characteristic predicted by the individual generates various customized healing plans through combination with the health feature level map of the healing knowledge base 400 and the disease-specific planning map. Of the various healing plans, a personalized healing plan can be chosen. The selected personalized healing plan can then be output.
And may further include the ability to feed back the actual execution healing behavior and user health feature changes for the selected custom healing. In addition, the health healing information output unit 160 may output the personalized health healing plan to the outside from the health healing planning unit 150. [ In addition, the health healing information output unit 160 may analyze the result of the pattern retrieval of cases similar to the personal health information received from the health information input unit 110, and output the result of the prediction.
2 is a view showing an embodiment of a health information preprocessing unit 120 according to an embodiment of the present invention. Referring to FIG. 2, the health information preprocessing unit 120 may include a health information feature extraction unit 121, a health information normalization unit 122, and a missing health information processing unit 123.
The health information feature extraction unit 121 receives individual health information through the health information input unit 110 (see FIG. 1), and extracts key or effective features from the inputted health information (S110). The health information of the user may include information on various features.
The healing knowledge base 400 stores key feature maps that significantly affect disease severity. For example, when the user selects hypertension as the disease selected by the health information input unit 110, there is a feature list that has an important effect on hypertension. The health information feature extraction unit 121 fetches the feature list (for example, systolic blood pressure, diastolic blood pressure, BMI, waist / hip ratio, hyperlipidemia etc.) from the healing knowledge base 400, Only the corresponding features from the information can be selected, and the selected features can be extracted with the health information feature.
If the name of the disease selected by the user or entered by the disease of interest is not in the list of features, key features corresponding to predefined major diseases (chronic diseases: hypertension, diabetes, myocardial infarction, hyperlipidemia, etc.) may be selected.
Thereafter, the health information normalization unit 122 may normalize the health information extracted by the health information feature extraction unit 121 (S120). The health information of the user is information including time series data of various lengths. Some fields of health information may be integer, or 'minor scores', by health. Some health information may include survey information in the form of 'yes' or 'no'.
Therefore, it is necessary to perform normalization (real form between 0 and 1) of various lengths and various types of data. Or a different time-series length for each user (for example, three-year value data, five-year value data, etc.), if the time-series length of the data for a specific disease analysis is at least 5, And a normalization process of a time series length using a representative value. The health information normalization unit 122 performs this process.
The missing health information processing unit 123 can process the missing information when there is missing health information. Since the health information may be omitted when inputting or collecting the health information by various types of health information input (or collection) devices, the missing health information processing unit 123 may process to retrieve and re-input the missing health information . If the health information entered by the user is omitted in this process, an interface for requesting re-input may be activated, and an interpolation method may be applied using the time-series health examination data inputted to process the missing part, Or a combination of these may complement the missing information.
FIG. 3 is a diagram illustrating an exemplary operation of the health information preprocessing unit 120 shown in FIG. 2 and 3, the processing operation of the health information can be proceeded as follows.
The health information preprocessing unit 120 performs health data filtering, performs health data filtering, analyzes the health big database 200 related to the filtered health data, A disease-related map, a health feature level map, an association feature map for each major health feature, a disease-specific planning map, and the like can be stored in the healing knowledge base 400. Here, the major health feature-related maps, disease-related maps, health feature level maps, major feature-related feature maps, disease-specific planning maps, and the like are obtained as a result of mining the health big database 200 (S110).
The health big database 200 is information collecting health data of various users and may include time series data of various lengths. The health information preprocessing unit 120 performs normalization (between 0 and 1 or 1 and 2) for certain data in order to generate the healing knowledge base 500 in the health big database 200 including the various lengths and various types of health data. 1) can be performed.
In addition, the missing data may be replaced with statistical values such as mean or median for the corresponding data of similar cases. Some health data may be generated with median values by closely interpolating the time interval or frequency between specific time series data (S120).
In order to retrieve a similar case in real time, the health big database 200 may divide and store health data with users having similar cases. At this time, a process of grouping into groups having similar health states can be performed. The criteria for similar groups of health status can be gender, age, etc., and can be a specific disease, such as with or without hypertension. In addition, groups may be distinguished by using a health condition classifier or the like as a healthy group or a dangerous group. In addition, the group may be subdivided into one or more conditions (S130).
FIG. 4 is a diagram illustrating an exemplary similar-case search unit 130 according to an exemplary embodiment of the present invention. Referring to FIG. 4, the similar case search unit 130 may include a feature weight calculation unit 131 and a feature similarity calculation unit 132.
