KR20220029972A - Support recipient health care product recommendation system using artificial intelligence and method thereof - Google Patents

Abstract

The present invention relates to a system and method for recommending health care products for a support target, and more specifically, to a system and method for recommending a health care products for a support target using artificial intelligence, which learns by applying a learning dataset mapping expert recommendation information including health care products information according to medical treatment history by experts and health care products information collected through online crawling to an artificial intelligence model, and applies user information including disease information to the learned artificial intelligence model to recommend optimal health care products to patients with disease information.

Description

Support recipient health care product recommendation system using artificial intelligence and method thereof

The present invention relates to a support target health care product recommendation system and method, and more particularly, expert recommendation information including health care product information according to the treatment history by experts and health care product information collected through online crawling. Artificial intelligence that learns by applying the mapped learning dataset to the artificial intelligence model, and recommends optimal health care products to patients with disease information by applying user information including disease information to the learned artificial intelligence model It relates to a system and method for recommending health care products for support target using

As cloud computing and big data environments that can access the Internet network anytime, anywhere are established, product recommendation systems that recommend products that are optimal to users in various environments based on big data have been developed and applied.

However, most of these conventional product recommendation systems are recommendations for the general public and general products, and a health care product recommendation system that recommends health care products related to diseases including diseases and disabilities to support recipients who have diseases or disabilities has not been developed. is not

The reason why a disease-related health care product recommendation system is not developed is that the distribution structure of health care products such as assistive devices for the disabled and welfare tools for the elderly in Korea is an incompletely competitive market structure by subsidies, and the This is because it is not easy to recommend because it is mainly done through door-to-door sales, which limits consumer choice.

In addition, since it targets users in special circumstances, that is, patients with diseases such as the disabled, it is difficult to access information on professional knowledge and user reviews on health care products. because it is getting worse.

As described above, it is difficult to access the distribution structure of health care products, expert knowledge on health care products, and user evaluations, it is difficult to map professional knowledge and health care products, and as information asymmetry in the market deepens, diseases There was a problem in that it was difficult to develop a system for recommending health care products for support recipients with this condition.

Therefore, in the distribution structure of health care products, it is possible to map professional knowledge on health care products and health care products, and to recommend health care products in consideration of the product evaluation of previous users who have used health care products. development is required.

Republic of Korea Patent Publication No. 10-2013-0054231 (published on May 14, 2013)

Therefore, it is an object of the present invention to learn by applying a learning dataset mapping expert recommendation information including health care product information according to medical treatment history by experts and health care product information collected through online crawling to an artificial intelligence model, and In providing a system and method for recommending health care products for support subjects using artificial intelligence that recommends optimal health care products to patients with diseases of disease information by applying user information including disease information to the learned artificial intelligence model. there is.

A health care product recommendation system for support target using artificial intelligence according to the present invention for achieving the above object is: connected to the Internet network through a wired/wireless data communication network to perform data communication with devices of the Internet network; a communication unit for performing data communication with the expert recommendation information providing unit and the support target terminal unit through the wired/wireless data communication network; A product DB that stores health care product information including health care product names for a number of health care products, and an expert who stores expert recommendation information including multiple treatment histories and recommended health care product information for each treatment specified by experts a storage unit including a recommendation DB; and collecting the expert recommendation information from the expert recommendation information providing unit through the communication unit, storing it in the expert recommendation DB, and crawling and collecting information on health care products online through the communication unit, collecting and analyzing the name of the health care product After generating and storing health care product information including After learning by applying the learning dataset to the model, the user information of the support target input from the support target terminal unit through the communication unit is applied to the health care product recommendation model to provide a recommended health care product corresponding to the user information After extraction, it is characterized in that it comprises a control unit for transmitting to the support target terminal unit.

The control unit collects the expert recommendation information from the expert recommendation information providing unit through the communication unit, stores it in the expert recommendation DB, and crawls and collects information about health care products online through the communication unit, collects and analyzes health After creating health care product information including the health care product name for the managed product and storing it in the product DB, the health care product information of the product DB and the expert recommendation DB expert with the health care product name of the health care product information as the key an information collection unit for generating a learning dataset by mapping the recommendation information, and obtaining user information of the support target inputted in real time from the support target terminal unit through the communication unit; a health care product recommendation model learning unit for learning by applying the learning dataset to a preset health care product recommendation model; And after extracting a health care product corresponding to the user information by applying the user information, health care product recommendation for transmitting health care product recommendation information including health care product information on the extracted health care product to the support target terminal unit It is characterized by including wealth.

