KR102198454B1 - System and method for recommending health checkup items based on machine learning - Google Patents

Abstract

The present invention relates to a system and a method for recommending health examination items based on machine learning, wherein the present invention is configured to: create a checkup data set based on a preset checkup guide and clinical results; create an independent examination learning model for each examination item using the examination data set; determine whether or not examination is necessary for each examination item based on medical information of a requester when receiving the medical information of the requestor; and integrate the results of the determination of whether or not the examination is necessary, which is determined for each examination item to send a requester (customer or user) for medical examination.

Description

System and method for recommending health examination items based on machine learning {SYSTEM AND METHOD FOR RECOMMENDING HEALTH CHECKUP ITEMS BASED ON MACHINE LEARNING}

The following embodiments relate to a technology for recommending health examination items by using medical information of a medical examination requester.

With more than 15 million people conducting health check-ups annually in Korea, the medical check-up industry through hospitals or medical check-up agencies is gradually expanding. However, the probability of disease occurrence differs according to the individual's age, sex, test cycle, and risk factors (family history, past medical history, eating habits), and there are difficulties in selecting the screening items due to lack of clear awareness of the screening items. Therefore, there is an urgent need for a method that can minimize the damage of false positives or false negatives and maximize the effect of the health checkup through a health checkup optimized for individual characteristics.
As an example of a technology for managing a conventional health checkup, Korean Patent Publication No. 2019-0007978 "Smart Health Checkup Coordinating System" has been proposed.

delete

Korean Patent Publication No. 2019-0007978 (published on January 23, 2019)

An object of the present invention is to provide a system and method for recommending health examination items based on machine learning.

The present invention provides examination items (colorectal cancer, lung cancer, cerebral cardiovascular, etc.) according to an individual's sex, age, test cycle, lifestyle (drinking, smoking, etc.), family history and past medical history, and symptoms (brain, chest, stomach, etc.) ) Or, we recommend testing equipment. In particular, it aims to recommend screening items in real time by creating a multiple screening learning model based on domestic and foreign screening data in order to increase the effectiveness of screening recommendations.

According to an embodiment of the present invention, a system for recommending health examination items based on machine learning includes: a examination data generation device for generating a examination data set based on a predetermined examination guide and a clinical result; A screening learning model generating device for generating an independent screening training model for each screening item using the checkup data set; And a checkup item recommendation server that receives medical information for a checkup requester, determines whether a checkup is necessary for each checkup item based on the checkup requester medical information, and integrates and transmits a result of determining whether a checkup is necessary for each checkup item. .

In this case, the medical examination requester information may include at least one of questionnaire information, last examination result information of the examination requester, and treatment information of the examination requester.

In this case, the examination learning model generating apparatus includes: a examination data selection unit for each examination item for selecting a examination characteristic and a predetermined examination label set for each examination item from the examination data set; And a machine learning unit that generates an independent examination learning model for each examination item by machine learning the examination data selected for each examination item through the examination data selection unit for each examination item.

In this case, the examination item recommendation server may include: a examination feature generator for generating a examination characteristic from the medical information of the examination requester; A check-up feature selection unit for each check-up item for selecting a check-up feature required for each check-up item from the check-up features generated through the check-up requester's medical information; A checkup prediction unit for each checkup item to determine whether a checkup is necessary using a checkup feature selected for each checkup item from the checkup requester's medical information and a checkup learning model corresponding to the checkup item; And a checkup prediction result collection unit that integrates the determination result of whether or not the checkup is necessary for each checkup item and provides it to a checkup requester.

In this case, the examination data set may include examination features and examination labels.

In this case, the examination feature may be information referenced to determine whether a examination is necessary including at least one of age, sex, examination period, risk factor, and symptom.

In this case, the examination label may be information indicating examination details indicating whether or not examination for each examination item or detailed examination items.

In this case, the examination learning model generation apparatus trains the examination data set with at least one machine learning algorithm among a decision tree, a random forest, and a support vector machine, Each independent screening learning model can be created.

Generating a checkup data set based on a preset checkup guide and clinical results according to an embodiment of the present invention; Generating an independent examination learning model for each examination item by using the examination data set; Receiving medical information of a person requesting a checkup; Determining whether a medical examination is necessary for each examination item based on the medical information of the examination requester; And integrating and transmitting a result of determining whether the examination is necessary or not determined for each examination item.

