System and method for recommending health functional foods to elderly patient with geriatric chronic disease

Abstract

The present invention relates to a system and method for recommending health functional foods to elderly patients with geriatric chronic diseases. According to the present invention, a method for recommending health functional foods, performed by a system for recommending health functional foods, comprises the steps of: receiving, from a target patient with a geriatric chronic disease, response data obtained from responses to a medical inquiry form or questionnaire; performing natural language processing on the response data; inputting the natural-language-processed response data into a first learning model to classify a mental disorder and a severity level of the target patient; receiving personal information and EMR data of the target patient; and inputting the personal information, EMR data , and information on the mental disorder and severity level of the target patient into a second learning model to recommend one or more health functional foods to the target patient.

Description

Health functional food recommendation system and method for elderly patients with chronic diseases

[0001] The present invention relates to a health functional food recommendation system and method for patients with chronic geriatric diseases, and more specifically, to a health functional food recommendation system and method that classifies patients suffering from chronic diseases according to the severity of mental illness and recommends health functional foods to the classified patients, thereby enabling more detailed treatment of the patients.

[0002] Existing diagnostic methods for mental illness utilized the complaints of patients and standardized questionnaires to ultimately determine the specific mental illness, relying on the knowledge and experience of psychiatrists based on the interpretation of the results.

[0003] Currently, to diagnose mental illness, patients are identified through self-report questionnaires or diagnosed by psychiatrists based on DSM-5 criteria. For example, in the case of anxiety disorders, self-report questionnaires such as the Beck Anxiety Inventory (BAI) and the State-Trait Anxiety Inventory (STAI) were used, while for depression, self-report questionnaires such as the Beck Depression Inventory (BDI) were utilized to identify the mental illness. However, self-report questionnaires have the drawback that results can be biased by the respondent's subjectivity.

[0004] Smart senior centers, which combine existing senior centers and non-face-to-face senior centers, have been distributed in Seogwipo City, Jeju Island, and other areas since January 2023, but the current situation remains at the level of introducing AI robots or metaverses.

[0005] As geriatric mental illnesses increase, the consumption of health functional foods to treat or prevent their deterioration is also important.

[0006] Generally, health functional foods, also referred to as health supplement foods or health functional foods, are foods manufactured using raw materials or ingredients (hereinafter referred to as "nutritional components") that possess beneficial functions for the human body and help maintain health. The health functional food market is expanding explosively due to increasing interest in health, and the market is expected to grow even further with the recent relaxation of sales facility standards for health supplements.

[0007] Furthermore, the nutritional components included in the aforementioned health functional foods are functional ingredients; depending on their type, they are classified into 'nutrient functions,' 'disease risk reduction functions,' and 'physiological functions,' with the physiological functions being further subdivided based on supporting data. Since health functional foods are required to include only nutritional components that have undergone safety evaluation by the Ministry of Food and Drug Safety, their safety is guaranteed to a certain extent; however, their effects may vary depending on individual physical characteristics, lifestyle habits, and health status.

[0008] Conventional health supplement recommendation systems consist of a process in which basic information such as name and contact details is registered, the functionality requested by the individual user is analyzed, health supplements are extracted based on the analysis results, and packages are assembled to recommend to the individual user. However, since such conventional recommendation systems are not based on separate diagnoses or prescriptions like pharmaceuticals, it was very difficult to accurately determine whether the acceptable intake of the nutrients contained in the actual recommended health supplements was suitable for the individual user.

[0009] Therefore, it is necessary to develop proper health supplement recommendation technology that meets individual user characteristics and adheres to acceptable intake levels.

[0010] The technology forming the background of the present invention is disclosed in Korean Registered Patent No. 10-2737689 (published Dec. 2, 2024).

[0011] The present invention aims to provide a health functional food recommendation system and method for elderly patients with chronic diseases, which classifies patients suffering from chronic diseases according to the severity of mental illness and recommends health functional foods to the classified patients, thereby enabling more detailed treatment of the patients.

[0012] According to an embodiment of the present invention, a method for recommending health functional foods performed by a health functional food recommendation system comprises the steps of: receiving response data from a patient with a geriatric chronic disease who has responded to a medical questionnaire or survey; processing the response data in natural language; inputting the naturally processed response data into a first learning model to classify the mental illness and severity of the corresponding patient; receiving personal information and EMR data regarding the patient; and inputting the patient's personal information, EMR information, mental illness and severity information into a second learning model to recommend one or more health functional foods to the patient.

