KR102162522B1 - Apparatus and method for providing personalized medication information - Google Patents

Abstract

The present disclosure relates to an apparatus and method for providing customized medical information based on user location information, by profiling user location information and attribute information, and processing medical information collected by crawling using an artificial intelligence learning model. By generating medical factor sets and medical institution information, it is possible to provide a medical information service with improved user convenience and option.

Description

Apparatus and method for providing customized medical information {APPARATUS AND METHOD FOR PROVIDING PERSONALIZED MEDICATION INFORMATION}

The present invention relates to an apparatus and method for providing customized medical information based on the user's location information, and more specifically, by profiling the identified user's location information, creating user life pattern information and user profile, and determining health based thereon. It relates to an apparatus and method for providing prediction and information of medical institutions.

The present invention also relates to an apparatus and method for recommending a medical institution and medical staff suitable for a user based on a user profile, subjective symptoms, and environmental information, and a method and system for generating customized medical institution and medical staff information using an artificial intelligence learning model. It is about.

Currently, medical services are becoming more specialized, subdivided, and complex, but patients still have to solve the acquisition of information for health management, self-measurement, diagnosis, treatment, etc., and selection of medical institutions and medical staff. In other words, it is common for patients to infer related diseases or current health conditions from symptoms they are aware of, and to take appropriate self-measures or to select medical institutions and medical staff by themselves or by advice of acquaintances. Related information can be collected through website or SNS information search, etc. to identify symptoms or health conditions, infer disease, self-measure, health care, medical institution or medical staff selection. However, such information collection takes a lot of effort, and the information that can be grasped from subjective symptoms is limited, there are limitations in the investigation and judgment of non-expert patients, and there are many cases of misjudgement. Therefore, individuals are likely to go through trial and error in order to determine their health status and to receive self-measures or appropriate medical services based on them, and symptoms may worsen or the treatment period may be missed during this process. As such, grasping and predicting an individual's health status, primary identification of necessary measures (including proactive self-measures), and appropriate medical measures (including matching appropriate medical institutions and medical staff) are urgent issues for individuals.

Conventional inventions also disclose technologies such as providing (remote) medical treatment by acquiring vital signs and symptoms such as body temperature and heart rate of a user measured through a user terminal in order to determine the health status of an individual. It is more of a means of assisting medical staff as part of a hospital's diagnostic system, rather than providing information and giving freedom of choice. Therefore, the system for appropriate self-measurement and patient's right to choose a medical institution in the pre-hospital selection stage has not been activated.

Meanwhile, a conventional hospital search apparatus such as Korean Patent Publication No. 10-1398383 is an invention for searching for a hospital suitable for a patient's condition, and provides a list of suitable hospitals in consideration of the patient's location, medical history, and biometric information. Since this is to extract a suitable hospital based on the patient's past medical history and biometric information at the time of treatment, it is possible to provide some meaningful medical institution matching information with reference to the patient's disease history or conventional health status. But. Since there is no consideration of the patient's accumulated lifestyle, current health status, and symptoms, and only the past hospital diagnosis history is considered, it is difficult to recommend a medical institution that meets the occurrence of a new disease and the current needs of the patient.

Furthermore, hospital information used for matching patient needs and medical institutions in the prior invention is mainly provided from medical institutions such as hospitals, and is not only extremely subjective, fragmentary, and limited, but also detailed information such as objective evaluation or more in-depth information, especially patient reviews. It is not effective because it is not reflected, and it is not possible to adequately respond to various needs of patients.

Registered Patent Publication No. 10-1398383

An object of the present invention is to solve the above problems and provide an apparatus and method for providing health information and medical service provider information for a user, thereby expanding the user's freedom of selection of medical services.

The present invention creates a user's life pattern based on the user's location information, predicts the trend of the generated life pattern information, and provides health prediction information of the future user, so that the user can manage health before the onset of disease. do.

The present invention also provides customized medical information based on multi-faceted feedback and information on various medical institutions and medical staff, including patient reviews, as well as information provided by medical institutions, thereby improving the quality of medical information provision services and managing personal health. And the convenience of using medical services.

In order to solve the above problems, a method for providing customized medical information performed by a medical information providing device according to an embodiment of the present invention uses an artificial intelligence life pattern learning model based on location information received from the user device. Generating pattern information; Generating a user profile based on the user life pattern information and user attribute information; Generating a set of medical factors including region, age, sex, time, feature facts, and disease using the artificial intelligence first learning model based on the collected medical information; Generating user health information for determining or predicting a health state of a user based on the user profile and the plurality of medical factor sets; And providing the generated user health information to a user device.

The artificial intelligence life pattern learning model is obtained by supervised machine learning using location information received from a user device, and the user attribute information includes the user's age, gender, height, weight, and blood type. The life pattern information includes a user's time-specific activity, a movement path or location, and a movement method, and the user profile includes the user's age, gender, region, exercise time, sleep time, and periodic activity details. The characteristic fact is a fact related to the disease, and preferably includes at least one of symptoms, exercise time, and sleep time.