In this case, the medical examination requester information may include at least one of questionnaire information, last examination result information of the examination requester, and treatment information of the examination requester.

In this case, generating an independent examination learning model for each examination item using the examination data set includes: selecting a examination characteristic and a predetermined examination label preset for each examination item from the examination data set; And generating an independent examination learning model for each examination item by machine learning the examination data selected for each examination item.

In this case, the step of determining whether a checkup is necessary for each checkup item based on the checkup requester medical information includes: generating a checkup characteristic from the checkup requester medical information; Selecting a required examination characteristic for each examination item from examination characteristics generated through the examination requester medical information; And determining whether a checkup is necessary using a checkup feature selected for each checkup item from the checkup requester medical information and a checkup learning model corresponding to each checkup item.

In this case, the examination data set may include a examination feature and a examination label.

In this case, the examination feature may be information referenced to determine whether a examination is necessary including at least one of age, sex, examination period, risk factor, and symptom.

In this case, the examination label may be information indicating examination details indicating whether or not examination for each examination item or detailed examination items.

In this case, the step of generating an independent examination learning model for each examination item using the examination data set includes at least one of a decision tree, a random forest, and a support vector machine. An independent examination learning model may be generated for each examination item by training the examination data set using a machine learning algorithm.

The present invention generates a checkup data set based on a preset checkup guide and clinical results, generates an independent checkup learning model for each checkup item using the checkup data set, and receives medical information of the checkup requester, It relates to a system and method for recommending health examination items based on machine learning that determines whether or not examination is necessary for each examination item, and sends the result of the determination of whether examination is necessary or not determined for each examination item to the examination requester (customer or user). In addition, by removing items with insufficient medical evidence and recommending items with a high probability of disease occurrence, there is an advantage of maximizing the effectiveness of health check-ups and reducing screening costs.

1 is a diagram showing a schematic configuration of a health examination item recommendation system according to an embodiment of the present invention.
2 is a diagram illustrating a configuration of an apparatus for generating a medical examination learning model in a health examination item recommendation system according to an exemplary embodiment of the present invention.
3 is a diagram illustrating a configuration of a checkup item recommendation server in a health checkup item recommendation system according to an embodiment of the present invention.
4 is a flowchart illustrating a process of recommending a health check-up item through a pre-processing process and a post-processing process in a health check-up item recommendation system according to an embodiment of the present invention.
5 is a flowchart illustrating a process of generating a medical examination learning model by the examination learning model generation apparatus of the health examination item recommendation system according to an embodiment of the present invention.
6 is a flowchart illustrating a process of determining whether or not a checkup is necessary in a checkup item recommendation server of a health checkup item recommendation system according to an embodiment of the present invention.
7 is a diagram illustrating an example of a checkup data set according to an embodiment of the present invention.

Hereinafter, exemplary embodiments will be described in detail with reference to the accompanying drawings. However, since various changes may be made to the embodiments, the scope of the rights of the patent application is not limited or limited by these embodiments. It should be understood that all changes, equivalents, or substitutes to the embodiments are included in the scope of the rights.

The terms used in the examples are used for illustrative purposes only and should not be interpreted as limiting. Singular expressions include plural expressions unless the context clearly indicates otherwise. In the present specification, terms such as "comprise" or "have" are intended to designate the presence of features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, but one or more other features. It is to be understood that the presence or addition of elements or numbers, steps, actions, components, parts, or combinations thereof, does not preclude in advance.

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 the embodiment belongs. Terms as defined in a commonly used dictionary should be interpreted as having a meaning consistent with the meaning in the context of the related technology, and should not be interpreted as an ideal or excessively formal meaning unless explicitly defined in this application. Does not.

In addition, in the description with reference to the accompanying drawings, the same reference numerals are assigned to the same components regardless of the reference numerals, and redundant descriptions thereof will be omitted. In describing the embodiments, when it is determined that a detailed description of related known technologies may unnecessarily obscure the subject matter of the embodiments, the detailed description thereof will be omitted.