The information collection unit crawls and collects product information on health care products online through the communication unit, analyzes it, and generates health care product information including the health care product name, manufacturer and price information, and stores it in the product DB a product information collection unit that collects the expert recommendation information from the expert recommendation information providing unit through the communication unit and stores the expert recommendation information collection unit in the expert recommendation DB, and the product using the health care product name of the health care product information as a key a learning dataset collecting unit including a learning dataset generating unit that creates a learning dataset by mapping the health care product information of the DB and the expert recommendation information of the expert recommendation DB; and a user information collection unit configured to acquire user information of a target subject to be supported in real time, input from the terminal unit to the support target through the communication unit.

The learning data set collection unit includes an evaluation information collection unit that collects evaluation information from a cloud sourcing unit that collects evaluation information on the health care product from users including support subjects who have used a plurality of different health care products. Further comprising, wherein the learning dataset generating unit is characterized in that the mapping of the health care product name to a key to generate a learning dataset further including the evaluation information.

The storage unit further includes a supplier DB for storing seller registration information including supplier (seller) information and health care product sales information including the price of one or more health care products sold by the supplier of the supplier information, wherein the The controller receives, from the health care product supply unit through the communication unit, supplier information for selling at least one health care product and seller registration information including the health care product sales information, stores it in the supplier DB, and registers it. Further comprising a registration unit, wherein the health care product recommendation unit searches the supplier DB for a supplier that supplies a health care product under the health care product name of the health care product information, and provides supplier information and price information for supplying the health care product It is characterized in that it is further included in the health care product recommendation information.

The health care products are health care products including poultice products, posture correction equipment, exercise equipment, eye health maintenance equipment, oxygen supply device, massager, protector, and seat training device, electric bed, pressure sore prevention article, non-slip member, loitering sensor, safety loss Welfare products including teeth, mobile toilets, walkers, walking sticks, vision aids, prosthetic/prosthetic hands, communication aids, posture aids, driving aids, wheelchairs, assistive devices including hearing aids, and hospital treatment devices, hospital measuring devices, and hospital holding devices , It is characterized in that it includes a medical device including a home measuring instrument, a home washing machine, and home herbal products.

The health care product information is characterized in that it includes a product name (health care product name), a model name (unique number), a manufacturer, a purchase price, an option, category classification information, and a release date.

The expert recommendation information includes gender code, age code, province/region (region) code, medical treatment start date, form code, treatment subject code, sick and injured code, number of days in the hospital, total number of prescription days, and health care product information. characterized by including.

In the method for recommending health care products for support subjects using artificial intelligence according to the present invention for achieving the above object, the control unit collects the expert recommendation information from the expert recommendation information providing unit through the communication unit and stores it in the expert recommendation DB, , After crawling and collecting information on health care products online through the communication unit, and analyzing it to create health care product information including the health care product name for the health care product and store it in the product DB, the health care product an information collection process of creating a learning dataset by mapping the information of the product DB and the expert recommendation information of the expert recommendation DB using the information; a learning process in which the controller applies the learning dataset to a preset health care product recommendation model through a health care product recommendation model learning unit to learn; and a user information acquisition process in which the control unit acquires user information of the support target inputted in real time from the support target terminal unit through the communication unit; and after the control unit extracts a health care product corresponding to the user information by applying the user information through the health care product recommendation unit, supports health care product recommendation information including health care product information on the extracted health care product It is characterized in that it includes a health care product recommendation process transmitted to the target terminal.

In the information collection process, the control unit crawls and collects product information on health care products online through the communication unit, analyzes it, and generates health care product information including the health care product name, manufacturer, and price information. Product information collection step of storing in the product DB; an expert recommendation information collecting step of the controller collecting the expert recommendation information from the expert recommendation information providing unit through the communication unit and storing the expert recommendation information in the expert recommendation DB; and a learning dataset creation step in which the controller maps the information of the product DB and the expert recommendation information of the expert recommendation DB to the health care product information as a key to generate a training dataset.

The information collection process includes an evaluation information collection step of collecting the evaluation information from the cloud sourcing unit that collects evaluation information on the health care product from users including the support target who has used a plurality of different health care products. Furthermore, the step of generating the learning dataset is characterized in that by mapping the health care product name to a key, the learning dataset further including the evaluation information is generated.

The method includes: the control unit receives seller registration information including supplier information for selling at least one health care product and the health care product sales information from the health care product supply unit through the communication unit and stores it in the supplier DB The method further comprising: a seller registration process to register, wherein the health care product recommendation process searches the supplier DB for a supplier that supplies a health care product under the health care product name of the health care product information, and supplies the health care product It is characterized in that supplier information and price information are further included in the health care product recommendation information.

