Artificial intelligence-based solution recommendation method and system for infant health management

Abstract

An artificial intelligence-based solution recommendation method for infant health management of the present invention comprises steps in which: a data collection unit collects life pattern data including nursing data or sleep data of infants; a life pattern extraction unit refines development data and life pattern data from the collected life pattern data, and matches the refined life pattern data and the refined development data to extract a plurality of monthly life patterns; a related factor extraction unit analyzes the extracted plurality of monthly life patterns to extract problems and life pattern related factors for each monthly life pattern; a similarity measurement unit measures the similarity between actual input data and the monthly life patterns; a training unit trains an artificial intelligence neural network model by inputting the actual input data and the monthly life patterns to the artificial intelligence neural network model, and derives a weight for each related factor; and a recommendation unit recommends non-therapeutic commercial drugs by means of large language model (LLM)-based symptom analysis.

Description

AI-based solution recommendation method and system for infant and toddler health management

[0001] The present invention relates to an AI-based solution recommendation method and system for infant and toddler health management, and more specifically, to an AI-based solution recommendation method and system for infant and toddler health management capable of recommending customized lifestyle pattern schedules for each infant and toddler based solely on simple survey results.

[0002] Sleep is a significant factor influencing growth during infancy and early childhood. In particular, research indicates that sleep has a major impact on obesity rates and growth hormone secretion. Therefore, ensuring adequate sleep time is crucial, and accurate information regarding when to feed and take naps is being requested to achieve this.

[0003] However, infants exhibit significant pattern changes in feeding amounts, feeding intervals, sleep amounts, and sleep duration at intervals of just a few weeks after birth. In other words, recommendations to ensure proper development and stable sleep for infants vary drastically. For instance, recommended feeding amounts and sleep durations differ depending on how much was fed and how much sleep was taken the previous day, and they can also vary significantly depending on whether the infant is currently 3 weeks old or 6 weeks old. These recommended feeding amounts and sleep durations cannot be accurately adjusted without the assistance of a professional.

[0004] Prior art includes Korean Registered Patent No. 10-0659695 (System and method for providing a service to check the growth and development stages of infants), but it merely discloses a technology in which statistical data is generated by statistically processing existing standard growth and development data regarding infant growth and development, and a comparative analysis of the infant's growth and development status is performed based on this generated statistical data upon a client's request to check the infant's growth and development stages.

[0005] The problem that the present invention aims to solve is to provide an AI-based solution recommendation method and system for infant health management that provides a customized lifestyle pattern schedule, including optimal sleep and feeding times for each infant, based solely on basic information about the infant and a simple survey result.

[0006] The method for recommending an AI-based solution for infant health management according to the present invention comprises: a data collection unit collecting lifestyle pattern data including feeding data or sleep data of an infant; a lifestyle pattern extraction unit refining developmental data and lifestyle pattern data from the collected lifestyle pattern data and matching the refined lifestyle pattern data and developmental data to extract multiple lifestyle patterns by month; an association factor extraction unit analyzing the extracted multiple lifestyle patterns by month to extract problems and lifestyle pattern association factors by month and lifestyle pattern; a similarity measurement unit measuring the similarity between actual input data and lifestyle patterns by month; a learning unit inputting actual input data and lifestyle patterns by month into an AI neural network model to perform learning and deriving weights for each association factor; a survey unit extracting questions regarding the association factor with the highest weight from a database and providing them to a user terminal; and a recommendation unit generating a customized lifestyle pattern timetable based on survey results input from the user terminal and providing it to the user terminal.

[0007]

[0008] According to the present invention, there is an effect of conveniently providing a customized lifestyle pattern schedule for each infant through a simple survey without the need to directly input feeding and sleep data of the infant.

[0009] In addition, it can automatically calculate recommended feeding amounts or times based on feeding amounts or times and provide them to the user in real time, thereby effectively assisting infants and toddlers in feeding by notifying them of the appropriate amount or time in a timely manner.