According to an embodiment of the present invention, a method for providing user-customized medical information performed by a medical information providing device is environmental information including temperature, humidity, fine dust concentration or weather information at a user location from the user device or an external data source. Obtaining a; In the step of generating the user profile, it is preferable to generate the user profile by adding environment information to the user life pattern information and user attribute information.

In the generating of the medical factor set, a phrase related to a medical factor is extracted from the medical information collected by the crawler by a natural language processing method, and a plurality of medical factors are performed using an artificial intelligence first learning model based on the extracted phrase. It may be a step of creating a set. The first artificial intelligence learning model may be obtained by performing machine learning based on a phrase including generation time, age, sex, exercise time, and related diseases among data collected by a crawler.

According to another embodiment of the present invention, a method for providing customized medical information performed by an apparatus for providing medical information includes: receiving a subjective symptom from the user device; Updating the user profile based on the subjective symptom; And generating medical service provider information including a user-customized medical institution and medical staff suitable for a user by using a third artificial intelligence learning model based on the user profile and the user health information. And providing the medical service provider information according to the ranking along with the user health information to a user device.

Based on the external information collected by the crawler, the medical institution information and the medical staff information are generated using an artificial intelligence second learning model obtained by machine learning the medical institution and medical staff information collected from external data sources including patient reviews, The artificial intelligence third learning model may be obtained by machine learning using the user health information, user profile, medical institution information, and medical staff information.

According to another embodiment of the present invention, a method for providing customized medical information may include generating a question based on the different location information when receiving location information different from the generated life pattern; Transmitting the question to the user terminal; Receiving answer information for the question information from the user terminal; And updating the user's life pattern based on the answer information.

Acquiring user characteristic information including at least one of information on a user's psychological state, drinking state, meal information, allergy information, and family history along with the subjective symptoms from the user device; And updating the user profile based on the user characteristic information.

The apparatus for providing customized medical information according to the present invention includes a processor and a storage unit for storing computer program instructions executed by the processor. The instructions include program codes that operate to execute the steps of the method. In addition, a computer program recorded on a computer-readable storage medium may be implemented as a computer program including instructions for performing the method.

Using the user's life pattern information, user-customized health information can be created, and the life pattern information can be modified and updated using the question-and-answer information with the user terminal, creating advanced user life patterns and user profiles. Can be obtained.

The present invention may generate a user life pattern based on the user's location information, and provide health prediction information of a future user based on the trend of the accumulated life pattern information.

The present invention also provides customized medical information based on a variety of medical-related information including patient reviews collected in real time, thereby improving the quality of medical information provision service and enhancing the convenience of personal health management and medical service use.

1 is a schematic diagram of a user-customized medical information providing system according to an embodiment of the present invention, and is a diagram schematically showing a relationship between a customized medical information providing device, a user device, a medical institution server, and an input terminal.
2 is a diagram schematically illustrating a method of providing medical information in a system for providing user-customized medical information according to an embodiment of the present invention.
3 is a schematic diagram of an architecture related to data processing of a medical information providing apparatus according to an embodiment of the present invention.
4 is a diagram showing an architecture of a medical information providing apparatus according to an embodiment of the present invention, centering on an artificial intelligence learning model for a medical institution.
5 is a flowchart of generating user life pattern information by processing user location data by the management unit 107 of the medical information providing apparatus according to an embodiment of the present invention.
6 is a flowchart illustrating a method of generating a question according to user location information and processing a user answer after generating a user's life pattern according to an embodiment of the present invention.
7 is a block diagram illustrating a process of collecting and processing data by a crawler according to an embodiment of the present invention.
8 is a diagram showing an example of collecting health-related articles from newspapers as raw data by a crawler.
FIG. 9 is a diagram showing another example of collecting health-related articles from newspapers as raw data by a crawler, and is a diagram showing a configuration in which a preprocessor extracts filtering data from article raw data and searches for a target to be crawled at the same time .
FIG. 10 is a diagram showing a new article retrieved by the target search of FIG. 9 and data extracted from this article by a preprocessor.
FIG. 11 is a diagram illustrating structured data generated by an artificial intelligence learning model from raw data of FIGS. 8 and 10 through a preprocessor in a table and a health prediction method.
12 is a diagram illustrating a health prediction method related to medical factor sets according to an embodiment of the present invention.
13 is a table showing an example of a medical factor set according to another embodiment of the present invention.
14 is a diagram showing a method of preprocessing medical data according to another embodiment of the present invention.
15 and 16 are diagrams showing, in a table, a health determination according to an exemplary embodiment of the present invention, and a method of generating and providing information about a medical institution and a medical staff accordingly.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art can easily implement the present invention. However, the present invention may be implemented in various different forms and is not limited to the embodiments described herein. In the drawings, parts irrelevant to the description are omitted in order to clearly describe the present invention, and similar reference numerals are assigned to similar parts throughout the specification.

Throughout the specification, when a part is said to be "connected" with another part, this includes not only "directly connected" but also "electrically connected" with another element interposed therebetween. . In addition, when a part "includes" a certain component, it means that other components may be further included rather than excluding other components unless specifically stated to the contrary.