The health care product information is characterized in that it includes a product name (health care product name), a model name (unique number), a manufacturer, a purchase price, an option, category classification information, and a release date.

The expert recommendation information includes gender code, age code, province/region (region) code, medical treatment start date, form code, treatment subject code, sick and injured code, number of days in the hospital, total number of prescription days, and health care product information. characterized by including.

The present invention collects expert recommendation information including a plurality of treatment details and health care product information corresponding to each treatment record, and crawls and analyzes health care product information for health care products sold or introduced on the online Internet. By constructing a learning dataset by creating health care product information and mapping it with expert recommendation information, it is possible to solve the problem of information asymmetry, and it has the effect of recommending optimal health care products to disease patients who are the target of support. .

In addition, the present invention collects product information and evaluation information on the health care product from previous users including support subjects who have used the health care product first, and further includes it in the learning dataset to learn the evaluation and It has the effect of being able to recommend the optimal health care product to the target recipient by reflecting the national support information.

In addition, the present invention has the effect of broadening the choice of support target by providing information including a plurality of sellers who provide and sell recommended health care products, and the price and function of each health care product sold by each seller. there is.

1 is a diagram showing the configuration of a communication system including a health care product recommendation system for support target using artificial intelligence according to the present invention.
2 is a diagram showing the configuration of a health care product recommendation system for support target using artificial intelligence according to the present invention.
3 is a flowchart illustrating a method for recommending health care products for support subjects using artificial intelligence according to the present invention.
4 is a flowchart illustrating a method of collecting information among methods of recommending health care products for support target using artificial intelligence according to the present invention.

Hereinafter, with reference to the accompanying drawings, the configuration and operation of a health care product recommendation system for support target using artificial intelligence according to the present invention will be described in detail, and a method for recommending health care products for support target in the system will be described.

1 is a diagram showing the configuration of a communication system including a health care product recommendation system for support target using artificial intelligence according to the present invention.

Referring to FIG. 1 , the support target health care product recommendation system 300 using artificial intelligence according to the present invention is connected to the support target terminal unit 100 and the expert recommendation information providing unit 200 through the wired/wireless data communication network 1 . to perform data communication according to the present invention, and to perform data communication by being connected to the cloud sourcing unit 400 and the health care product supply unit 500 according to an embodiment.

The wired and wireless data communication network 1 is an Internet network including a Wi-Fi network and a local area network (LAN), a mobile communication network such as 3G, 4G, 5G, a WiBro network, and electric power. It is a data communication network in which any one or more of a wire network and a dedicated network are combined, and a plurality of various devices are connected to the wired/wireless data communication network 1 by either wired or wireless to perform data communication with each other.

The support target terminal unit 100 is a terminal unit of the support target who needs health care products such as the disabled, the elderly, etc. according to the present invention or receives support from the state. , a personal computer, a computer terminal 120 such as a notebook computer, and the like.

The support target terminal unit 100 receives user information from the support target or transmits pre-stored user information to the health care product recommendation system 300 , and receives and displays health care product recommendation information corresponding to the user information.

The health care products may be classified into health products, welfare products, auxiliary devices, and medical devices. The health products may be poultice products, posture correction products, exercise equipment, eye health products, oxygen supply devices, accessories, massagers, protectors, seat training devices, and the like. The welfare product may be an electric bed, a pressure sore prevention product, an anti-slip product, a roaming sensor, a safety handle, a mobile toilet, a walker, a cane, and the like. In addition, the assistive device may be a vision aid, a prosthetic/prosthetic device, a communication aid, a posture aid, a driving aid, a wheelchair, a hearing aid, and the like. Medical equipment is classified into hospital use and home use, and hospital use includes a treatment device, a measuring instrument, a maintenance device, and the like, and there may be a measuring instrument, a washing machine, oriental medicine supplies for home use.

The user information may include information such as age, gender, disability information, disease information, and region.

The expert recommendation information providing unit 200 collects treatment details according to arbitrary diseases and diseases from a plurality of hospitals and doctors, and expert recommendation information to which information on recommended or recommended health care products required or recommended in response to each treatment history is mapped is generated and stored at a predetermined time interval, that is, the second cycle, and provided to the health care product recommendation system 300 of the present invention. The recommended health care product information may be included in the medical treatment history, but the recommended health care product information necessary for the missing medical treatment history may be inputted from a medical professional and registered.

The expert recommendation information providing unit 200 may be a server of the National Health Insurance Corporation that simply provides only treatment details, and in this case, the health care product recommendation system 300 receives treatment received from the expert recommendation information providing unit 200 . It may be configured to receive and register a corresponding health care product from a medical professional for each treatment detail in the history DB.