[0010] FIG. 1 is a flowchart illustrating an artificial intelligence-based solution recommendation method for infant health management according to an embodiment of the present invention.

[0011] FIG. 2 is a configuration diagram illustrating an artificial intelligence-based solution recommendation system for infant health management according to an embodiment of the present invention.

[0012] FIGS. 3 and FIGS. 4 are drawings illustrating a judgment method using cosine similarity according to an embodiment of the present invention.

[0013] FIGS. 5 to 9 are drawings illustrating application service operation screens according to an embodiment of the present invention.

[0014]

[0015] Specific structural or functional descriptions of embodiments according to the concept of the present invention disclosed herein are provided merely for the purpose of explaining embodiments according to the concept of the present invention, and embodiments according to the concept of the present invention may be implemented in various forms and are not limited to the embodiments described herein.

[0016] Embodiments according to the concept of the present invention may be subject to various modifications and may take various forms; therefore, embodiments are illustrated in the drawings and described in detail in this specification. However, this is not intended to limit the embodiments according to the concept of the present invention to specific disclosed forms, and includes all modifications, equivalents, or substitutions that fall within the spirit and scope of the present invention.

[0017] The terms used herein are used merely to describe specific embodiments and are not intended to limit the invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described herein, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0018] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings attached to this specification.

[0019]

[0020] FIG. 1 is a flowchart illustrating an artificial intelligence-based solution recommendation method for infant health management according to an embodiment of the present invention.

[0021] Referring to FIG. 1, the artificial intelligence-based solution recommendation method for infant health management according to the present invention first collects lifestyle pattern data including feeding data or sleep data of an infant (S110). At this time, the feeding data includes the number and interval of feeding during the day and the number and interval of feeding at night, and the sleep data includes the number of naps, the duration of the nap, the total sleep time, the start time of night sleep, the end time of night sleep, the duration, and the interval of the previous sleep. The lifestyle pattern data may include weaning food data, and the weaning food data may include the number of weaning foods and the interval of weaning foods.

[0022] The lifestyle pattern extraction unit (120) can refine the developmental data and lifestyle pattern data from the collected lifestyle pattern data, and can extract multiple lifestyle patterns by month by matching the refined lifestyle pattern data and developmental data (S120). The lifestyle pattern extraction unit (120) can refine the data by organizing the extracted lifestyle pattern data into similar patterns and deleting patterns among the organized similar patterns that are judged to be special cases that cannot be recommended. The lifestyle pattern extraction unit (120) can extract lifestyle patterns by month through the above refinement process. For example, lifestyle patterns by month as shown below can be extracted. The lifestyle pattern extraction unit (120) can extract multiple lifestyle patterns by month.

[0023] The related factor extraction unit (130) can analyze the extracted multiple lifestyle patterns by age and extract problems and lifestyle pattern related factors by age and lifestyle pattern (S130). The related factor extraction unit (130) can first extract problems by age and lifestyle pattern, organize the related factors used to determine the problems, and organize the frequency of occurrence of problems by age.

[0024] The related factor extraction unit (130) can extract related factors based on the overlapping related factors among the related factors related to the problems by extracted age and lifestyle pattern and the related factors related to the lifestyle pattern (S130). The related factor extraction unit (130) can determine nine related factors as major related factors, such as the number of naps, total nap time, total nighttime nap time, total sleep time, total number of feedings, feeding interval, weaning interval, number of weaning meals, and number of nighttime feedings, and can match the problems with the multiple lifestyle patterns extracted by the lifestyle pattern extraction unit.

[0025] The similarity measurement unit (140) can measure the similarity between actual input data and the lifestyle pattern by age (S140). The similarity measurement unit (140) can scale the time in each column of the standard lifestyle pattern of infants and toddlers by age by converting it into a decimal number table. The similarity measurement unit (140) can scale the user's actual input data in the same way. The similarity measurement unit (140) can set the variables of the standard lifestyle pattern of infants and toddlers by age to A1 to A9, set the variables of the user's actual input data to B1 to B9, and calculate the cosine similarity by applying a separate scale control table. At this time, the scale control table can be set as weights to C1 to C9, and the variables C1 to C9 may be weights based on expert knowledge by age and the frequency of major problem occurrences. The similarity measurement unit (140) can measure the similarity according to the similarity calculation formula below.