In addition, in describing the constituent elements of the embodiment, terms such as first, second, A, B, (a) and (b) may be used. These terms are only used to distinguish the component from other components, and the nature, order, or order of the component is not limited by the term. When a component is described as being "connected", "coupled" or "connected" to another component, that component may be directly connected or connected to that other component, but another component between each component It should be understood that may be “connected”, “coupled” or “connected”.

Components included in one embodiment and components including common functions will be described using the same name in other embodiments. Unless otherwise stated, descriptions in one embodiment may be applied to other embodiments, and detailed descriptions in the overlapping range will be omitted.

Hereinafter, a system and method for recommending health examination items based on machine learning according to an embodiment of the present invention will be described in detail with reference to FIGS. 1 to 7.

1 is a diagram showing a schematic configuration of a health examination item recommendation system according to an embodiment of the present invention.

Referring to FIG. 1, the health examination item recommendation system may be largely composed of a pre-treatment system 110 and a post-treatment system 120. In this case, the preprocessing system 110 may be configured to include a screening data generating device 111 and a screening learning model generating device 113.

The examination data generating device 111 generates a examination data set 112 based on a predetermined examination guide and a clinical result. In this case, the clinical results may include previous examination results of examinees (customers) managed by the examination center operating the health examination item recommendation system. In other words, the examination data used in the examination learning model can be updated using the examination results of the managed customers.

The examination data set may include examination features and examination labels. At this time, the characteristics of the examination include age, gender, examination cycle, risk factors (family history, past medical history, eating habits, alcohol consumption, smoking, etc.) and symptoms (symptoms for each area such as brain, chest, stomach, etc.). This is information that is referenced to determine In addition, the examination label is information indicating examination details indicating the presence or absence of examination for each examination item, such as colon cancer, lung cancer, and cerebrovascular disease, or detailed examination items.

7 is a diagram illustrating an example of a checkup data set according to an embodiment of the present invention.

7 is an example of a data set relating to the presence or absence of a lung cancer test and the presence or absence of a cerebrovascular test. -There are lung cancer, BMI (body　mass　index), and the screening label includes lung cancer test items, brain cardiovascular test items, etc.

Referring to FIG. 7, the examination data may be expressed as a examination feature vector indicating an input attribute and a examination label vector indicating an output attribute.

　 = {x₁, x₂, …, x_m}와 같이 표기되며 n차원의 검진 라벨 벡터는

= {y₁, y₂, …, y_n}로 표기된다. The m-dimensional examination feature vector is

= {x ₁ , x ₂ ,… , x _m } and the n-dimensional examination label vector is

= {y ₁ , y ₂ ,… , y _n }.

In the example of FIG. 7, the examination feature vector indicating the examination characteristic has information on the customer and a health state, and may have a numerical value or a category value.

In addition, the examination label vector representing the examination label may have a binary classification such as whether examination is necessary (0/1) or a multiclass classification that is classified into three or more classes.

At this time, the multi-class may be used when detailed test items are included.For example, if there is a detailed cervical cytology or human papillomavirus test for cervical cancer test, the test result value 0 is no test required, 1 is It can be expressed as a Pap smear, 2 for HPV, and 3 for Pap smear and HPV.

Returning to the description of FIG. 1 again, the examination learning model generation device 113 generates independent examination learning models 130, 131, 132, and 133 for each examination item using the examination data set 112. In this case, a more detailed description of the screening learning model generating device 113 will be described later with reference to FIG. 2.

The post-processing system 120 may include a screening item recommendation server 121 and a terminal 122.

The terminal 122 receives medical information of a customer (user) of a medical examination requester and transmits it to the examination item recommendation server 121. In this case, the medical examination requester information may include at least one of interview information, information on a previous examination result of the examination requester, and treatment information of the examination requester.

The examination item recommendation server 121 receives the medical information of the customer through the terminal 122, determines whether or not examination is necessary for each examination item based on the medical information of the examination requester, and determines whether or not the examination determined for each examination item is necessary. The examination item recommendation list 124 generated by integrating the results is provided to the examination requester (customer or user) or the administrator (doctor, nurse, medical related person). In this case, a more detailed description of the examination item recommendation server 121 will be described later with reference to FIG. 3.