The health care product recommendation system 300 receives expert recommendation information including treatment details or treatment details and recommended health care product information from the expert recommendation information providing unit 200, and in the former case, recommended health care products for treatment details Information is input from a medical expert, expert recommendation information is generated, and the database is stored, and in the latter case, the received expert recommendation information is converted into a database and stored. The expert recommendation information may be collected in units of a second cycle, and the second cycle may be 1 year, 1 month, or the like.

The medical history may include information such as Table 1 below.

number Item name explanation One gender code 1 (male), 2 (female) 2 age code group by age 5 3 try code State code of residence 4 Start date of treatment Year, month, day of visit 5 format code Classification of types of treatment for patients treated at medical and health institutions 6 Medical course code 26 types of medical course codes 7 main casualty code Korean standard disease and cause of death classification 8 wounded code additional corporal 9 Number of days in hospital day of hospitalization 10 Total number of days of prescription When a prescription is issued

In addition, the health care product recommendation system 300 collects health care product information by performing crawling on health care products among the information existing in the Internet network through the wired/wireless data communication network 1, and includes necessary information elements. Health care product information is created and stored as a database. The health care product information may include information as shown in Table 2 below.

number Item name explanation One product name Product name 2 model name unique number 3 manufacturing company In the case of imported goods, the name of the importer 4 purchase price retail price 5 option Add column by product 6 Main Category Category classification 7 middle class 8 subcategory 9 Release date point of sale

When the collection of the information is completed, the health care product recommendation system 300 maps health care product information and expert recommendation information to the product name of the health care product information as a key to create a learning dataset, and is applied to a preset health care product recommendation model. After application and learning, the health care product recommendation model is applied to the learned health care product recommendation model to extract the health care product corresponding to the user information by applying the user information of the support target input from the support target terminal unit, and then generate health care product recommendation information to support the above It is transmitted to the target terminal unit 100 .

The cloud sourcing unit 400 recruits a plurality of support recipients through any one or more of online and offline, and allows the recruited support recipients to input evaluation information about the health care products they have used through online and offline. The evaluation information for each of a plurality of support targets is collected, and the collected evaluation information is provided to the health care product recommendation system 300 .

The evaluation information may be configured in the form shown in Table 3 below.

number Item name explanation One gender code 1 (male), 2 (female) 2 age code group by age 5 3 try code City and province code of residence (local information) 4 model name Health care product name (unique number) 5 purchase price product purchase price 6 date of purchase When to buy thread 7 subsidy Government subsidy amount 8 grade out of 5 9 type of disability Determination of disability 10 disease, disease disease, in the presence of disease

In Table 3, the model name is shown as a health care product name, but the model name and the health care product name may be separately included.

The health care product supply unit 500 may be a terminal of a seller (or “supplier”) who sells health care products or a server of a sales company, and includes supplier information, health care product name, selling price, performance, etc. The seller registration information is transmitted to the health care product recommendation system 300 to perform supplier registration.

When the cloud sourcing unit 400 is configured, the health care product recommendation system 300 further configures the evaluation information received from the cloud sourcing unit 400 in a learning dataset, and then applies it to the health care product recommendation model to learn carry out

In addition, when the health care product supply unit 500 is configured, the health care product recommendation system 300 searches for a supplier that supplies, that is, sells, a recommended health care product among registered suppliers before generating health care product recommendation information. It will be provided to the support target terminal 100 by further including it in the health care product recommendation information.

2 is a diagram showing the configuration of a health care product recommendation system for support target using artificial intelligence according to the present invention.

Referring to FIG. 2 , the support target health care product recommendation system 300 includes a storage unit 310 , a communication unit 320 , and a control unit 330 .

The storage unit 310 includes a program area for storing a control program for controlling the overall operation of the support target health care product recommendation system 300 according to the present invention, and a temporary area for temporarily storing data generated during the execution of the control program. , and a data area for permanently storing data required for the control program and data generated during execution of the control program. In the data area, according to the present invention, support target DB 311, supplier DB 312, product DB 313, expert recommendation DB 314, evaluation DB 315, dataset DB 316, etc. may be configured. will be able

The support target DB 311 stores member information including user information of the support target. The user information may be included in member information, or may be configured to be input whenever a support target requests for recommendation of a health care product.

The supplier DB 312 stores seller registration information for each supplier. As described above, the seller registration information may include supplier information, health care product name, model name, function information, sales price, and the like.

The product DB 313 stores health care product information including health care product names for a plurality of health care products. The health care product information includes the information of Table 1 described above.

The expert recommendation DB 314 stores expert recommendation information including a plurality of treatment details and recommended health care product information for each treatment history designated by the expert. The medical history includes the information of Table 2 described above.