The feature weight calculation unit 131 can calculate the weight of each feature of the health information. In the embodiment, the feature weight calculation unit 131 may assign a different weight to each feature of the health information. As described in FIGS. 2 and 3, the main features for searching for similar cases through the preprocessing process are selected. In addition, normalization has been underway for the analysis of these characteristics of health information.
The feature similarity calculation unit 132 can calculate the similarity of the user's health information to individual health cases belonging to the extracted similar case group. The 1: 1 similarity can be calculated using information on disease-specific association weights extracted from the healing knowledge base 400 (eg, for diabetes, blood glucose is weighted at 0.8 and age is weighted at 0.3) have.
For example, blood glucose levels? 0.8 is the weighted value of feature. The feature similarity calculation unit 132 compares the weight calculation value of each characteristic of the user's health information with the weight value of each feature of the similar case group and calculates the similarity value of the cases in the similar case group by the similarity calculation formula have.
In an embodiment, the similar-case searching unit 130 can search for one similar case. In another embodiment, the similar-case searching unit 130 may search for a plurality of similar cases.
It will be understood that the similar case search unit 130 according to the embodiment of the present invention is not limited to the calculation of similarity per feature. The similar case searching unit 130 according to the embodiment of the present invention may include any kind of process for searching for a similar health case for a user's health case. That is, the similar case searching unit 130 may include a process of searching for a health case showing a similar health condition to the user. The similar case searching unit 130 extracts a health care group similar to the user's health information from the health group database 300 (see FIG. 1), and extracts the user health information And a step of prioritizing by calculating the degree of similarity with 1: 1.
Specifically, the similar case searching unit 130 may convert the health information of the user into the group information for the health characteristics associated with the disease of interest. For example, if the blood pressure value is 120, the health information of the user can be converted into group information in which the blood pressure of the user is 120 among the groups stored in the health group database 300.
The data of the group matched from the health group database 30 can be extracted based on the health information of the user converted into the group information. For example, assuming that the input value of the user is "33 years old, blood sugar: 115, blood pressure: 123, disease of interest: diabetes", and the diabetes according to the healing knowledge base 400 is related to age and blood glucose level, A health case corresponding to "Age: 30, blood glucose level: fifth group" can be extracted from the database 300 as a user's similar case group.
FIG. 5 is a diagram illustrating an exemplary method of calculating the similarity of the similar case searching unit 130 shown in FIG. Referring to FIGS. 4 and 5, the similarity calculation method of the similar case search unit 130 can be performed as follows.
The similar case searching unit 130 receives a query of the user's own health information and a disease of interest from the user, and preprocesses the user's health information in response to the query. This preprocessing process may be similar to the health data filtering process described in FIG. A missing data restoration operation, a normalization operation of health feature values, and an interpolation operation of time series health information may be performed when the user's health feature is missing (S210).
The user can refer to key health feature information associated with the entered disease of interest with reference to the disease-specific key health feature association map in the healing knowledge base 400. [ A group to search for similar cases of users in the health group database 3050 may be selected using the user's disease and health information of interest (S220). The similarity degree between the selected group cases and the user health information in the health group database 300 may be calculated (S230). After calculating the similarity, the rank of the similarity degree may be calculated and given using the weighting information of the health feature information stored in the healing knowledge base 400 (S240). Thereafter, the upper similar cases may be output as many times as the system determines (S250).
FIG. 6 is a diagram illustrating an exemplary operation of the health pattern analysis and future health prediction unit 140 according to the embodiment of the present invention. Referring to FIG. 6, the health pattern analysis and future health prediction method of the future health prediction unit 140 may proceed as follows. The similar case grouping corresponding to the personal information is performed (S310), and the related health factor for each group is analyzed (S320). Thereafter, the future health measurement value can be predicted (S330).
The health pattern analysis and future health prediction unit 140 may receive an output value of the similar case search unit 130 as an input. That is, an individual's health measurement and related health factors to be predicted and similar cases can be provided as input values. The health pattern analysis and future health prediction unit 140 may be provided with the following input values. In the following, we will explain systolic blood pressure as an example of health measurement. However, it should be understood that the health measurements are not limited thereto, and that the health measurements of the present invention are applicable to all health measurements.
The input value of the individual is as follows. First, the input value may include a time series set of health metrics to be predicted, for example, systolic blood pressure (hereinafter "blood pressure"). That is, when the current time is t, {blood pressure (t-ti),? , The blood pressure (t-2), the blood pressure (t-1), and the blood pressure (t)}. Second, the input value may include a set of health factors associated with blood pressure (e.g., {drinking habits, smoking habits, exercise habits, BMI, total cholesterol,?) I will describe it as / useless, smoking / void.
In addition, the similar case input data may basically include a similar case set, i.e., {similar case 1, case 2,?, Case n}. Each similar case has the following form.