Hereinafter, embodiments of the present invention will be described in detail.

1 is a diagram schematically showing a relationship between a customized medical information providing apparatus 100 and a user apparatus 200, an external medical data source, and an input terminal of a user-customized medical information providing system according to an embodiment of the present invention. FIG. 2 is a schematic diagram illustrating a method of providing customized medical information according to an embodiment of the present invention in terms of data transmission between a medical information providing device and a user device and external data sources.

A system for providing user-customized medical information according to an embodiment of the present invention will be described with reference to FIGS. 1 and 2. A medical information providing device including user location information and user attribute information from the user device 200 through a communication network When transmitted to 100, the medical information providing device 100 generates a user profile by profiling the user information, and generates determination or prediction information of the user's health condition based on the user profile and provides it to the user device. . In general, profiling is an automated process that collects information about an individual or group of individuals, analyzes their characteristics and behavior, and makes predictions and evaluations. The user profile is a data set generated through profiling based on user information, and includes data such as a user's age, gender, exercise amount, sleep time, region, labor amount, walking time, and sitting time.

According to another embodiment of the present invention, the apparatus 100 for providing medical information transmits a question to the user device based on the user information, and updates a user profile using the answer and the user information.

On the other hand, in order to extract a set of medical factors that serve as a standard for determining and predicting a user's health status, the medical information providing apparatus 100 collects medical and medical-related structured and unstructured data, and machine learning using the collected data. A first learning completion model (structurer AI in FIG. 3) that extracts the quantified medical factors through is acquired. The apparatus 100 for providing medical information generates health information for determining or predicting a user's health status based on the user profile and the plurality of medical factor sets, and provides it to the user device.

According to another embodiment of the present invention, when a user inputs a symptom into the user device 100 and requests appropriate hospital information, the medical information providing device 100 is based on the input symptom and the user profile. Is determined and the appropriate medical service provider, i.e. medical institution and medical staff information is generated and transmitted to the user device.

The user device 200 is an information processing device possessed by a user and includes a processor, a memory, a communication module, a GPS module providing location information, an acceleration sensor capable of acquiring the user's exercise information, a 3-axis sensor, a gyro sensor, and the like. can do. The user device is, for example, a smart phone, a wearable device such as a smart watch, a tablet PC, and the like, and any information processing device capable of communicating with the outside and detecting a location is not limited by the type.

The apparatus 100 for providing medical information of the present invention may be implemented as a single server or a distributed data processing system including one or more networks. When implemented as a single server, the medical information providing apparatus 100 is a computer on which computer-readable program code or process implementation instructions can be executed, and includes a processor, a storage unit (memory/permanent storage), a communication unit, and an input/output unit. . A processor is used to execute instructions for software that can be loaded into memory, and may be a set of one or more processors or a multiprocessor core (depending on the specific implementation). In addition, the processor may be implemented using one or more disparate processor systems in which the main processor is on a single chip, and as another example, it may be a symmetric multiple processor (SMP) system including several processors of the same type.

The storage may include, for example, memory and persistent storage, the memory may be random access memory or other suitable volatile or nonvolatile storage device, and the persistent storage may take various forms depending on the particular implementation. For example, the permanent storage may be a hard drive, an SSD, a flash memory, a rewritable optical disk, a rewritable magnetic tape, or a combination of the above, and may be a removable hard drive. The communication unit provides communication with other data processing systems or devices, and may be, for example, a network interface card, and may provide communication through the use of wired, wireless communication links, or both. The input/output unit includes various input/output devices such as a display, a touch panel display, a keyboard, and a printer.

Hereinafter, an embodiment of the present invention will be described in more detail with reference to FIGS. 3 to 5. 3 is an architectural diagram illustrating the system of FIGS. 1 and 2 in terms of operation and functionality. The apparatus 100 for providing customized medical information according to an embodiment of the present invention includes a management unit 107 that collects and processes data from the user device 100 and works with the user device, and a database 101 that stores filtered data. , A crawler for crawling data from an external data source, a database 104 storing raw data collected from an external data source, a pre-processor 109 that preprocesses raw data from the database 104 ), a medical information processing engine 110, a structured database 102, a prediction database 103, and a post-processing unit 121.

The management unit 107 processes user information collected from the user device 200 to generate a user profile and stores it in the database 101. The medical information processing engine 110 generates user health information that determines or predicts a user's health status based on the user profile and stores it in the prediction database 103, and the user health information is passed through the post-processing unit 121. It is sent to the user device. The management unit 107 may be implemented as, for example, a management application program loaded into a memory and executed by a processor. In addition, to the user device, user information including user identification information, location information of the user device, and user attribute information is transmitted to the customized medical information providing device 100 in connection with the management application program, and processed information from the medical information providing device It is preferable that a mobile application program to be provided is installed. The mobile application program also interlocks with the management application program to provide a GUI displaying a specific question received from the management application program on the user device or display a text message, and recall the user's response input by a method such as a touch. Transfer to the management department.