[0013] The above response data includes text data or voice data, and if the response data is voice data, it may further include a step of converting the voice data into text data using a STT (Speech to Text) algorithm.

[0014] The step of classifying the mental illness and severity of the subject patient may classify the subject patient's mental illness into depression, anxiety disorder, or bipolar disorder, and classify the severity into any one of the labels of severe, moderate, mild, and normal.

[0015] The above personal information may include at least one of the following: the gender, age, history of chronic diseases, family history, types of preferred foods, types of disliked foods, and information on foods that must not be eaten by the subject patient.

[0016] The step of recommending a health functional food to the above-mentioned patient may provide the recommended health functional food and a health functional food contraindicated for concomitant use to the above-mentioned patient simultaneously.

[0017] The method may further include a step of training the first learning model by setting the natural language processing results of response data performed on multiple elderly patients with mental illness as input data, and setting the corresponding depression, anxiety disorder, bipolar disorder, and severity as output data.

[0018] The method may further include a step of training the second learning model by setting the mental illness and severity, personal information, and EMR information of multiple elderly patients having at least one of depression, anxiety disorder, or bipolar disorder as input data, and setting the type of health functional food recommended and currently being taken as output data.

[0019] According to another embodiment of the present invention, a health functional food recommendation system for patients with chronic geriatric diseases comprises: an input unit that receives response data from a patient with a chronic geriatric disease who has responded to a medical questionnaire or survey, as well as personal information and EMR data regarding the patient; an analysis unit that processes the response data in natural language; a mental illness classification unit that inputs the naturally language processed response data into a first learning model to classify the mental illness and severity of the corresponding patient; and a health functional food recommendation unit that inputs the patient's personal information, mental illness, and severity information into a second learning model to recommend one or more health functional foods to the patient.

[0020] According to the present invention, the automation of artificial intelligence models can minimize the consumption of manpower in clinical settings, and since no separate equipment is required, there are no spatial limitations. Furthermore, it enables precise prediction and diagnostic classification of the severity of depression and normality without requiring a psychiatrist to directly evaluate records containing patients' primary mental complaints. By streamlining the diagnostic work of psychiatrists, it allows them to focus on the essential treatment of patients with depression and enables them to concentrate more on researching new methods for treating patients with mental illnesses.

[0021] Furthermore, the present invention not only predicts mental illness but also enables proactive care of the disease by recommending health functional foods to the patient.

[0022] Furthermore, patients suffering from chronic diseases can be treated more subdivided by classifying them according to the severity of mental illness, particularly depression, and recommending health functional foods to each patient for each case.

[0023] To provide detailed recommendations for health functional foods, the performance of recommendation models can be enhanced by including structured data in addition to natural language data; furthermore, since structured data includes information on nutrients that should not be recommended to patients, side effects caused by health functional foods can be minimized.

[0024] In particular, because it efficiently recommends health functional foods to consume, patients can minimize the financial burden of treatment, and by classifying patients with chronic diseases according to the severity of their mental illness, it allows for the efficient identification of recommended health functional foods.

[0025] In addition, it has the advantage of being applicable to most patients because it analyzes chronic diseases (hypertension, cancer, diabetes, etc.) that the elderly are generally susceptible to.

[0026] FIG. 1 is a diagram showing a health functional food recommendation system for elderly patients with chronic diseases according to an embodiment of the present invention.

[0027] Figure 2 is a diagram illustrating the configuration of the health functional food recommendation system shown in Figure 1.

[0028] FIG. 3 is a flowchart illustrating a method for recommending health functional foods according to an embodiment of the present invention.

[0029] Figure 4 is an example of label processing for a natural language response to explain step S310 of Figure 3.

[0030] Figure 5 is a diagram illustrating the process of classifying the severity of mental illness in elderly patients with chronic diseases at step S350 of Figure 3.

[0031] FIGS. 6a to 6c are drawings for explaining step S370 of FIG. 3.

[0032] Figure 7 is an example diagram showing recommended dietary supplements and dietary supplements contraindicated for concomitant use for depression among mental illnesses in patients with dementia symptoms.

[0033] Then, with reference to the attached drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily implement the invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification have been given similar reference numerals.