2 is a diagram illustrating a configuration of an apparatus for generating a medical examination learning model in a health examination item recommendation system according to an exemplary embodiment of the present invention.

Referring to FIG. 2, the apparatus 113 for generating a medical examination learning model may include examination data selection units 211, 212 and 213 and a machine learning unit 220.

The examination data selection units 211, 212, 213 select a predetermined examination characteristic and a predetermined examination label for each examination item assigned to each examination data set.

The machine learning unit 220 machine learns the examination data selected for each examination item (preset examination characteristics and data of a preset examination label) through the examination data selection units 211, 212, 213, and independent examination learning for each examination item. Models 131, 132, and 133 are created.

At this time, the machine learning unit 220 trains the examination data set with at least one machine learning algorithm from among a decision tree, a random forest, a support vector machine, etc. You can create an independent examination learning model.

In FIG. 2, the machine learning model for health checkup can be expressed as follows.

　 = {x₁, x₂, …, x_m}로 표현되고, n차원의 검진 라벨 벡터가

= {y₁, y₂, …, y_n}로 표현되고, 검진 특징 벡터

　의 부분집합

₁,

₂, …,

_n　(여기서

_i　⊆

, 1 ≤ i ≤n) 이 주어져 있을 때, 제i 학습 모델 f_i (1 ≤ i ≤n　)는 아래와 같이 표기할 수 있다.The m-dimensional examination feature vector

= {x ₁ , x ₂ ,… , x _m }, and the n-dimensional examination label vector

= {y ₁ , y ₂ ,… , y _n }, and the examination feature vector

Subset of

₁ ,

₂ ,… ,

_n (where

_i ⊆

, 1 ≤ i ≤ n), the i-th learning model f _i (1 ≤ i ≤ n) can be expressed as follows.

₁)로 표현되고, 제2 검진 학습 모델(132)은 y₂= f₂(

₂)로 표현되고, 제n 검진 학습 모델(133)은 y_n= f_n(

_n)로 표현될 수 있다. The first checkup learning model 131 is y ₁ = f ₁ (

₁ ), and the second examination learning model 132 is y ₂ = f ₂ (

₂ ), and the n-th examination learning model 133 is y _n = f _n (

It can be expressed as _n ).

_i, y_i}를 검진 데이터 집합(112)에서 추출하는 작업을 수행한다.The examination data selection unit 211, 212, 213 is the training data required for generating the examination learning model f _i in the machine learning unit 200 {

Extracting _i , y _i } from the examination data set 112 is performed.

_i　= {폐암 검사 주기, 금연기간, 연령, 흡연갑년}을 추출하는 작업을 수행할 수 있다. 건강 검진의 기계 학습을 위해서는 검진 특징의 개수(차원 수)를 줄이는 것이 중요한데, 본 발명의 검진 데이터와 같이 검진 특징의 수가 많고 한정된 검진 데이터 집합의 경우, 검진 특징의 상당수가 원하는 학습 결과의 생성에 유용하지 않으며 한정된 데이터 집합으로 인해 학습 알고리즘이 잡음에 과적합 될 수 있기 때문에 검진 데이터 선택부의 역할이 매우 중요하다. 이때, 검진 데이터 선택부(211, 212, 213)에서 추출하는 검진 특징은 사용자(관리자 또는 관련종사자 등)에 의해서 검사항목별로 사전에 기설정될 수 있다.For example, the screening data selection unit is related to the lung cancer test y _i to generate a lung cancer learning model f _i .

_i = {Lung cancer test cycle, smoking cessation period, age, smoking pack year} can be extracted. For machine learning of medical examination, it is important to reduce the number of examination features (the number of dimensions).In the case of a limited examination data set with a large number of examination features, like the examination data of the present invention, a significant number of examination features are required to generate desired learning results. The role of the examination data selector is very important because it is not useful and the learning algorithm may overfit to noise due to a limited data set. At this time, the examination features extracted from the examination data selection units 211, 212, 213 may be preset for each examination item by a user (administrator or related worker, etc.).

_i, y_i}를 이용하여 자동적으로 검진 항목별로 제i 검진 학습 모델(f_i)을 생성한다. The machine learning unit 220 includes examination data corresponding to the training data {

_{Using i} , y _i }, an _i th examination learning model (f _i ) is automatically generated for each examination item.