The mobile application program executed by the processor of the user device transmits the user location information to the management unit 107 of the medical information providing device 100 in the following manner.

1. Receives user location information from the GPS module of the user device at regular time intervals, and transmits a log including information of {time, location} to the management unit 107 in real time (real-time location transmission mode), or accumulated logs ( log) in a package at a fixed time interval.

2. Analyzes the user log and transmits the log to the management unit in real time from 7 am to 9 am and 6 pm to 8 pm Monday to Friday estimated as the moving time zone.

3. When the position change detected by GPS is more than a certain speed, for example, 4 km/h, it is determined that it is moving using a vehicle or other means of transportation, and automatically converts to a real-time location transmission mode to manage location information in real time. Transfer to (107).

4. Analyzes the user log and stores the log as a package at a time when it is determined that there will be no movement, for example, Monday to Friday, 9 to 14:00, which are business hours, and transmits it in bulk later. However, in this case, if the moving distance or moving speed is less than a certain standard, for example, if it is 0.1 km/h, the location data is not transmitted, and the log is refined to store the stop time and stop position ('no movement') data. Send the log.

The user-customized medical information providing device 100 processes user location information received from the user device 200 to generate user life pattern information (time, activity, movement method, location, etc.). That is, the management unit 107 of the medical information providing apparatus 100 learns the time and location information of the received log or log package using the artificial intelligence learning model to obtain an artificial intelligence life pattern learning model, and uses the learning model. To generate user life pattern information. However, since this learning is by online learning, it is possible to continue learning even after the initial setting, so that decision-making and learning can be made at the same time.

The learning of the life pattern learning model is by supervised learning, and the location information of the user U1 by time of a specific user device Up1={{time1, location1}, {time2, location2}, ... } It learns by using the log including the data of the user's activity and movement method, but you can refer to the following.

-Location information displayed on weekends, holidays and sleeping hours is recognized as the user's home

-If location information on the move is collected during rush hour, it is recognized as being on the way

-If it appears as fixed location information during working hours, it is identified as the location of the workplace

-If the moving location information appears at the time of work and there is no moving location information afterwards, it is recognized as the user's home

-Comparing the commuting route with the route of public transportation provided by public APIs, it is possible to understand the route of movement, but whether the user's means of transportation is public transportation, whether it is a vehicle, bicycle, motorcycle, etc.

-Measure the moving speed according to the change in location to determine walking, jogging, bicycle, motorcycle, car, railroad, etc., but determine the means of transportation by referring to the public transportation route (for example, walking is 1 to 4 km/h). ; Jogging is 4-20 km/h, bicycle 20-40 km/h, transportation 40 km/h or more)

Using the learned AI life pattern learning model, the user's Up1={{time1, location1}, {time2, location2},… Ua1={{time=8 o'clock ~ 8:20, activity=coming, walking, movement=Mapo-gu oo-ro oo → Mapo-gu 00 station}, {time=8:20~9 City, activity=coming to work, public transportation, traffic line=oo-ro oo in Mapo-gu → Samseong-ro 00 station}, {time=9:00~18:30, activity=work, no movement, location= Samseong-ro 00 building},… , {Time=22:30 ~ 21:00, activity=coming work, car, traffic line= Samseong-ro 00 building → Mapo-gu oo-ro oo}, {time=22:00 ~ 22:40, activity=exercise, Jogging, movement = 00-ro, Mapo-gu → oo-ro, Mapo-gu oo},… } Is created. For example, a standard machine learning algorithm can be applied to statistically learn the weights of the life pattern algorithm. Examples of suitable machine learning in the implementation of the invention include supervised machine learning algorithms such as SVM (Support Vector Machine), neural network learning, or decision tree learning algorithm. The life pattern Uai (i = user 1, user 2,…) is basically generated on a daily basis, but may be generated on a weekly or monthly basis.

According to another embodiment of the present invention, a user's exercise method, exercise time, amount of exercise, walking time, sitting time, and the like may be calculated using a motion measurement sensor including an acceleration sensor of the user device. This is due to various conventional motion tracking methods and can be used by a person of ordinary skill in the art without difficulty. The result values (exercise amount, exercise time, exercise method) of another application program using sensor data are transmitted to the management unit by interlocking with the mobile application of the user device, or the management unit 107 collects sensor measurement data of the user device. You can also analyze it yourself.

In addition, the management unit 107 may continuously receive the user location information and update the user life pattern even after generating the user life pattern based on the user location information from a specific user device. In particular, when receiving location information different from the previously created lifestyle pattern (eg, moving to a location different from the existing movement route at 7 pm on Wednesday), a question may be generated based on the different location information. For example, when the user moves to a place other than the one recognized as home at 7 pm on Wednesday, it asks for the purpose of the move and marks it. In this case, the management unit 107 asks a categorized question whether the purpose of the movement is an activity such as a drinking party or drinking, watching a performance such as a play or a movie, or an exercise such as fitness, swimming pool, or yoga. Create a question and send this question to the user's device to provide a question to the user. The mobile application program of the user's device provides a GUI (Graphic User Interface) for asking the user the above question and inputting an answer, and the user selects a category (dinner, drink, watch a performance, exercise, etc.) and/or selects specific activity details. You can enter. The activity input as the answer is added to the user's life pattern information generated by the management unit, and is stored as data for learning the artificial intelligence life pattern learning model, and is also used for artificial intelligence learning.