[0034] Throughout the specification, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected," but also cases where they are "electrically connected" with other components interposed between them. Furthermore, when a part is described as "including" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0035] FIG. 1 is a diagram showing a health functional food recommendation system for elderly patients with chronic diseases according to an embodiment of the present invention.

[0036] As shown in FIG. 1, a health functional food recommendation system (100) according to an embodiment of the present invention can be connected to a user terminal (200) via a network.

[0037] A health functional food recommendation system (100) according to an embodiment of the present invention can be connected to a user terminal (200) via a wired, wireless, or wired-wireless combined network to transmit and receive mutual information. The wireless network may include at least one of RF, WLAN, Wi-Fi, and Bluetooth methods, and various known wireless network methods may be used.

[0038] The health functional food recommendation system (100) according to the embodiment of the present invention may be implemented as an online platform, such as a web server or app server, that provides an artificial intelligence-based health functional food recommendation service based on mental illness to a network-connected user terminal (200), or may be implemented in the form of an application program, application, etc., on the user terminal, etc.

[0039] A health functional food recommendation system (100) according to an embodiment of the present invention may provide a health functional food recommendation service platform for mental illness, implemented as an application or a web, to a network-connected user terminal (200). The service platform may be an application program running in an application or web environment.

[0040] In this way, the health functional food recommendation system (100) can be implemented as an application program executed on a platform server or user terminal (200) that provides a health functional food recommendation service according to mental illness, and the user terminal (200) can be connected to the system (100) via a network while the relevant application program is running to receive the relevant service.

[0041] The user terminal (200) may include a device capable of exchanging information by connecting to a wired or wireless network, such as a PC, desktop, smartphone, tablet, or notebook. Here, the user terminal (200) may correspond to a terminal on the side of a patient with an elderly chronic disease or a terminal on the side of a medical staff.

[0042] Figure 2 is a diagram illustrating the configuration of the health functional food recommendation system shown in Figure 1.

[0043] As shown in FIG. 2, a health functional food recommendation system (100) according to an embodiment of the present invention includes a learning unit (110), an input unit (120), an analysis unit (130), a mental illness classification unit (140), and a health functional food recommendation unit (150). Here, the operation of each unit (110 to 150) and the data flow between each unit can be controlled by a control unit (not shown).

[0044] This health functional food recommendation system (100) may be physically configured and implemented as a computer device including a processor, memory, user interface input / output device and storage device, network input / output unit, etc., or may be implemented as an application program running on a computer device or user terminal.

[0045] First, the learning unit (110) sets the natural language processing results of survey data conducted on multiple elderly patients who have at least one mental illness among depression, anxiety disorder, and bipolar disorder as input data, and sets the labeling values ​​corresponding to the depression, anxiety disorder, bipolar disorder and severity (e.g., Severe, Moderate, Mild, and Normal) as output data to train the first learning model.

[0046] Additionally, the learning unit (110) sets the mental illness and severity, personal information, and EMR information of multiple elderly patients who have at least one of depression, anxiety disorder, or bipolar disorder as input data, and sets the type of health functional food recommended and currently being taken as output data to train a second learning model.

[0047] Next, the input unit (120) receives survey data from a patient who has a chronic geriatric disease including at least one of dementia, Parkinson's disease, and stroke, in which the patient responds to a medical questionnaire or survey.

[0048] Here, the response data includes text data or voice data.

[0049] In addition, the input unit (120) receives personal information and EMR data regarding the target patient.

[0050] Here, personal information includes at least one of the following: the patient's gender, age, history of chronic diseases, types of preferred foods, types of disliked foods, and information on foods that must not be eaten.

[0051] Next, the analysis unit (130) processes the survey data into natural language to generate embedding vector values.

[0052] At this time, if the response data is voice data, the analysis unit (130) converts the voice data into text data using a STT (Speech to Text) algorithm.

[0053] And the mental illness classification unit (140) inputs the natural language processed embedding vector value into the first learning model to classify the mental illness and severity of the corresponding target patient.

[0054] The health functional food recommendation unit (150) inputs personal information, mental illness and severity information of a patient with a chronic geriatric disease into a second learning model and recommends one or more health functional foods to the patient.

[0055] Hereinafter, a method for recommending health functional foods according to an embodiment of the present invention will be described through FIGS. 3 to 7.

[0056] FIG. 3 is a flowchart illustrating a method for recommending health functional foods according to an embodiment of the present invention.