For example, a screening learning model for lung cancer test, a screening learning model for colon cancer test, and a screening learning model for stomach cancer test may be separately generated.

In the present invention, when new examination data or examination conditions are modified/added, a new learning model is created, and an _i- th examination learning model (f _i ) is generated for each examination label (y _i ), so that the execution time is small and an independent model is created. This is possible. As an example of the machine learning unit, supervised learning classification algorithms such as decision tree, random forest, and support vector machine are used, and the optimal learning model can be generated by changing the parameters of the algorithm. The final examination learning model can be saved in a single file format and used when recommending examination items.

3 is a diagram illustrating a configuration of a checkup item recommendation server in a health checkup item recommendation system according to an embodiment of the present invention.

Referring to FIG. 3, the examination item recommendation server 121 includes a examination feature generation unit 310, a examination characteristic selection unit 321, 322, 323, a examination prediction unit 331, 332, 333, and a examination prediction result collection unit. It may include 340.

The examination characteristic generation unit 310 generates examination characteristics from medical information of the examination requester.

The examination feature selection unit 321, 322, 323 selects a examination characteristic required for each examination item from examination characteristics generated through medical information of the examination requester.

The examination prediction units 331, 332, and 333 determine whether examination is necessary using examination characteristics selected for each examination item from the medical information of the examination requester and a examination learning model corresponding to each examination item.

The examination prediction result collection unit 340 integrates the determination result of whether the examination is necessary or not determined for each examination item and provides it to the customer (user).

Looking at the operation of the configuration of FIG. 3 with reference to the example of FIG. 7, the examination feature generation unit 310 may receive a result of an examination of the health status of the examination requester (customer), which is one of the medical information of the examination requester, from the customer. have. At this time, examples of questionnaire information are {"ID": 1, "Age": 43, "Gender":"M", "Height": 169, "Weight": 92, "Amount of smoking per day": 20, "Smoking period ":25,… } Etc. may be paperweight information.

= {y'₁, y'₂, …, y'_n}을 고객에게 전달한다. 검진 추천 리스트의 예로서 JSON 형태의 {"ID":1, "폐암검사":0, "간암검사":0, "대장암검사":0, "뇌심혈관검사":1, "자궁경부암검사":3, …} 로 표현할 수 있다. 검사 결과 값은 이진 데이터 혹은 세 개 이상의 클래스로 분류하는 다중 클래스를 가질 수도 있다.The examination characteristic generation unit 310 may derive examination characteristics from the examination information provided by the customer. For example, based on height/weight, BMI, daily smoking amount and smoking period are used to calculate the year of smoking. Meanwhile, the examination feature selection units 321, 322, and 323 extract examination features for each individual examination item. For example, in relation to lung cancer screening, features such as lung cancer test cycle, smoking cessation period, age, and smoking pack year are extracted. The examination prediction units 331, 332, and 333 predict whether a specific examination item needs examination based on examination characteristics corresponding to each examination item and examination learning model corresponding to each examination item. Finally, the examination prediction result collection unit 340 collects individual examination prediction results and recommends

_{= {Y '1, y'} 2, ... Delivers, y _'n} to the customer. As an example of a list of recommended examinations, {"ID":1, "Lung Cancer Test":0, "Liver Cancer Test":0, "Colon Cancer Test":0, "Cerebral Cardiovascular Test":1, "Cervical Cancer Test"":3,… It can be expressed as }. The test result value may have binary data or multiple classes classified into three or more classes.

Hereinafter, a method according to the present invention configured as described above will be described with reference to the drawings.

4 is a flowchart illustrating a process of recommending a health check-up item through a pre-processing process and a post-processing process in a health check-up item recommendation system according to an embodiment of the present invention.

Referring to FIG. 4, the health examination item recommendation system generates a examination data set based on a predetermined examination guide and clinical results (410). In this case, the clinical results may include previous examination results of examinees (customers) managed by the examination center operating the health examination item recommendation system. In other words, the examination data used in the examination learning model can be updated using the examination results of the managed customers. In addition, the examination data set may include examination features and examination labels. Here, the screening characteristics include age, sex, test cycle, risk factors (family history, past medical history, eating habits, alcohol consumption, smoking, etc.) and symptoms (symptoms for each area such as brain, chest, stomach, etc.) This is information that is referenced to determine In addition, the examination label is information indicating examination details indicating the presence or absence of examination for each examination item, such as colon cancer, lung cancer, and cerebrovascular disease, or detailed examination items.