[0026] AA = SQRT[(A1*C1) 2 +(A2*C2) 2 +...+(A9*C9) 2 ]

[0027] BB = SQRT[(B1*C1) 2 +*B2*C2) 2 +... +(b9*C9) 2 ]

[0028] MM = [(A1*C1*B1*C1) + (A2*C2*B2*C2)+...+(A9*C9*B9*C9)]

[0029] Similarity = AA * BB / MM

[0030] Variables A1 to A9 of the standard lifestyle patterns of infants and toddlers by age and variables B1 to B9 of the user's actual input data may be nine associated factors of the number of naps, total nap time, total nighttime sleep time, total sleep time, total number of feedings, feeding interval, weaning interval, number of weanings, and number of nighttime feedings.

[0031] The learning unit (150) inputs actual input data and lifestyle patterns by month into an artificial intelligence neural network model to perform learning and can derive weights for each associated factor (S150). The learning unit (150) can perform learning by inputting variable values ​​C1 to C9 corresponding to the weights into the neural network model. At this time, learning can be performed between the user's actual input data and the lifestyle schedule proposed by an expert, and optimal weights can be derived for each month and problem.

[0032] The recommendation unit can recommend over-the-counter drugs through symptom analysis based on LLM (Large Language Model) (S160).

[0033] The survey unit (160) can extract a question regarding the association factor with the highest weight from the survey DB (170) and provide it to the user terminal (200) (S170). The survey unit (160) can extract a user's lifestyle pattern timetable through general questions including the user's age and the problem to be solved, and can provide a customized lifestyle pattern timetable to the user without data input through a question regarding the association factor with the highest weight derived from an artificial intelligence learning model. The survey unit (160) can extract a question regarding the association factor with the highest weight from among multiple questions stored in the survey DB (170) and provide it to the user.

[0034] The recommendation unit (180) can generate customized lifestyle patterns and non-prescription commercial drug ingredients based on the survey results input from the user terminal (200) and provide them to the user terminal (S180).

[0035]

[0036] FIG. 2 is a configuration diagram illustrating an artificial intelligence-based solution recommendation system for infant health management according to an embodiment of the present invention. The configuration of the system (10) shown in FIG. 2 is merely a simplified example. In one embodiment of the present invention, the operating server and user terminal of the system may include other configurations for performing a computing environment, and only some of the disclosed configurations may constitute the system (10).

[0037] Referring to FIG. 2, an artificial intelligence-based solution recommendation system (10) for infant health management consists of an operating server (100) and a user terminal (200). The operating server (100) consists of a data collection unit (110), a lifestyle pattern extraction unit (120), an association factor extraction unit (130), a similarity measurement unit (140), a learning unit (150), a survey unit (160), a survey DB (170), a recommendation unit (180), and a control unit (190).

[0038] The data collection unit (110) can collect lifestyle pattern data including feeding data or sleep data of infants and toddlers. The data collection unit (110) can collect feeding data or sleep data of infants and toddlers directly entered by users of the application service, and can also collect data in the form of sentences, such as memos entered by users. The feeding data includes the number and interval of feeding during the day and the number and interval of feeding at night, and the sleep data includes the number of naps, the duration of the nap, the total sleep time, the start time of night sleep, the end time of night sleep, the duration, and the interval of the previous sleep. The lifestyle pattern data may include weaning food data, and the weaning food data may include the number of weaning foods and the interval of weaning foods.