The management unit 107 creates a user profile based on the lifestyle pattern and user attribute of a specific user. The user profile is a data set, which includes the user's age, gender, region, exercise time, exercise amount, sleep time, activity history, and includes working hours, walking time, sitting time, alcohol consumption, leisure time, etc. It may further include data. Age and gender are based on the user attribute, and are based on values (birth date, gender) entered by the user when registering as a member according to a form requested by the application program of the user device. The exercise time, exercise amount, sleep time, periodic activity details, working time, walking time, sitting time, drinking amount, leisure activity time, etc. are generated by the management unit from life pattern information. The data set of the user profile is required to correspond to the factors of the medical factor set. Accordingly, when a new factor is added, the medical information processing engine 110 requests the management unit 107 to add new data to the profile, and accordingly, may request an additional factor for generating life pattern information.

In addition, the user device may be a variety of wearable devices or a smart device interworking with the wearable device, and the wearable device may include a sensor that senses a heartbeat and a body temperature of the wearer. In this case, the user device may transmit biometric signals such as the user's heart rate, body temperature, blood sugar, and blood pressure to the management unit, and the management unit may process the biosignal and add it to the user profile.

Meanwhile, the medical information processing engine 110 of the medical information providing apparatus 100 generates user-customized health information and medical service provider information based on a user profile stored in the database 101 and provides it to the post-processing unit 121, The post-processing unit transmits the health information and medical service provider information to the user device. Hereinafter, the medical information processing engine 110 obtains a medical factor structured learning model (artificial intelligence first learning model) and a predictive learning model through machine learning using data collected from an external data source, and user-customized health information And a method of generating medical service provider information will be described in detail.

First, the medical information processing engine uses medical information announced by public institutions, Health Insurance Review and Assessment Service, universities, hospitals, research institutions, etc. as standard data, but crawling can be applied to collect additional data. The reference data may be structured or unstructured data.

For example, from the Health Insurance Review and Assessment Service, the standard data of "health condition standard (reference data): 40s / male / exercise 3 times a week / health for 30 minutes or more each time / drinking once a week" can be received. The reference data is stored in a database (not shown) and then processed into meta data to be generated as a medical factor set and stored in the structured database 102.

Meanwhile, various atypical medical-related data such as newspaper articles, blogs, SNS, and news are collected by the crawler, stored in the raw data database 104, and then processed by the preprocessor 109 (see FIG. 3). 7 is a flow chart of data processing in the crawler and raw data database 104. That is, the medical information providing apparatus 100 of the present invention collects various medical and medical service related information from hospitals, medical institutions, research institutes, papers, news channels, media, blogs, SNS, public institutions, medical related sites, etc. by crawling. Then, through machine learning using the medical data and medical service-related data, an artificial intelligence first learning model (structurer AI of FIG. 3) that extracts quantified medical factors is obtained. For example, data displayed in tables, diagrams, and graphics are also refined and processed in the same way.

8 to 10 exemplarily show a method of collecting and pre-processing atypical medical data. Referring to FIG. 8, when a newspaper company website is added to the crawl list, the robot manager collects raw data of related articles in the medical, social, and health aspects of the website and stores it in the database 104. The preprocessor extracts only phrases (sentences enclosed in red) containing medical data through natural language analysis of the raw data. Referring to FIG. 9, the preprocessor extracts a medical-related phrase from another article, and detects a research institution, department, doctor,'investigation','announcement', etc. from this phrase and performs additional search or crawling. FIG. 10 is a web page searched or tracked by a hyperlink from the syntax of FIG. 9, and the preprocessor extracts the obesity-related phrases of the elderly aged 60 or older. Data extracted by the preprocessor 109 is stored in the database 101 and processed by the medical information processing engine 110.

The medical information processing engine 110 includes an artificial intelligence first learning model (Structurer AI) 111, a health prediction unit 112, an artificial intelligence second learning model 113, and a medical institution matching unit 114. The medical information processing engine 110 acquires an artificial intelligence first learning model that establishes a medical factor set by learning medical and medical-related data stored in the database 101, and uses this to obtain a medical factor set from the collected medical data. Create. Since this learning is by online learning, continuous learning can be performed even after the initial setting, so that the medical factor set generation and learning can be performed at the same time.