Then, the health examination item recommendation system generates an independent examination learning model for each examination item using the examination data set (420). In this case, a more detailed description of step 420 will be described later with reference to FIG. 5.

In addition, when the health examination item recommendation system receives the medical information of the customer (user) of the medical examination requester (430), it determines whether examination is necessary for each examination item based on the medical information of the examination requester (440). A more detailed description of step 440 will be described later with reference to FIG. 6. In this case, the medical examination requester information may include at least one of interview information, information on a previous examination result of the examination requester, and treatment information of the examination requester.

In addition, the health examination item recommendation system integrates the determination result of whether the examination is necessary or not determined for each examination item and transmits it to the examination requester (customer or user) or the administrator (doctor, nurse, medical personnel) (450).

5 is a flowchart illustrating a process of generating a medical examination learning model by the examination learning model generation apparatus of the health examination item recommendation system according to an embodiment of the present invention.

Referring to FIG. 5, the screening learning model generation apparatus 113 selects a screening feature set and a preset test label for each test item from a test data set (510).

In addition, the examination learning model generation apparatus 113 generates an independent examination learning model for each examination item by machine learning the examination data selected for each examination item (520).

In step 520, the examination learning model generation device 113 trains the examination data set with at least one machine learning algorithm from among a decision tree, a random forest, and a support vector machine. An independent examination learning model can be created for each examination item.

6 is a flowchart illustrating a process of determining whether or not a checkup is necessary in a checkup item recommendation server of a health checkup item recommendation system according to an embodiment of the present invention.

Referring to FIG. 6, the examination item recommendation server 121 generates a examination characteristic from medical information of a examination requester (610).

Then, the examination item recommendation server 121 selects a necessary examination characteristic for each examination item from the examination characteristic generated through the examination requestor medical information (620).

In addition, the examination item recommendation server 121 determines whether a examination is necessary using a examination feature selected for each examination item from the medical information of the examination requester and a examination learning model corresponding to each examination item (630).

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

The software may include a computer program, code, instructions, or a combination of one or more of these, configuring the processing unit to behave as desired or processed independently or collectively. You can command the device. Software and/or data may be interpreted by a processing device or to provide instructions or data to a processing device, of any type of machine, component, physical device, virtual equipment, computer storage medium or device. , Or may be permanently or temporarily embodyed in a transmitted signal wave. The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

As described above, although the embodiments have been described by the limited drawings, a person of ordinary skill in the art can apply various technical modifications and variations based on the above. For example, the described techniques are performed in a different order from the described method, and/or components such as a system, structure, device, circuit, etc. described are combined or combined in a form different from the described method, Alternatively, even if substituted or substituted by an equivalent, an appropriate result can be achieved.

Therefore, other implementations, other embodiments and claims and equivalents fall within the scope of the following claims.

110: pretreatment system
111: examination data generating device
112: examination data set
113: checkup learning model generation device
120: post-treatment system
121: checkup item recommendation server
122: terminal
123: Medical information of the requester
124: list of recommendations for examination items
130: screening learning model
131: first examination learning model
132: second examination learning model
133: nth examination learning model
[One]

Claims (17)