The evaluation DB 315 stores evaluation information evaluated by a plurality of support subjects received from the cloud sourcing unit 400 . The evaluation information includes the information of Table 3 described above.

The dataset DB 316 stores a training dataset. The learning dataset includes health care product information and expert recommendation information, and may further include evaluation information according to an embodiment.

The communication unit 320 is connected to the wired and wireless data communication network 1 by any one of wired and wireless, and connected to the wired and wireless data communication network 1, the support target terminal 100, the expert recommendation information providing unit 200, the cloud sourcing unit ( 400), the health care product supply unit 500, and the control unit 330 to perform data communication, and to communicate with a plurality of devices constituting the Internet network through the wired/wireless data communication network 1 do.

The control unit 330 includes an information collection unit 340, a seller registration unit 350, a health care product recommendation model learning unit 360, and a health care product recommendation unit 370, a health care product recommendation system according to the present invention Controls the overall operation of (300).

The information collection unit 340 includes a learning dataset collection unit 341 and a user information generation unit 342 , the expert recommendation information providing unit 200 , the cloud sourcing unit 400 and the Internet through the communication unit 320 . The learning dataset configuration information and user information for configuring the learning dataset are collected from the websites on and stored in the storage unit 310 .

Specifically, the learning data set collecting unit 341 includes a product information collecting unit 343 , an expert recommendation information collecting unit 344 , and a learning data set generating unit 346 , and according to an embodiment, evaluation information It further includes a collection unit 345 .

The product information collection unit 343 accesses the wired/wireless data communication network 1 through the communication unit 320 in a first cycle, crawls and collects information on health care products from a plurality of websites on the Internet network, and collects the collected information. By analyzing the data, health care product information including the information elements of Table 1 is generated and stored in the product DB 313 . The second cycle may be one day, one week, one month, or the like.

The expert recommendation information collecting unit 344 collects expert recommendation information from the expert recommendation information providing unit 200 by accessing the expert recommendation information providing unit 200 through the communication unit 320 and the wired/wireless data communication network 1 in a second cycle. and stored in the expert recommendation DB 314 . The expert recommendation information collected from the expert recommendation information providing unit 200 may include only a plurality of treatment details, or may include a plurality of treatment details and recommended health care products for each treatment history. The second cycle may be a year, a month, or the like.

In the latter case, the expert recommendation information collection unit 344 provides the medical treatment details to a preset expert terminal (not shown), and receives the recommended health care product information for each treatment history according thereto to the expert recommendation DB 314 . It would be desirable to be configured to store.

The evaluation information collection unit 345 accesses the cloud sourcing unit 400 through the communication unit 320 and the wired/wireless data communication network 1 in a third cycle to receive a plurality of evaluation information collected in advance, or to the cloud in a third cycle. A plurality of evaluation information provided by the sourcing unit 400 is received and stored in the evaluation DB 315 . The third cycle may be one day, one week, one month, or the like.

The learning dataset generating unit 346 uses the health care product information and expert recommendation information collected through the product information collection unit 343 and the expert recommendation information collection unit 344 as a key using the health care product name, etc. to create a learning dataset by mapping, or according to another embodiment, a learning dataset is created using health care product information, expert recommendation information, and evaluation information as a health care product name as a key and stored in the dataset DB 316 .

The learning dataset generation unit 346 collects health care product information from the product information collection unit 343 , collects expert recommendation information from the expert recommendation information collection unit 344 , and collects evaluation information from the evaluation information collection unit 345 . It would be desirable to update the training dataset whenever either one of them occurs.

The user information collection unit 342 obtains user information from the support target terminal unit 100 of the accessed support target and outputs it to the health care product recommendation unit 370 . The user information may be information obtained in real time or information included in member information registered in advance.

The seller registration unit 350 receives the seller registration information from the plurality of health care product supply units 500 and stores it in the supplier DB 312 to register the seller, that is, the supplier.

The health care product recommendation model learning unit 360 has a preset health care product recommendation model that is an artificial intelligence model, and receives a learning data set from the learning data set generating unit 346 or stored in the data set DB 316 . The latest learning dataset is loaded and applied to the health care product recommendation model for learning.

The health care product recommendation model learning unit 360 outputs the learned health care product recommendation model to the health care product recommendation unit 370 .