The table of FIG. 11 shows examples of a set of medical factors generated using the artificial intelligence first learning model. For example, the medical information processing engine uses the first artificial intelligence learning model based on the pre-processed article in FIG. 8, and MC1 = {Time = January 2017, Region = Seoul/Gyeonggi, Host Organization = Gangbuk Samsung Hospital , Gender=male, age=40s, sample=160,000, average=43%, disease=fat liver, daily exercise time=30 minutes}. However, here, the exercise time per day = 30 minutes is not based on the article of FIG. 8, but is based on the information of the user's life pattern and user profile. That is, the average value of exercise time in the region = Seoul/Gyeonggi, gender = male, age = 40s collected from a plurality of user devices is added. That is, the first artificial intelligence learning model learns to estimate the missing factor value using not only medical information, but also user life pattern information and statistical values of the user profile. As such, some unstructured data or medical factors that cannot be derived from structured data can be obtained through the learning process of the first artificial intelligence learning model. Alternatively, it is possible to obtain the first artificial intelligence learning model by using various estimated values such as statistics collected from the user device or the location of the host institution. The plurality of medical factor sets generated above are stored in the structured database 102. The medical factor set may include information such as time, region, sex, age, characteristic facts (symptoms, fine dust concentration) related to a specific disease, disease, treatment, amount of exercise, walking time, and a host organ. Characteristic facts may include environmental information such as average fine dust concentration in a specific area, accident or natural disaster information such as volcanic eruptions in a specific area or at a specific time. For example, if you are dealing with medical news collected by crawling, for example, news that the rate of cancer patients is high in a specific area, news that the groundwater in a specific area is contaminated with radioactive materials, the weight of a specific area will increase. I can. In this case, news data may be added for learning of the first artificial intelligence learning model, and a weight may be assigned to a region value.

By tracking the past history of public data based on the creation time and region of the above raw data, environmental information (fine dust concentration, C02 concentration, climate, accidents such as human resources, heat waves, cold waves, natural disasters such as typhoons, humidity, etc.) It can further bind to the set of medical factors. More specifically, the medical information processing engine collects fine dust concentration, CO2 concentration, weather, and climate information provided to the public API based on the creation time and area data from the medical factor set generated by the artificial intelligence first learning model. It can be added to the set of medical factors. For accidents or natural disasters, you can request the crawler to collect accident information from URLs such as news and media websites based on time and region. The collected data can be added to the set of medical factors through a process similar to medical data.

Meanwhile, the prediction unit 112 generates health determination and prediction information of the user based on the user profile and a plurality of medical factor sets generated through the artificial intelligence first learning model. The prediction unit 112 may be implemented as a logic algorithm including calculating a degree of matching between the user profile and a plurality of medical factor sets. However, when the user profile is accumulated, the collected medical data is vast, and the medical factor set generated therefrom is also large, the prediction unit is preferably implemented as an artificial intelligence learning model that generates a matching degree. That is, the artificial intelligence health prediction learning model is acquired by machine learning based on a user profile and a set of medical factors, and is advanced by continuous learning.

When the matching degree between the user profile and the medical factor set is generated using the artificial intelligence health prediction learning model, health determination or prediction information is generated according to the ranking of the matching degree, and the health determination and the user device through the post-processor 121 Transmit prediction information.

The medical information providing apparatus 100 is an artificial intelligence first learning completion model (structurer AI in FIG. 3) that collects medical and medical-related structured and unstructured data, and extracts quantified medical factors through machine learning using the collected data. To obtain. The apparatus 100 for providing medical information generates and provides the user's health information to the user device based on the user profile and the plurality of medical factor sets.

The health determination information is, for example, that the user profile of a specific user is {40s / male / walking twice a week / 30 minutes / moving by car when moving to commute / drinking more than 3 times a week}, and the matching degree is the best among the set of medical factors. If the factor set is {40s / male / exercise 3 times a week / health for 30 minutes or more each time / drinking once a week}, then {lack of exercise * / obesity caution ***** / alcoholism recommendation *****} Health information may be provided. In addition, it is possible to analyze the trend of changes in the user's health status by using information on the user's lifestyle accumulated over time, and through this, a health warning that quantifies the possibility of obesity or fatty liver after 6 months or 1 year is previewed. It can predict and provide it to the user device. More specifically, the management unit 107 generates a future user life pattern in advance from the accumulated user life pattern information, reflects it in the creation of a user profile at a specific time in the future, and uses the user profile at a specific time in the future to provide medical information. The processing engine 110 generates health prediction information.

Furthermore, when a user medical history is input through an input terminal or a user medical history is collected through a hospital or an insurance server, the medical information providing apparatus may add this medical history to the user profile. It is possible to determine or predict more advanced health status by using the user profile to which the user's medical history has been added.

According to still another embodiment of the present invention, an apparatus and method for determining a health status based on a user profile and recommending a medical institution and medical staff suitable therefor are provided. As described above, the user-customized medical information providing device generates user life pattern information and user profile from user information, and the health status of the user based on the set of medical factors and the user profile generated using the first artificial intelligence learning model. Is determined.

Meanwhile, the health information providing apparatus 100 of the present invention may collect structured data provided by an institution, etc. to store and process data related to medical institutions and medical staff. However, since this information is fragmentary and limited, it is limited to meet the diverse needs of patients. Therefore, according to an embodiment of the present invention, in addition to the collection of the structured data, the crawler collects various types of data including information such as patient reviews, medical staff experiences, papers, medical equipment, etc., and the medical information processing engine ( 110) acquires the second artificial intelligence learning model 113 through machine learning using this.