First, each similar case may include a time series set of health measurements. In this case, the similar case can be displayed in a form different from the input data of the individual. If the input data of the individual is the past time series based on the present, in the case of the similar case, the future time series is based on the present. That is, it is a set of blood pressure time series in the form of {blood pressure (t), blood pressure (t + 1), blood pressure (t + 2),?, Blood pressure (t + ti)
It can be strange that future time series values exist based on the present. Here, the future is based on the similar case mapping and the relative time is applied. For the sake of convenience of explanation, for example, if a person's input data blood pressure time series set is personal (blood pressure (2001), blood pressure (2002),? , Blood pressure (2005)} (the number in parenthesis is the year). The similar case for this is similar (blood pressure (1996), blood pressure (1996),? , Blood pressure (2000), blood pressure (2001),?, Blood pressure (2003)}. At this time, two time series are used to calculate the individual 's blood pressure (2001), blood pressure (2002),? (2000), blood pressure (2001), blood pressure (2000), and blood pressure (2000) are assumed to be similar to the blood pressure (T + 1), blood pressure (t + 2),?, Blood pressure (t + ti)}.
Second, each similar case may include a set of health factors associated with blood pressure, i. E., A set containing the same kind of personal input data. That is, if the personal input data has the form of {drinking / no smoking, smoking / no smoking}, the similar case also has the form of drinking / smoking, smoking / no smoking.
FIG. 7 is a flow chart illustrating an exemplary case grouping operation shown in FIG. 6; FIG. Referring to Figs. 1 and 7, the similar case grouping operation can be proceeded as follows.
First, a regression analysis operation for each similar case may be performed (S410). Here, the regression analysis operation can be performed by various methods. In an embodiment, a linear regression analysis operation may be performed. If the actual value is f, a regression analysis operation can be performed using the formula f '= ax + b. The x value can be a relative year value. That is, the time series set of similar cases is similar to {(blood pressure (t),? , The blood pressure (t + 1), the blood pressure (t + 2),?, And the blood pressure (t + ti)}, the set of x values is {0, 1, 2,? , ti} or {1, 2, 3,?, ti + 1}. Here, the set of f values is similar to the blood pressure (t),? , Blood pressure (t + 1), blood pressure (t + 2),?, Blood pressure (t + ti)}. When the above-described regression analysis operation is performed, a regression analysis value f 'can be derived.
Then, based on the distance d between the actual value f and the regression analysis value f ', the weight w may be calculated (S420). Thereafter, the grouping may be performed using the actual value set {f} and the weight set {w} of similar cases (S430).
8 is a diagram illustrating an exemplary linear regression analysis according to an embodiment of the present invention. Referring to FIG. 8, when the regression analysis operation is completed, the regression analysis value f 'can be obtained. Let the absolute value of the difference between the f 'value and the f value, that is, the distance, be d. The weight w can be calculated based on the d value. This weight w can be designed to be inversely proportional to the value of d. Then, it can be designed to have a value between [0, 1]. For example, the weight (w) can be defined as Equation (1).

Where dmax is the maximum value of the absolute value of the difference between f and f '.
On the other hand, grouping can be performed using the set of actual values {f} and the set of weights {w} of similar cases existing at (t + ti). The group of similar cases can be roughly divided into three categories: rising / holding / falling group. The rising group is a case where the value is raised based on the existing value. Similarly, a hold / descend group refers to a group in which a value is maintained or elevated.
9 is a diagram illustrating an exemplary similar grouping according to an exemplary embodiment of the present invention. Referring to FIG. 9, a method of using w value when grouping is as follows. Grouping of data is also referred to as clustering.
In an embodiment, clustering may use a K-means method. The K-means method divides the data into k clusters. More specifically, the K-means method calculates the representative value of the cluster and repeats the process of dividing the data based on the representative values. When determining its own cluster, each data is divided by determining its own cluster with the shortest distance between itself and the representative value of the cluster.
Here, when the representative value of the cluster is calculated, the weight value (w) is used. The larger the weight (w) is, the more the contribution is made to the representative value of the cluster. This is because the fact that the weight w has a large value means that the variability of the data is small. Conversely, the fact that the weight (w) has a small value means that the data is highly volatile. The representative value Ck of the cluster can be calculated by the following equation.

The clustering algorithm is as important as the method of calculating representative values. Since the similar case grouping method of the present invention is defined as a clearly defined cluster of rise / sustain / fall, an initialization procedure suitable for these is needed.
Firstly, the representative value of the maintenance group (cluster) is fixed to zero. When ascending and descending groups are ranked in ascending order based on the absolute value of the distance from the representative value (0) of the holding group (for example), data not contained in the 1/3 position quantile (Cluster) data).