A checkup data generation device that generates a checkup data set based on a preset checkup guide and clinical results;
A screening learning model generating device for generating an independent screening training model for each screening item using the checkup data set; And
A checkup item recommendation server that receives medical information of a checkup requester, determines whether a checkup is necessary for each checkup item based on the checkup requester medical information, and integrates and transmits the result of the determination of whether a checkup is necessary for each checkup item
Including,
The examination learning model generation device,
A checkup data selection unit for each checkup item for selecting a checkup characteristic and a preset checkup label for each checkup item from the checkup data set; And
A machine learning unit that generates an independent examination learning model for each examination item by machine learning the examination data selected for each examination item through the examination data selection unit for each examination item
Including,
The examination item recommendation server,
A checkup feature generator for generating a checkup feature from the medical information of the checkup requester;
A check-up feature selection unit for each check-up item for selecting a check-up feature required for each check-up item from the check-up features generated through the check-up requester's medical information;
A checkup prediction unit for each checkup item to determine whether a checkup is necessary using a checkup feature selected for each checkup item from the checkup requester's medical information and a checkup learning model corresponding to the checkup item; And
A checkup prediction result collection unit that integrates the determination result of whether or not the checkup is necessary for each checkup item and provides it to a checkup requester
Including
A system for recommending health screening items based on machine learning.
The method of claim 1,
The medical information of the person requesting the examination,
Including at least one of questionnaire information, past examination result information of the examination requester, and treatment information of the examination requester
Health examination item recommendation system.
delete delete The method of claim 1,
The examination data set,
Including examination features and examination labels
A system for recommending health screening items based on machine learning.
The method of claim 5,
The above examination features,
Information that is referenced to determine whether a medical examination is necessary including at least one of age, sex, examination frequency, risk factors, and symptoms.
A system for recommending health screening items based on machine learning.
The method of claim 5,
The examination label,
Information indicating whether or not a medical examination is performed for each examination item or detailed examination items
A system for recommending health screening items based on machine learning.
The method of claim 1,
The medical examination learning model generation device,
Train the examination data set with at least one machine learning algorithm among a decision tree, a random forest, and a support vector machine to generate an independent examination learning model for each examination item.
A system for recommending health screening items based on machine learning.
Generating a checkup data set based on a preset checkup guide and clinical results in a health checkup item recommendation system;
Generating an independent examination learning model for each examination item by using the examination data set in the health examination item recommendation system;
Receiving medical information for a medical examination requester from the health examination item recommendation system;
Determining, in the health examination item recommendation system, whether examination is necessary for each examination item based on the medical information of the examination requester; And
Integrating and transmitting a determination result of whether a medical examination is necessary or not determined for each examination item in the health examination item recommendation system
Including,
In the health examination item recommendation system, generating an independent examination learning model for each examination item using the examination data set,
Selecting a predetermined examination characteristic and a predetermined examination label for each examination item from the examination data set through each examination data selection unit for each examination item included in the health examination item recommendation system; And
Machine learning the examination data selected for each examination item to generate an independent examination learning model for each examination item
Including,
In the health checkup item recommendation system, the step of determining whether a checkup is necessary for each checkup item through each checkup feature selection unit for each checkup item based on the medical information of the checkup requester,
Generating a checkup feature from the checkup requester medical information;
Selecting, by each examination item-specific examination feature selection unit included in the health examination item recommendation system, a required examination characteristic for each examination item from among examination characteristics generated through the medical information of the examination requester; And
Determining whether a checkup is necessary using a checkup characteristic selected for each checkup item and a checkup learning model corresponding to each of the checkup items by each checkup prediction unit for each checkup item included in the health checkup item recommendation system
Including
A method of recommending health screening items based on machine learning in the health screening item recommendation system.
The method of claim 9,
The medical information of the person requesting the examination,
Including at least one of questionnaire information, past examination result information of the examination requester, and treatment information of the examination requester
A method of recommending health screening items based on machine learning in the health screening item recommendation system.
delete delete The method of claim 9,
The examination data set,
Including examination features and examination labels
A method of recommending health screening items based on machine learning in the health screening item recommendation system.
The method of claim 13,
The above examination features,
Information that is referenced to determine whether a medical examination is necessary including at least one of age, sex, examination frequency, risk factors, and symptoms.
A method of recommending health screening items based on machine learning in the health screening item recommendation system.
The method of claim 13,
The examination label,
Information indicating whether or not a medical examination is performed for each examination item or detailed examination items
A method of recommending health screening items based on machine learning in the health screening item recommendation system.
The method of claim 9,
In the health examination item recommendation system, generating an independent examination learning model for each examination item using the examination data set,
Train the examination data set with at least one machine learning algorithm among a decision tree, a random forest, and a support vector machine to generate an independent examination learning model for each examination item.
A method of recommending health screening items based on machine learning in the health screening item recommendation system.
A computer-readable recording medium, wherein a program for executing the method of any one of claims 9, 10, and 13 to 16 is recorded.

Images