[0057] As shown in FIG. 3, first, the learning unit (110) sets the natural language processing results of response data performed on multiple elderly patients with mental illness as input data, and sets the corresponding depression, anxiety disorder, bipolar disorder and severity as output data to train a first learning model for determining mental illness (S310).

[0058] Here, mental illness includes at least one of depression, anxiety disorder, and bipolar disorder, and severity is classified into severe, moderate, mild, and normal, each with a single label.

[0059] That is, as shown in Table 1 below, it is labeled with a value of 3 for severe, 2 for moderate, 1 for mild, and 0 for normal.

[0060]

[0061] Figure 4 is an example of label processing for a natural language response to explain step S310 of Figure 3.

[0062] In particular, Figure 4 is an example showing the results of natural language responses and labeling of severity for multiple elderly patients with depression among multiple mental illnesses.

[0063] As shown in FIG. 4, according to an embodiment of the present invention, a health functional food recommendation system (100) refines the content of natural language responses to a survey targeting elderly patients who have mental illness or are in a normal state among a plurality of elderly patients, and inputs the results of labeling the data sets regarding which mental illness was diagnosed and the severity compared to the natural language expression.

[0064] At this time, as shown in Fig. 4, the dataset consists of pid (anonymized subject identification), re-examination_diagnosis_name, natural language expression content, and label (diagnosed disease).

[0065] The analysis unit (130) tokenizes natural language expressions of mental illnesses using natural language programming techniques for artificial intelligence learning and counts frequent expressions for the tokenized natural language expressions of mental illnesses such as depression and normal groups.

[0066] Then, the learning unit (110) performs artificial intelligence learning on the first learning model by linking the data in which the severity of mental illnesses such as depression and normal group expressions are counted with existing labels.

[0067] Then, a first learning model can be obtained in which learning is performed on words or expressions that are frequently used according to the severity of mental illness.

[0068] Next, the learning unit (110) sets the mental illness and severity, personal information, and EMR information of multiple elderly patients who have at least one of depression, anxiety disorder, or bipolar disorder as input data, and sets the type of health functional food recommended and currently being taken as output data to train a second learning model for health functional food recommendation (S320).

[0069] Here, personal information includes at least one of gender, age, history of chronic diseases, family history, types of preferred foods, types of disliked foods, and information on foods to avoid. Additionally, EMR information includes at least one of medical records, previous medical history and surgical history, diagnosis and prescription information, prescribed drug information, symptom information, treatment and surgery data, vital sign data, blood test data, genetic test results, insurance claim information, lifestyle pattern information, and smoking and drinking habit information.

[0070] Once the learning for the first and second learning models is completed in this manner, the testing process proceeds.

[0071] First, the input unit (120) receives response data from a patient with a chronic geriatric disease who has answered a medical questionnaire or survey (S330).

[0072] Here, chronic geriatric diseases include at least one of dementia, Parkinson's disease, and stroke.

[0073] The input unit (120) receives data in which the patient responds to a questionnaire related to mental illness directly by hand or receives input via voice.

[0074] Here, the questionnaire may include the PHQ-9 (Patient Health Questionnaire-9) for confirming depression, the GAD-7 (Generalized Anxiety Disorder-7) for determining anxiety disorders, the PCL-5 (PTSD Checklist for DSM-5) for evaluating post-traumatic stress disorder (PTSD), the MDQ (Mood Disorder Questionnaire) for screening bipolar disorder or manic-depressive disorder, the MMPI (Minnesota Multiphasic Personality Inventory) for evaluating overall mental health and personality traits, the BAI (Beck Anxiety Inventory) for determining the level of anxiety, and the Beck Depression Inventory (BDI) for evaluating depression.

[0075] In addition, the survey can be conducted via voice to account for the possibility that the target patients may be elderly.

[0076] Next, the analysis unit (130) processes the input response data into natural language (S340).

[0077] At this time, the analysis unit (130) performs natural language processing on the input sentence immediately when the response data is sentence data, and when the response data is voice data, it can convert the voice data into text data using an STT (Speech to Text) algorithm and then perform natural language processing.

[0078] Since the response data of the target patient is unstructured data, the analysis unit (130) can perform natural language processing on the unstructured response data using a tokenization technique.

[0079] And, the mental illness classification unit (140) inputs the natural language processed embedding vector value into the first learning model to classify the mental illness and severity of the corresponding target patient (S350).

[0080] Figure 5 is a diagram illustrating the process of classifying the severity of mental illness in elderly patients with chronic diseases at step S350 of Figure 3.