Data having a positive distance value is assigned to a rising group (cluster), and data having a negative distance value can be assigned to a falling group (cluster) among non-maintaining group (cluster) data. Representative value calculation and data cluster partitioning can be repeated using the initialized clusters. However, the representative value of the maintenance group (cluster) can always be fixed to 0 even in the iterative calculation.
FIG. 10 is a diagram illustrating an example of individual health factor-based prediction through analysis of group-related health factors according to an embodiment of the present invention. In Fig. 10, health feature = health measurement, association health feature = health factor.
Assuming that the associated health factors of systolic blood pressure are alcohol and tobacco (smoking), the factors for each group can be analyzed as shown in FIG.
For example, there are 85 similar cases in the G1 group, with 80 cigarettes and 5 cigarettes. And drinking is 45 and drinking is 40. The analysis of the distribution of health factors in each group is the analysis of the associative health factors by group.
Finally, future health measurement forecasting operations can be summarized as follows: t + t _i (A kind of centroid) of each group at the time of birth can be used to predict the future trend of health measurement (reflecting factor) by reflecting the same data as my important health factor. Future trend forecasts, that is, forecasting of future health metrics, take two steps. The first step is to predict future health metrics that reflect the influence of individual factors. Second, the final health metrics that reflect the importance of health factors are predicted.
11 is a diagram illustrating an embodiment of future health measurement prediction according to an embodiment of the present invention. In Fig. 11, in the prediction query, that is, my health factor is assumed to be tobacco (smoking) is nothing and drinking is good. As shown in FIG. 11, it is assumed that a total of 85 people belong to the G1 group, its representative value is 140, there are no cases belonging to the G2 group, 15 persons belong to the G3 group, and its representative value is 100.
First of all, one of my health factors, the tobacco factor, reflects five people in the G1 group and ten people in the G3 group. Reflecting the representative values of the G1 and G3 groups and this distribution, the predicted value of the future health measurement (for example, blood pressure) reflecting the tobacco factor is 113 as shown in FIG.
In a similar way, the predicted value of the future health measurement (here, blood pressure) reflecting the drinking factor is 136. This is a prediction of future health measurements reflecting the influence of individual health factors.
An example of the final future health measurement prediction that reflects the importance of health factors is as follows. First, it is assumed that the importance of health factors is calculated in advance.
FIG. 12 is a diagram illustrating an example of a final future health measurement prediction that reflects the importance of health factors according to an embodiment of the present invention. Referring to FIG. 12, when the importance of drinking is 0.8 and the importance of cigarettes is 0.2, the predicted value of the final future health measurement is 117.6 by combining the predictive value of future health metrics reflecting the influence of individual factors and the importance of health factors .
On the other hand, the forecasting method reflecting data volatility started from calculating the representative value of the cluster by reflecting the volatility of the data itself. When performing the regression analysis operation, the difference (d) between the actual value (f) and the predicted value (f ') means the variability of the data. The greater the difference value d, the greater the variability of the data. The prediction method of the present invention can reflect differential effects in prediction value calculation rather than reflecting the same influence as data having a small degree of variability. More specifically, the volatility data has a small influence on the calculation of the predicted value, and the data with a small variation has a large influence, and is reflected in the prediction value calculation. This influence can be reflected as a weight of a form similar to Equation (1).
13 is an exemplary diagram illustrating data variability according to an embodiment of the present invention. Referring to FIG. 13, in the prediction of future health measurement values reflecting the influence of individual factors, a prediction method reflecting the variability of the data itself is as follows
A predictive value based on a similar case that does not smoke in the G1 group (no tobacco) can be calculated as shown in Equation (3).

here,

The blood pressure predictive value based on the non-smoking case of the G3 group can also be calculated by a method similar to that of Equation (3).
The prediction value of the blood pressure of the tobacco radish including all of the similar cases of the G1 group and the G3 group can be calculated as shown in Equation (4).

The blood pressure prediction value reflecting the similar case of drinking inducement can also be calculated by a method similar to the equation (4). Finally, the final blood pressure prediction value reflecting smoking (smoking) and drinking habits can be calculated by Equation (5).

The method of predicting future health metrics according to an embodiment of the present invention can be used to predict future health metrics based on pseudo time series cases based on time series data (for example, 5-year blood glucose data time series data) . For example, you can predict 'your blood sugar will have some range in the future'. Short-term forecasts such as 1, 2, and 3 years future are possible as well as long-term forecasts after 5 and 10 years.