The health care product recommendation unit 370 receives and applies the health care product recommendation model learned from the health care product recommendation model learning unit 360, and generates a health care product recommendation event when a health care product recommendation event occurs. After obtaining user information on the support target, which is the subject of the request, is applied to the health care product recommendation model, health care product recommendation information corresponding to the user information is generated, and then provided to the corresponding support target. The health care product recommendation event may be generated when a support target accesses through the support target terminal unit 100 and requests a health care product recommendation, or may be configured to occur automatically at a certain period. In the latter case, it will be possible to provide a service that provides optimal health care product recommendation information to registered support recipients at regular intervals.

The health care product recommendation information may include information such as health care product information, national support information, sales price of the health care product, and support amount, which is optimal for the user who is a support target in response to the region in which the user lives.

The health care product recommendation model of the health care product recommendation unit 370 may output only the health care product name or may further output other information.

The health care product recommendation unit 370 searches the product DB 313 for health care product information including the selling price corresponding to the health care product name or from the supplier DB 312 to search the health care product name, sale price, supplier, It may be configured to output health care product recommendation information including information such as performance information.

3 is a flowchart illustrating a method for recommending health care products for support target using artificial intelligence according to the present invention.

Referring to FIG. 3 , first, the control unit 330 collects health care product-related information through any one or more of various websites of the Internet network, the expert recommendation information providing unit 200 , and the cloud sourcing unit 400 . (S111). The health care product-related information may include health care product information and expert recommendation information, and may further include evaluation information according to embodiments.

When the health care product-related information is collected, the control unit 330 analyzes, classifies, and maps the health care product-related information to generate a learning dataset (S113).

When the learning data set is generated, the controller 330 applies the generated learning data set to the health care product recommendation model to learn the health care product lottery model (S115).

When the health care product recommendation model is learned, the control unit 330 monitors whether a health care product recommendation event occurs (S117), and when a health care product recommendation event occurs, the support target connected through the support target terminal 100 Requests the user to input user information (S119).

After the user information input request, the control unit 330 checks whether user information is input (S121), and when user information is input, applies the user information to the learned health care product recommendation model (S123).

If the support target is a member of the health care product recommendation service and has previously registered user information, the user information may be obtained by loading the registered user information. However, since disability information and disease information in user information can change frequently, it would be desirable to receive input whenever a health care product recommendation event occurs.

When the user information is applied to the learned health care product recommendation model, the control unit 330 monitors whether health care product recommendation information is output from the health care product recommendation model (S125), and when the health care product recommendation information is output, the health The management product recommendation information is transmitted to the support target terminal unit 100 (S127).

After the transmission of the health care product recommendation information, the control unit 330 monitors whether a request for supplier information for supplying the recommended health care product is generated from the support target terminal unit 100 (S129), and supplies the recommended health care product When a request for information on a supplier, that is, a seller, occurs, at least one or more suppliers who supply optimal recommended health care products are searched for in the supplier DB 312, and the retrieved supplier information is provided to the applicant terminal unit 100 (S131) ).

On the other hand, according to another embodiment, when the health care product recommendation information is output from the health care product recommendation model, the controller 330 checks information such as the provider information provision option of the corresponding support target in advance, and if it is set, the health care product recommendation information It may be configured to automatically search for a provider that provides health care products recommended in , and then include it in the health care product recommendation information and provide it to the support target terminal 100 .

In addition, when there is one or few values output from the health care product recommendation model, the control unit 330 stores additional information such as supplier information, selling price information, health care product performance information, national support information, and evaluation information to the supplier DB ( 312), the product DB 313, the expert recommendation DB 314, the evaluation DB 315, etc. may be extracted and included in the health care product recommendation information and provided to the applicant target terminal 100.

4 is a flowchart illustrating a method of collecting information and a method of configuring a learning dataset among the methods of recommending health care products for support target using artificial intelligence according to the present invention.

Referring to FIG. 4 , when the system is driven, the controller 330 initializes the information collection reference time and then counts the information collection reference time (S211).

When the information collection reference time starts to be counted, the control unit 330 determines whether it is the first period (S213), the second period (S215), or the third period (S217) based on the counted information collection reference time.

When it is determined that it is the first cycle, the controller 330 crawls health care product-related information from a plurality of websites online, that is, the Internet network (S219), and then analyzes the crawled health product-related information to only receive necessary information elements. By extracting health care product information, it is stored in the product DB 313 and DB is performed (S221).

On the other hand, in the second cycle, the control unit 330 collects or generates expert recommendation information to which the expert's treatment details and health care product information corresponding to the treatment details are mapped (S223).

And in the third cycle, the control unit 330 collects evaluation information on health care products through the cloud sourcing unit 400 (S225).

The control unit 330 determines whether health care product information and expert recommendation information are collected among the health care product information, expert recommendation information, and evaluation information in a finer cycle than the first cycle, second cycle, and third cycle (S227) ), whether at least the health care product information and expert recommendation information are collected, whether it is the first collection (S229), and whether one or more information among the three information elements is updated after the configuration of the learning dataset (S231).