[0081] The learning unit (110) inputs the natural language processed response data of the target patient into the first learning model, and the first learning model outputs, based on the results of prior learning, which mental illness the target patient has among depression, anxiety disorder, and bipolar disorder, and the respective probabilities for the severity of the corresponding mental illness as severe, moderate, mild, and normal.

[0082] That is, according to an embodiment of the present invention as shown in FIG. 5, response data from a patient with a chronic geriatric disease including at least one of dementia, Parkinson's disease, and stroke can be analyzed to determine whether there is depression, anxiety disorder, or bipolar disorder and the severity thereof.

[0083] The learning unit (110) analyzes the input natural language expressions regarding the reason why the probability of a mental illness such as depression or a normal level of severity was obtained, and presents the result of such a predicted diagnosis result through an artificial intelligence capable of explaining which response expression was used, thereby supporting the identification of causes and the calculation of the treatment range for the precise diagnosis of mental illness such as depression.

[0084] In this way, when the mental illness and severity of the target patient are output, the input unit (120) receives personal information and EMR data regarding the target patient from the target patient or medical staff (S360).

[0085] In other words, it receives not only the personal information of the target patient but also EMR data information corresponding to the clinical case.

[0086] And, the health functional food recommendation unit (150) inputs the patient's personal information, EMR information, mental illness and severity information into a second learning model and recommends one or more health functional foods to the patient (S370).

[0087] That is, the health functional food recommendation unit (150) adds the target patient's personal information and EMR information to the mental illness and severity information output from the first learning model and inputs it into the second learning model to provide recommended health functional foods and health functional foods contraindicated for concomitant use.

[0088] FIGS. 6a to 6c are drawings for explaining step S370 of FIG. 3.

[0089] FIGS. 6a to 6c provide examples of target patients having symptoms of dementia, and as seen in FIGS. 6a to 6c, the health functional food recommendation system (100) according to an embodiment of the present invention provides a list of recommended health functional foods and contraindicated health functional foods according to the severity of each mental disease for target patients having symptoms of dementia.

[0090] Figure 7 is an example diagram showing recommended dietary supplements and dietary supplements contraindicated for concomitant use for depression among mental illnesses in patients with dementia symptoms.

[0091] As shown in FIG. 7, there are quite a few types of health functional foods recommended based solely on the severity of chronic and mental illnesses, but among them there are some that should not be consumed depending on the patient's condition. That is, since the effects on patients with various mental illnesses may differ depending on the type of health functional food, the health functional food provision system (100) according to the embodiment of the present invention can enhance the treatment effect on the patient by recommending recommended health functional foods and prohibited health functional foods, respectively, according to the patient's mental illness and severity.

[0092] In addition, the recommended list of health supplements in the health supplement provision system (100) according to an embodiment of the present invention consists of health supplements that can ultimately treat the patient most efficiently, and provides information on the recommended health supplements including information on priority.

[0093] As such, according to the embodiments of the present invention, the automation of the artificial intelligence model can minimize the consumption of manpower in clinical settings, and since no separate equipment is required, there are advantages such as no spatial limitations. Furthermore, it is possible to precisely predict and classify the severity of depression and normality without a psychiatrist directly evaluating records containing patients' primary mental complaints; by streamlining the diagnostic work of psychiatrists, this enables them to focus on the essential treatment of patients with depression and allows them to concentrate more on researching new methods for treating patients with mental illnesses.

[0094] Furthermore, the present invention not only predicts mental illness but also enables proactive care of the disease by recommending health functional foods to the patient.

[0095] Furthermore, patients suffering from chronic diseases can be treated more subdivided by classifying them according to the severity of mental illness, particularly depression, and recommending health functional foods to each patient for each case.

[0096] To provide detailed recommendations for health functional foods, the performance of recommendation models can be enhanced by including structured data in addition to natural language data; furthermore, since structured data includes information on nutrients that should not be recommended to patients, side effects caused by health functional foods can be minimized.

[0097] In particular, because it efficiently recommends health functional foods to consume, patients can minimize the financial burden of treatment, and by classifying patients with chronic diseases according to the severity of their mental illness, it allows for the efficient identification of recommended health functional foods.

[0098] In addition, it has the advantage of being applicable to most patients because it analyzes chronic diseases (hypertension, cancer, diabetes, etc.) that the elderly are generally susceptible to.