FIG. 14 is a view showing an embodiment of the operation of the health healing planner 150 according to the embodiment of the present invention. Referring to FIG. 14, the health healing planing unit 150 may proceed as follows.
The health healing planing unit 150 receives the results from the health pattern analysis and future health prediction unit 140 and generates information for improving the health pattern of the individual based on the healing knowledge base 400 . That is, the health healing planning unit 150 can design a health healing plan that meets personal health conditions and patterns by utilizing the knowledge of improvement planning according to each disease risk level in the healing knowledge base 400.
The health healing planner 150 may further include a health feature level map stored in the healing knowledge base 400 based on the health pattern analysis and the future health characteristic change values of the user predicted by the future health predictor 140, , And disease-specific planning maps, you can design a health-healing plan that fits your personalized health status and patterns.
The health healing information output unit 160 can output a health healing plan corresponding to the designed personalized health state and pattern to the outside such as a user display device.
The healing knowledge base 400 stores health characteristic level maps, disease-specific planning maps, and the like. By referring to these, a customized plan can be created according to the numerical value of the health characteristic expected by the user.
The health feature level map may store reference information about normal, risk, anomalies, etc. of important health features. For example, the health characteristics map may include information on the normal range of the blood pressure, the range of the risk group, the range of the abnormal range, and the normal range and abnormal range of the blood glucose.
Disease-specific planning maps can store information about dietary, exercise and lifestyle habits needed to improve a particular disease. For example, disease-specific planning maps can store food groups that diabetics should not eat, and how to calculate calorie counts according to body weight and height.
The change pattern of the health characteristic predicted by the individual can be variously combined with the health feature level map of the healing knowledge base 400 and the disease-specific planning map to generate various customized healing plans (S610). A healing plan suitable for an individual among various healing plans may be selected (S620). The selected personalized healing plan may then be output (S630). An actual healing action for the selected personal healing and a function for feedback of the user's health feature change may be added (S640).
The health healing information output unit 160 receives the personalized health healing plan of the health healing planer 150 and outputs the personalized health healing plan to the outside. In addition, the health healing information output unit 160 may analyze the results of the patterns retrieved from cases similar to the personal health information received from the health information input unit 110, and output the predicted result values.
15 is a flowchart illustrating an exemplary health management method of a health information service system according to an embodiment of the present invention. Referring to FIGS. 1 to 15, the health care method is as follows.
The health information input unit 110 may receive personal health information (for example, sex, age, height, weight, blood pressure, blood glucose, body mass, etc.) through various health information input (or collection) devices or a collection path (S710) . The health information preprocessing unit 120 may preprocess individual health information received from the health information input unit 110 (S720). At this time, the health information preprocessing unit 120 extracts and normalizes the health information of the main feature among the inputted health information. Of course, missing health information may be re-entered if there is missing health information.
The similar case searching unit 130 calculates the weight w of each feature of the preprocessed health information and calculates the similarity of each feature in the health group database 300 based on the calculated weight w , One or more similar cases may be searched (S730).
Thereafter, the health pattern analysis and future health prediction unit 140 uses the knowledge map for analyzing and predicting the health information change pattern stored in the healing knowledge base 400, Analyze the pattern, and predict the health pattern of the corresponding individual. In particular, the health pattern analysis and future health prediction unit 140 may perform a prediction on future health measurement values based on similar cases as illustrated in FIG. 6 (S740).
Thereafter, the health healing planing unit 150 receives the results from the health pattern analysis and future health predicting unit 140, and uses the knowledge of the improvement planning according to the disease-specific risk stored in the healing knowledge base 400, A health healing plan can be designed according to health condition and pattern (S750).
Then, the health healing information output unit 160 may receive and output health healing information (i.e., health healing plan) from the health healing planing unit 150 (S760).
16 is an exemplary illustration of a network environment 1000 including an electronic device 1100 according to various embodiments. 16, an electronic device 1100 includes a bus 1110, a processor 1120, a memory 1130, an input / output interface 1140, a display 1150, a communication interface 1160 and an additional function module 1170, . ≪ / RTI >
The bus 1110 may be a circuit that interconnects the components described above and communicates (e.g., control messages, etc.) between the components described above. Processor 1120 may be coupled to other components (e.g., memory 1130, input / output interface 1140, display 1150, communication interface 1160, (E.g., at least one of the decrypted data 1170 and the like), decrypt the received command, and execute computation and data processing according to the decrypted command.
Memory 1130 may be received from processor 1120 or other components (e.g., at least one of input / output interface 1140, display 1150, communication interface 1160, add-on module 1170, Lt; RTI ID = 0.0 > 1120 < / RTI > or other components. The memory 1130 may include programming modules such as, for example, a kernel 1131, a middleware 1132, an application programming interface (API) 133, or an application 1134. Each of the above-described programming modules may be composed of software, firmware, hardware, or a combination of at least two of them.