The raw data related to medical institutions and medical staff collected by the crawler are stored in the database 101 through a pre-processing process of extracting only the natural language processing and related phrases in the same manner as medical data. A medical institution and a medical staff data set in which the second artificial intelligence learning model is structured are generated through machine learning based on phrases including medical thesis and patient reviews. For example, medical institution information of a hospital for a specific disease A HA1={location=00, Ilwon-dong, Gangnam-gu, Seoul; Medical department=00; Related medical equipment=PET CT; Medical staff=00, 00, 00; Number of annual treatment patients=000; Overall review score=4.52; Average patient age=40s; Preferred gender=male; Service satisfaction=4.12; Cure rate=45.3%; Special feature = mental and rehabilitation service; Key keyword for patient review = fast, accurate, OO; ??}, Medical Staff Information on Disease A of a Specific Doctor DA1={Location= 00, Ilwon-dong, Gangnam-gu, Seoul; Medical department=00; Medical equipment=PETRON; Affiliated institution = 00 hospital; Number of annual treatment patients=000; Overall review score=4.72; Average patient age=40s; Gender=male; Service satisfaction=4.42; Cure rate=55.3%; Specifics = 00 surgery; Key keyword for patient review=accurate, kind; ??}to be. These items are exemplary and can be added as needed.

For the medical institution information and medical staff information, information on medical institutions announced by public institutions, Health Insurance Review and Assessment Service, universities, hospitals, research institutions, etc., and data collected through crawling are additionally added, and the second artificial intelligence learning model is used. And stored in the medical institution database 105.

Meanwhile, the medical information processing engine 110 further includes a matching unit 114, and the matching unit extracts medical institution information and medical staff information suitable therefor based on the determination of the health status and the user profile. Medical institution information on a specific disease or health condition is a data set, location, department, medical equipment, medical staff, number of patients processed per year, overall review score, average age of patients, preferred gender, service satisfaction, cure rate, special details, patient review core Include keywords. Medical staff information is a data set of a medical person such as a specific doctor or physical therapist. Location, medical department, disease, affiliated institution (hospital), number of patients treated per year, overall review score, average age of patients, gender, service satisfaction, cure rate, special details, Includes key keywords for patient reviews and treatment schedule. Based on the user profile and at least one health determination information generated by the prediction unit 112 based on the user profile, the matching unit 114 compares the medical institution information and the medical staff information stored in the medical institution database 105 Matching medical institutions and medical staff information are generated in the order of high matching. The matching degree is calculated by giving weight to a determined health state, for example, a specific disease, but symptoms, age, region, sex, exercise amount, etc. included in the user profile can also be considered, and the weight applied accordingly is artificial. It is determined through intelligence learning. That is, the matching unit is an artificial intelligence third learning model, which is obtained through machine learning using medical institution information, medical staff information, and user profile.

15 and 16 show a health state determined based on a profile of a specific user, that is, a predicted disease and matching medical institutions and medical staff suitable therefor. In other words, it is shown in the table that matching medical institution and medical staff information is generated according to the ranking of the expected disease and by appropriate ranking according to the matching degree.

A brief description will be given of a process for receiving a medical information providing service from the side of a user device according to another embodiment of the present invention. The mobile application program installed on the user device transmits a log or log package of user location information to the medical information providing device, receives user-customized health determination or prediction information from the medical information providing device, and sends the health information to the display or audio output unit of the user device. It provides decision or prediction information. At this time, user-customized health information is provided according to the ranking.

When a user feels a specific subjective symptom and wants to obtain hospital information, the user can input the symptom using the GUI of the mobile application program. The input symptoms are transmitted to a medical information providing device along with other user information, added to a user profile, and a user-customized health determination and medical institution and medical information are generated and transmitted to the user device.

According to another embodiment of the present invention, the collection priority for collecting medical institution and medical staff information by crawling is selected by the following method. The medical information processing engine 110 classifies a distance based on the location information received from the user device as an origin, and determines the degree of overlap using the location information of multiple users as the origin. The distance may be, for example, 1Km or 5Km. The above-mentioned medical institutions with a high degree of overlap are preferentially targeted for crawling, and information is collected, and this crawling must be continued even after the initial setting, and even in this case, the location information from multiple user devices is given priority to medical institutions with high degree of overlap as the origin. Data is collected by ranking and the process is repeated.

A plurality of health factor sets are generated using the artificial intelligence first learning completion model, and user-customized health status determination or health prediction information is provided to the user.

Meanwhile, programs stored in the memory of the apparatus 100 for providing medical information of the present invention may be classified into a plurality of modules according to their functions, wherein the plurality of modules are software, not hardware, and functionally operated modules. It means, and may be classified into, for example, a UI module and a touch screen module.

Some embodiments may also be implemented in the form of a recording medium including instructions executable by a computer, such as a program module executed by a computer. Computer-readable media can be any available media that can be accessed by a computer, and includes both volatile and nonvolatile media, removable and non-removable media. Further, the computer-readable medium may include a computer storage medium. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

In addition, in the present specification, the “unit” may be a hardware component such as a processor or a circuit, and/or a software component executed by a hardware configuration such as a processor.