When one of the three information elements is updated, the control unit 330 updates the learning dataset reflecting the updated information element (S231).

The updated learning dataset is applied to the health care product recommendation model again to re-lear the health care product recommendation model to improve performance.

On the other hand, it is common knowledge in the art that the present invention is not limited to the typical preferred embodiments described above, but can be improved, changed, replaced, or added in various ways without departing from the spirit of the present invention. Those who have will be able to understand it easily. If implementation by such improvement, change, substitution, or addition falls within the scope of the appended claims below, the technical idea should also be regarded as belonging to the present invention.

100: terminal unit for support target 200: expert recommendation information providing unit
300: health care product recommendation system 310: storage
311: Support target DB 312: Provider DB
313: Product DB 314: Expert Recommendation DB
315: evaluation DB 316: dataset DB
320: communication unit 330: control unit
340: information collection unit 341: training dataset collection unit
342: user information collection unit 343: product information collection unit
344: expert recommendation information collection unit 345: evaluation information collection unit
346: training dataset generator 350: seller register
360: Health care product recommendation model learning unit
370: health care product recommendation unit 400: cloud sourcing unit
500: Health care product supply department

Claims (14)

It is connected to the Internet network through a wired/wireless data communication network to perform data communication with devices of the Internet network, and to perform data communication with an expert recommendation information providing unit and a support target terminal unit through the wired/wireless data communication network. communication department;
A product DB that stores health care product information including health care product names for a number of health care products, and an expert who stores expert recommendation information including multiple treatment histories and recommended health care product information for each treatment specified by experts a storage unit including a recommendation DB; and
The expert recommendation information is collected from the expert recommendation information providing unit through the communication unit and stored in the expert recommendation DB, and information on health care products is crawled online through the communication unit, collected and analyzed to obtain the health care product name. After generating the health care product information including the information and storing it in the product DB, a learning dataset is created by mapping the health care product information of the product DB and the expert recommendation information of the expert recommendation DB, and a preset health care product recommendation model After learning by applying the learning dataset to the communication unit, the user information of the support target input from the support target terminal unit through the communication unit is applied to the health care product recommendation model to extract recommended health care products corresponding to the user information Then, the support target health care product recommendation system using artificial intelligence, characterized in that it comprises a control unit for transmitting to the support target terminal unit.
According to claim 1,
The control unit is
The expert recommendation information is collected from the expert recommendation information providing unit through the communication unit and stored in the expert recommendation DB, and information on health care products is crawled online through the communication unit, collected and analyzed to provide information on health care products. After generating health care product information including the health care product name and storing it in the product DB, the health care product information of the product DB and the expert recommendation information of the expert recommendation DB are mapped using the health care product name of the health care product information as a key an information collection unit for generating a learning data set and acquiring user information of a subject to be supported in real time, which is inputted in real time from the terminal unit for support through the communication unit;
a health care product recommendation model learning unit for learning by applying the learning dataset to a preset health care product recommendation model; and
After extracting a health care product corresponding to the user information by applying the user information, a health care product recommendation unit that transmits health care product recommendation information including health care product information about the extracted health care product to the support target terminal unit Support target health care product recommendation system using artificial intelligence, characterized in that it includes.
3. The method of claim 2,
The information collection unit,
Product information collecting unit that crawls and collects product information on health care products online through the communication unit, and then analyzes it to generate health care product information including the health care product name, manufacturer, and price information and stores it in the product DB ,
an expert recommendation information collecting unit for collecting the expert recommendation information from the expert recommendation information providing unit through the communication unit and storing it in the expert recommendation DB; and
a learning dataset collecting unit including a learning dataset generator configured to generate a learning dataset by mapping the health care product information of the product DB and the expert recommendation information of the expert recommendation DB to the health care product name of the health care product information as a key; and
Support target health care product recommendation system using artificial intelligence, characterized in that it comprises a user information collection unit for obtaining user information of the support target input in real time from the support target terminal unit through the communication unit.
4. The method of claim 3,
The learning data set collection unit,
Further comprising an evaluation information collection unit that collects the evaluation information from the cloud sourcing unit that collects evaluation information on the health care product from users including the support target who have tried a number of different health care products,
The training data set generation unit,
Support target health care product recommendation system using artificial intelligence, characterized in that by mapping the health care product name to a key to generate a learning dataset further including the evaluation information.
3. The method of claim 2,
The storage unit,
Further comprising a supplier DB for storing seller registration information including supplier (seller) information, health care product sales information including the price of one or more health care products sold by the supplier of the supplier information,
The control unit is
A seller registration unit that receives, from the health care product supply unit through the communication unit, seller registration information including supplier information for selling at least one health care product and the health care product sales information, stores it in the supplier DB, and registers it including,
The health care product recommendation unit,
Searching the supplier DB for a supplier that supplies a health care product under the health care product information of the health care product information, and providing supplier information and price information for supplying the health care product by further including the health care product recommendation information in the health care product recommendation information Support target health care product recommendation system using artificial intelligence, characterized in that.