The kernel 1131 may include system resources (e.g., bus 1110) used to execute operations or functions implemented in other programming modules, e.g., middleware 1132, API 1133, ), The processor 1120, the memory 1130, and the like). The kernel 1131 may also provide an interface for accessing and controlling or managing individual components of the electronic device 1100 in the middleware 1132, the API 1133, or the application 1134.
The middleware 1132 can act as an intermediary for the API 1133 or the application 1134 to communicate with the kernel 1131 to exchange data. The middleware 1132 may also be coupled to the system resources (e.g., bus 1110) of the electronic device 1100 to at least one of the applications 1134, for example in connection with work requests received from the application 1134. [ (E.g., at least one of scheduling, load balancing, etc.) using a method such as assigning a priority that can be used (e.g., at least one of processor 1120, memory 1130, etc.) .
The API 1133 is an interface for the application 1134 to control the functions provided by the kernel 1131 or the middleware 1132 and may include at least one of file control, window control, image processing, Lt; / RTI > interface or function (e.
In various embodiments, the application 1134 may be a multimedia messaging service such as a Short Message Service (MMS) / Multimedia Messaging Service (MMS) application, an email application, a calendar application, an alarm application, a health care application An application that measures at least one), an environmental information application (e.g., an application that provides at least one of air pressure, humidity, temperature information, etc.), and the like.
Additionally or alternatively, the application 1134 may be an application related to the exchange of information between the electronic device 1100 and an external electronic device (e.g., electronic device 1200). Applications associated with information exchange may include, for example, a notification relay application for communicating specific information to an external electronic device, or a device management application for managing an external electronic device.
For example, the notification delivery application may send notification information from another application (e.g., an SMS / MMS application, an email application, a healthcare application, or an environmental information application) of the electronic device 1100 to an external electronic device 1200) or the like).
Additionally or alternatively, the notification delivery application may receive notification information from, for example, an external electronic device (e.g., electronic device 1200) and provide it to the user. The device management application may provide functionality (e.g., external electronic device itself (or some component) for at least a portion of an external electronic device (e.g., electronic device 1200, etc.) (Eg, at least one of turning on / off of the display, turning on / off of the display, adjusting the brightness (or resolution) of the display), managing an application running on the external electronic device, : At least one of installation, deletion, update, etc.).
In an embodiment, the application 1134 may include an application designated according to attributes (e.g., the type of electronic device, etc.) of an external electronic device (e.g., electronic device 1200). For example, if the external electronic device is an MP3 player, the application 1134 may include an application related to music playback.
Similarly, if the external electronic device is a mobile medical device, the application 1134 may include applications related to health care. In an embodiment, application 1134 may include at least one of an application specified in electronic device 1100 or an application received from an external electronic device (e.g., server 1300, electronic device 1200, etc.).
The input / output interface 1140 is configured to communicate commands or data input from a user via an input / output device (e.g., at least one of a sensor, a keyboard, a touch screen, etc.) via a bus 1110, The communication interface 1160, the additional function module 1170, and the like.
For example, the input / output interface 1140 may provide the processor 1120 with data about the user's touch input via the touch screen. The input / output interface 1140 is connected to the input / output interface 1140 via a bus 1110. The input / output interface 1140 is connected to the input / output interface 1140 via a bus 1110, Data can be output through an input / output device (e.g., at least one of a speaker, a display, etc.). For example, the input / output interface 1140 may output voice data processed through the processor 1120 to a user through a speaker.
Display 1150 can display various information (e.g., at least one of multimedia data, text data, etc.) to the user.
The communication interface 1160 can connect communications between the electronic device 1100 and an external electronic device (e.g., electronic device 1200, server 1300, etc.). For example, the communication interface 1160 may be connected to the network 1001 via wireless or wired communication to communicate with external electronic devices.
The wireless communication may include, for example, wireless fidelity (WiFi), WiFi direct, Bluetooth, near field communication (NFC), global positioning system (GPS), cellular communication (e.g., LTE, LTE- CDMA, WCDMA, UMTS, WiBro, or GSM), and the like. The wired communication may include at least one of, for example, a universal serial bus (USB), a high definition multimedia interface (HDMI), a recommended standard 232 (RS-232), a plain old telephone service (POTS)
In an embodiment, the network 1001 may be a telecommunications network. The communication network may include at least one of a computer network, an internet, an internet of things, a telephone network, and the like. (E.g., transport layer protocol, data link layer protocol, physical layer protocol, etc.) for communicating between the electronic device 1100 and the external electronic device may include an application 1134, an API 1133, a middleware 1132 ), A kernel 1131, a communication interface 1160, and the like.