The above description of the present invention is for illustrative purposes only, and those of ordinary skill in the art to which the present invention pertains will be able to understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not limiting. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as being distributed may also be implemented in a combined form.

The scope of the present invention is indicated by the claims to be described later rather than the detailed description, and all changes or modified forms derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention. do.

100: medical information providing device 200: user device
101: database 102: structured database
103: prediction database 104: raw data database
105: medical institution database 107: administration
110: medical information processing engine 111: artificial intelligence first learning model
112: prediction unit 113: artificial intelligence second learning model
114: matching unit

Claims (8)

As a method of providing user-customized medical information performed by a medical information providing device,
Generating user life pattern information using an artificial intelligence life pattern learning model based on the location information received from the user device;
Generating a user profile based on the user life pattern information and user attribute information;
Generating a set of medical factors including region, age, sex, time, feature facts, and disease using the artificial intelligence first learning model based on the collected medical information;
Generating user health information for determining or predicting a health state of a user based on the user profile and the plurality of medical factor sets; And
Including; providing the generated user health information to the user device,
The artificial intelligence life pattern learning model is obtained by supervised machine learning using location information received from a user device,
The user attribute information includes the user's age and gender, and the user life pattern information includes the user's time-based activity, movement path or location, and the user profile includes the user's age, gender, region, exercise time, and sleep. It includes time and a history of regular activities, and the characteristic fact is a fact related to the disease, and includes at least one of symptoms, exercise time, and sleep time,
Obtaining environment information including temperature, humidity, fine dust concentration or weather information at a user location from the user device or an external data source; further comprising,
In the step of generating the user profile, the user profile is generated by adding environment information to the user life pattern information and user attribute information. The method of claim 1,
Receiving subjective symptoms from a user device; And
Updating the user profile based on the subjective symptoms; further comprising, a method for providing user-customized medical information. As a method of providing user-customized medical information performed by a medical information providing device,
Generating user life pattern information using an artificial intelligence life pattern learning model based on the location information received from the user device;
Generating a user profile based on the user life pattern information and user attribute information;
Generating a set of medical factors including region, age, sex, time, feature facts, and disease using the artificial intelligence first learning model based on the collected medical information;
Generating user health information for determining or predicting a health state of a user based on the user profile and the plurality of medical factor sets; And
Including; providing the generated user health information to the user device,
The artificial intelligence life pattern learning model is obtained by supervised machine learning using location information received from a user device,
The user attribute information includes the user's age and gender, and the user life pattern information includes the user's time-based activity, movement path or location, and the user profile includes the user's age, gender, region, exercise time, and sleep. It includes time and a history of regular activities, and the characteristic fact is a fact related to the disease, and includes at least one of symptoms, exercise time, and sleep time,
In the step of generating the medical factor set, a phrase related to the medical factor is extracted from the medical information collected by the crawler by a natural language processing method, and a plurality of medical factor sets are performed using an artificial intelligence first learning model based on the extracted phrase. Method for providing customized medical information, characterized in that the step of generating. The method of claim 1,
Generating medical service provider information including user-customized medical institutions and medical staff based on the user profile and the user health information; And
Providing the medical service provider information according to the ranking along with the user health information to a user device; further comprising, a method for providing customized medical information for a user. As a method of providing user-customized medical information performed by a medical information providing device,
Generating user life pattern information using an artificial intelligence life pattern learning model based on the location information received from the user device;
Generating a user profile based on the user life pattern information and user attribute information;
Generating a set of medical factors including region, age, sex, time, feature facts, and disease using the artificial intelligence first learning model based on the collected medical information;
Generating user health information for determining or predicting a health state of a user based on the user profile and the plurality of medical factor sets; And
Including; providing the generated user health information to the user device,
The artificial intelligence life pattern learning model is obtained by supervised machine learning using location information received from a user device,
The user attribute information includes the user's age and gender, and the user life pattern information includes the user's time-based activity, movement path or location, and the user profile includes the user's age, gender, region, exercise time, and sleep. It includes time and a history of regular activities, and the characteristic fact is a fact related to the disease, and includes at least one of symptoms, exercise time, and sleep time,
Generating a question based on the different location information when receiving location information different from the generated user life pattern information;
Transmitting the question to the user device;
Receiving answer information for the question information from the user device; And
Updating the user life pattern information based on the answer information; further comprising, a method for providing customized medical information for a user. The method according to claim 1 or 2,
Acquiring user characteristic information including at least one of information on a user's psychological state, drinking state, meal information, allergy information, and family history, along with subjective symptoms from the user device; And
The method of providing customized medical information, further comprising updating the user profile based on the user characteristic information. As a user-customized medical information providing device,
Processor,
And a storage unit for storing computer program instructions executed by the processor,
The instructions include program codes that operate to execute the steps of the method of claim 1. A computer program recorded in a computer-readable storage medium, comprising instructions for performing the method of any one of claims 1 to 5.

Images