6. The method of claim 5,
The health care product is
Health products including poultice products, posture correction, exercise equipment, eye health maintenance equipment, oxygen supply device, massage device, protector,
Welfare products including electric beds, bedsore prevention products, non-slip members, loitering sensors, safety holders, mobile toilets, walkers, and canes,
assistive devices, including sight aids, prosthetic/prosthetic devices, communication aids, postural aids, driving aids, wheelchairs, and hearing aids;
Support target health care product recommendation system using artificial intelligence, characterized in that it includes medical devices including hospital treatment devices, hospital measuring instruments, hospital maintenance devices, home measuring instruments, home washing machines, and home oriental medicinal products.
6. The method of claim 5,
The health care product information,
Product name (health care product name), model name (unique number), manufacturer, purchase price, option, category classification information, support target health care product recommendation system using artificial intelligence, characterized in that it includes the release date.
6. The method of claim 5,
The expert recommendation information is
Gender code, age code, province (residence) code, medical treatment start date, form code, medical course code, common disease code, injured disease code, number of days in the hospital, total number of prescription days and health care product information A system for recommending health care products to support recipients using artificial intelligence.
The control unit collects the expert recommendation information from the expert recommendation information providing unit through the communication unit, stores it in the expert recommendation DB, and crawls and collects information on health care products online through the communication unit, collects it, and analyzes it into health care products After generating health care product information including the health care product name for Korea and storing it in the product DB, the health care product information is used as a key to map the information in the product DB and the expert recommendation information in the expert recommendation DB to create a learning dataset information gathering process;
a learning process in which the controller applies the learning dataset to a health care product recommendation model preset through the health care product recommendation model learning unit to learn; and
a user information acquisition process in which the control unit acquires user information of the support target inputted in real time from the support target terminal unit through the communication unit; and
After the control unit extracts a health care product corresponding to the user information by applying the user information through the health care product recommendation unit, health care product recommendation information including health care product information on the extracted health care product is provided to the support target Support target health care product recommendation method using artificial intelligence, characterized in that it includes a health care product recommendation process transmitted to the terminal unit.
10. The method of claim 9,
The information collection process is
The control unit crawls and collects product information on health care products online through the communication unit, analyzes it, and generates health care product information including the health care product name, manufacturer, and price information, and stores it in the product DB information gathering step;
an expert recommendation information collecting step of the controller collecting the expert recommendation information from the expert recommendation information providing unit through the communication unit and storing the expert recommendation information in the expert recommendation DB; and
Support target using artificial intelligence using artificial intelligence, characterized in that the controller maps the information of the product DB and the expert recommendation information of the expert recommendation DB to the health care product information as a key to generate a learning dataset How to recommend health care products.
11. The method of claim 10,
The information collection process is
Further comprising an evaluation information collection step of collecting the evaluation information from the cloud sourcing unit that collects evaluation information on the health care product from users including the support target who have tried a number of different health care products,
The step of generating the training data set includes:
Support target health care product recommendation method using artificial intelligence, characterized in that by mapping the health care product name to a key to generate a learning dataset further including the evaluation information.
11. The method of claim 10,
Seller registration in which the control unit receives seller registration information including supplier information for selling at least one health care product from the health care product supply unit through the communication unit and seller registration information including the health care product sales information, stores it in the supplier DB and registers it further comprising the process,
The health care product recommendation process is
Searching the supplier DB for a supplier that supplies a health care product under the health care product information of the health care product information, and providing supplier information and price information for supplying the health care product by further including the health care product recommendation information in the health care product recommendation information A method for recommending health care products for support target using artificial intelligence, characterized in that.
13. The method of claim 12,
The health care product information,
Product name (health care product name), model name (unique number), manufacturer, purchase price, option, category classification information, support target health care product recommendation method using artificial intelligence, characterized in that it includes the release date.
13. The method of claim 12,
The expert recommendation information is,
Gender code, age code, province (residence) code, medical treatment start date, form code, medical course code, common disease code, injured disease code, number of days in the hospital, total number of prescription days and health care product information A method of recommending health care products for support recipients using artificial intelligence.

Images