In an embodiment, the add-on module 1170 can support the operation of the electronic device 1100 by performing at least one of the actions (or functions) implemented in the electronic device 1100. [ For example, the server 1300 may include an add-on server module 1320 that may support an add-on module 1170 implemented in the electronic device 1100. For example, the add-on server module 1320 may include at least one component of the add-on module 1170 to perform at least one of the operations performed by the add-on module 1170 (e.g., an agent, etc.) can do.
The additional function module 1170 processes at least some of the information obtained from other components (e.g., processor 1120, memory 1130, input / output interface 1140, communication interface 1160, etc.) And can be used in a variety of ways. For example, the add-on module 1170 may be configured to enable the electronic device 1100 to interact with other electronic devices (e.g., electronic device 1200, server 1300, etc.) using the processor 1120, At least some functions of the electronic device 1100 can be controlled. The add-on module 1170 may be integrated into the processor 1120 or the communication interface 1160. In an embodiment, at least one configuration of the additional function module 1170 may be included in the server 1300 (e.g., an additional function server module 1320) and may be provided from the server 1300 to the additional function module 1170 At least one operation to be implemented may be supported.
Meanwhile, the present invention can provide health services to various kinds of electronic devices.
17 is a view illustrating an exemplary health service providing system according to an embodiment of the present invention. Referring to FIG. 17, the health service providing system 2000 may include a plurality of terminal devices 2100 and 2200 and a health service providing server 2300.
Each of the plurality of terminal devices 2100 and 2200 may be implemented with the electronic device described in Figs.
The health service provision server 2300 may be configured to store time series big data for providing health services and to provide health services according to a request of at least one of the terminal devices 2100 and 2200.
Flowchart illustrations, or block diagrams in accordance with embodiments of the present invention, combinations thereof, may be implemented by computer program instructions. These instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus. When executed by a processor, instructions may generate means for implementing functions, activities or special events in each block of the diagrams or a combination thereof.
In this regard, each block or block diagram of the flow chart may correspond to a module, segment, or portion of code. Where a portion of the code may include one or more instructions for implementing a particular logic function. On the other hand, the function associated with any block may deviate from the order mentioned in the drawings. For example, although two blocks are shown in succession, they may actually be performed at the same time, or sometimes in reverse.
It should be understood by those skilled in the art that the technical spirit of the present invention can be implemented as an apparatus, system, method or computer program product. Thus, what is referred to as circuits, modules, configurations, or systems in accordance with the present invention may be implemented in hardware or implemented in software (including firmware, resident software, microcode) or a combination of hardware and software . The combination of hardware and software may include a computer program product embodied in a computer readable medium having computer readable program code.
The above-described contents of the present invention are only specific examples for carrying out the invention. The present invention will include not only concrete and practical means themselves, but also technical ideas which are abstract and conceptual ideas that can be utilized as future technologies.

100: Health care device
110: health information input unit
120: Health Information Preprocessing Department
130: Similar case search unit
140: Health Pattern Analysis and Future Health Prediction Department
150: Health Healing Planning Division
160: health healing information output unit
200: Health Big database
300: Health Group Database
400: Healing Knowledge Base
10: Health service delivery system

Claims (21)

For future health prediction methods of health care devices:
Receiving health information of a user including time series data;
Pre-processing the health information by normalizing the length of the time series data included in the health information;
Retrieving similar cases including time series data from the health database using the pre-processed health information;
Analyzing the health pattern of the user and predicting a future health measurement value based on regression analysis of the time series data included in the similar cases;
Designing a health healing plan corresponding to the analyzed health pattern or the predicted future health measurement; And
And outputting health healing information corresponding to the health healing plan to a display device,
The step of predicting the future health measurement comprises:
Performing similar case grouping corresponding to the health information of the user;
Analyzing the health factors associated with each group; And
Predicting the future health measurement using similar case group information and the health factors, and
Wherein proceeding with the similar case grouping comprises:
Performing a regression analysis operation on the time series data included in the similar cases;
Calculating first weights that are inversely proportional to differences between the regression analysis values and the actual values resulting from the regression analysis operation; And
And proceeding with the similar case grouping using the set of actual values of the similar cases and the set of the first weights. The method according to claim 1,