[0099] The present invention has been described with reference to embodiments illustrated in the drawings, but this is merely illustrative, and those skilled in the art will understand that various modifications and equivalent alternative embodiments are possible therefrom. Accordingly, the true technical scope of protection of the present invention should be determined by the technical spirit of the appended claims.

Claims

1. In a method for recommending health functional foods performed by a health functional food recommendation system, A step of receiving response data from medical questionnaires or surveys from target patients with geriatric chronic diseases, Step of natural language processing of the above response data, A step of inputting natural language processed response data into a first learning model to classify the mental illness and severity of the corresponding target patient, A step of receiving personal information and EMR data regarding the aforementioned patient, and A method for recommending health functional foods, comprising the step of inputting the personal information, EMR information, mental illness and severity information of the aforementioned patient into a second learning model to recommend one or more health functional foods to the aforementioned patient.

2. In Claim 1, The above response data includes text data or voice data, If the above response data is voice data, A method for recommending health functional foods that further includes the step of converting the above voice data into text data using a STT (Speech to Text) algorithm.

3. In Claim 1, The step of classifying the mental illness and severity of the above-mentioned patient is, The mental illness of the above-mentioned patient is classified into depression, anxiety disorder, and bipolar disorder, and A method for recommending health functional foods that classifies the above severity into one of the labels: Severe, Moderate, Mild, and Normal.

4. In Claim 1, The above personal information is, A method for recommending health functional foods that includes at least one of the following: the gender, age, history of chronic diseases, family history, types of preferred foods, types of disliked foods, and information on foods that must not be eaten of the above-mentioned patient.

5. In Claim 1, The step of recommending health functional foods to the above-mentioned patients is, A method for recommending health functional foods by simultaneously providing recommended health functional foods and health functional foods contraindicated for concomitant use to the above-mentioned target patient.

6. In Claim 3, A method for recommending health functional foods, further comprising the step of training a first learning model by setting the natural language processing results of response data performed on multiple elderly patients with mental illness as input data, and setting the corresponding depression, anxiety disorder, bipolar disorder, and severity as output data.

7. In Claim 6, A method for recommending health functional foods, further comprising the step of training a second learning model by setting the mental illness and severity, personal information, and EMR information of multiple elderly patients having at least one of depression, anxiety disorder, or bipolar disorder as input data, and setting the type of health functional food recommended and currently being taken as output data.

8. In a health functional food recommendation system for patients with geriatric chronic diseases, An input unit that receives response data from a patient with a geriatric chronic disease who has answered a medical questionnaire or survey, as well as personal information and EMR data regarding the said patient. Analysis unit that processes the above response data into natural language, A mental illness classification unit that inputs natural language processed response data into a first learning model to classify the mental illness and severity of the corresponding target patient, and A health functional food recommendation system comprising a health functional food recommendation unit that recommends one or more health functional foods to the target patient by inputting the target patient's personal information, mental illness, and severity information into a second learning model.

9. In Claim 8, The above response data includes text data or voice data, If the above response data is voice data, The above analysis unit is, A health functional food recommendation system that converts the above voice data into text data using a STT (Speech to Text) algorithm.

10. In Claim 8, The above classification of mental disorders is, The mental illness of the above-mentioned patient is classified into depression, anxiety disorder, and bipolar disorder, and A health functional food recommendation system that classifies the above severity into one of the labels: Severe, Moderate, Mild, and Normal.

11. In Claim 8, The above personal information is, A health functional food recommendation system comprising at least one of the following: the gender, age, history of chronic diseases, family history, types of preferred foods, types of disliked foods, and information on foods that must not be eaten of the above-mentioned patient.

12. In claim 8, The above-mentioned health supplement recommendation section is, A health functional food recommendation system that simultaneously provides recommended health functional foods and health functional foods contraindicated for concomitant use to the above-mentioned target patients.

13. In Claim 10, A health functional food recommendation system further comprising a learning unit that sets the natural language processing results of response data performed on multiple elderly patients with mental illnesses as input data, and sets the corresponding depression, anxiety disorder, bipolar disorder, and severity as output data to train the first learning model.

14. In Claim 13, The above learning unit is, A health functional food recommendation system that sets the mental illness and severity, personal information, and EMR information of multiple elderly patients having at least one of depression, anxiety disorder, or bipolar disorder as input data, and sets the types of health functional foods recommended and currently being taken as output data to train the second learning model.

Images