The health information includes at least one of sex, age, height, weight, waist circumference, systolic blood pressure, diastolic blood pressure, pre-eclampsia, total cholesterol, triglyceride, HDL cholesterol, LDL cholesterol, hemoglobin, urinary glucose, (Serum gioti) ALT, Gamma japitii, presence or absence of the past illness of the user, presence or absence of a disease patient in the user's family, smoking status, past smoking period, past day smoking amount, A method comprising at least one of the current smoking period, the current smoking amount, drinking habit, the amount of one drinking, the intense exercise days more than one week, the exercise days more than one week, the walking days more than 30 minutes, and the body mass . 3. The method of claim 2,
The step of retrieving the similar cases may include:
Calculating a second weight for each feature of the health information;
Grouping the users' health status into a similar group using the second weight;
Calculating a similarity between at least one user's health state of the grouped group and the user's health state; And
And retrieving the similar cases from the health database using the second weight and the similarity. 3. The method of claim 2,
The step of predicting the health measurement comprises:
Analyzing and predicting a health information change pattern into a healing knowledge base storing the similar cases; And
And performing matching and recognition of the user ' s health information with the similar cases. delete The method according to claim 1,
The above-mentioned health measurement is based on the following criteria: height, weight, waist circumference, systolic blood pressure, diastolic blood pressure, pre-eclampsia, total cholesterol, triglyceride, HDL cholesterol, LDL cholesterol, hemoglobin, urinary glucose, (Geothi) AST, (serum geothi) ALT, and gamma-lipoprotein,
The above health factors include sex, age, presence of previous disease, presence of family disease, smoking status, past smoking period, past day smoking amount, current smoking period, current smoking amount, drinking habit, Wherein the method comprises information relating to at least one of a number of days of intense exercise, a median exercise period of more than 30 minutes per week, a number of days of walking more than 30 minutes per week, and a body mass. delete The method according to claim 6,
Wherein the similar case grouping is a case in which a similar case is assigned to any one of the rising, maintaining, and falling groups. delete The method according to claim 6,
Wherein the similar case grouping is performed by calculating representative values of two or more clusters and dividing the similar cases into two or more cluster sets based on the representative values. The method according to claim 1,
The step of predicting the health measurement may further include calculating the importance of each of the health factors,
And combining the predicted health metric and the importance to calculate a final health metric predicted value. delete The method according to claim 1,
Wherein the health database is a personal health information case database. A similar case retrieval unit for retrieving similar cases including time series data corresponding to health information of a user including time series data from a health group database;
Analyzing a health pattern of the health information based on regression analysis of the time series data included in the similar cases, grouping the similar cases, analyzing health factors of the grouped similar cases, A health pattern analysis and future health prediction unit for predicting a future health measurement value corresponding to the health information using a factor;
A health healing planner for designing a health healing plan corresponding to the analyzed health pattern or the predicted health metric; And
And a display device for displaying the health healing plan,
The health pattern analysis and future health prediction unit includes:
Performing a regression analysis operation on the time series data included in the similar cases,
Calculating first weights that are inversely proportional to differences between the regression analysis values and the actual values resulting from the regression analysis operation, and
And groups the similar cases using a set of the actual values of the similar cases and a set of the first weights. 15. The method of claim 14,
A health information input unit receiving the health information; And
Further comprising a health information preprocessing unit for filtering the input health information, normalizing the filtered health information, and grouping the normalized health information. 15. The method of claim 14,
And the similar case searching unit searches the similar cases using the second weight and the similarity corresponding to the health information. delete 15. The method of claim 14,
Wherein the regression analysis operation is a linear regression analysis operation,
Wherein the grouping of the similar cases is determined to be one of an ascending group, a maintenance group, and a descending group. 15. The method of claim 14,
Wherein the health care apparatus is a wearable apparatus. Receiving health information of a user including time series data;
Pre-processing the health information by normalizing the time series data included in the health information;
Retrieving similar cases including time series data from the health database using the pre-processed health information;
Calculating first weighting factors inversely proportional to differences between regression analysis values and actual values according to a result of the regression analysis operation, performing a regression analysis operation on the time series data included in the similar cases, Grouping the similar cases using the set of actual values of the similar cases and the set of the first weights, analyzing the health factor of the grouped similar cases, and analyzing the health pattern of the user using the analyzed health factor Analyzing and predicting future health measurements;
Designing a health healing plan corresponding to the analyzed health pattern or the predicted health measurement; And
And outputting health healing information corresponding to the health healing plan to a display device. A computer-readable recording medium having recorded thereon a program for executing by a computer. 21. The method of claim 20,
Wherein the health pattern analysis and the future health metrics are predicted,
Grouping the similar cases of the person-public health database into a plurality of clusters;
Arranging the plurality of clusters in order of distance between time series; And
And using a center value for each of the sorted clusters as a prediction candidate value. ≪ Desc / Clms Page number 24 >

Images