[0039] The lifestyle pattern extraction unit (120) can refine the developmental data and lifestyle pattern data from the collected lifestyle pattern data, and can extract multiple lifestyle patterns by month by matching the refined lifestyle pattern data and developmental data. The lifestyle pattern extraction unit (120) can refine the data by organizing the extracted lifestyle pattern data into similar patterns and deleting patterns among the organized similar patterns that are judged to be special cases that cannot be recommended. The lifestyle pattern extraction unit (120) can extract lifestyle patterns by month through the above refinement process.

[0040] The associated factor extraction unit (130) can extract associated factors based on overlapping associated factors among the associated factors related to problems by extracted age-specific lifestyle patterns and associated factors related to lifestyle patterns. The associated factor extraction unit (130) can determine nine associated factors—number of naps, total nap time, total nighttime nap time, total sleep time, total feeding frequency, feeding interval, weaning interval, number of weaning meals, and number of nighttime feedings—as major associated factors, and match the problems with multiple lifestyle patterns extracted by the lifestyle pattern extraction unit. In the present invention, nine associated factors are determined as major associated factors, but the number of associated factors is not limited.

[0041] The similarity measurement unit (140) can measure the similarity between actual input data regarding infants and toddlers and lifestyle patterns by age. The actual input data regarding infants and toddlers may include feeding data or sleep data. The similarity measurement unit (140) can scale the time in each column of the standard lifestyle patterns of infants and toddlers by age by converting it into a decimal number table. The similarity measurement unit (140) can scale the user's actual input data in the same way. The similarity measurement unit (140) can calculate cosine similarity by setting the variables of the standard lifestyle patterns of infants and toddlers by age to A1 to A9, setting the variables of the user's actual input data to B1 to B9, and applying a separate scale control table. At this time, the scale control table can be set as weights to C1 to C9, and the variables C1 to C9 may be weights based on expert knowledge by age and the frequency of occurrence of major problems. The similarity measurement unit (140) can measure similarity according to the similarity calculation formula below.

[0042] AA = SQRT[(A1*C1) 2 +(A2*C2) 2 +...+(A9*C9) 2 ]

[0043] BB = SQRT[(B1*C1) 2 +*B2*C2) 2 +... +(b9*C9) 2 ]

[0044] MM = [(A1*C1*B1*C1) + (A2*C2*B2*C2)+...+(A9*C9*B9*C9)]

[0045] Similarity = AA * BB / MM

[0046] The learning unit (150) can perform learning by inputting actual input data and lifestyle patterns by month into an artificial intelligence neural network model and derive weights for each associated factor. The learning unit (150) can perform learning by inputting C1 to C9 variable values ​​corresponding to weights into the neural network model. At this time, learning can be performed between the user's actual input data and the lifestyle schedule proposed by an expert, and optimal weights can be derived for each month and problem. The above neural network model may be a deep neural network, and a deep neural network (DNN) may refer to a neural network that includes multiple hidden layers in addition to input and output layers. Using a deep neural network allows for the identification of latent structures of data. That is, it is possible to identify the latent structures of photos, text, videos, voice, and music (e.g., what object is in the photo, what is the content and emotion of the text, what is the content and emotion of the voice, etc.). Deep neural networks may include convolutional neural networks (CNN), recurrent neural networks (RNN), autoencoders, Generative Adversarial Networks (GAN), restricted Boltzmann machines (RBM), deep belief networks (DBN), Q networks, U networks, Siamese networks, Generative Adversarial Networks (GAN), etc. The description of deep neural networks described above is merely illustrative and the present disclosure is not limited thereto.

[0047] Neural networks can be trained using at least one of supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. The training of a neural network may be the process of applying knowledge to the neural network to perform a specific action.

[0048] Neural networks can be trained to minimize the error in their output. The training process involves repeatedly inputting training data into the network, calculating the error between the network's output and the target for the training data, and updating the weights of each node by backpropagating the error from the output layer to the input layer in a direction that reduces the error. In the case of supervised learning, training data is used where the correct answer is labeled for each data point (i.e., labeled training data), whereas in the case of unsupervised learning, the correct answer may not be labeled for each training data point. For instance, in the case of supervised learning for data classification, the training data may consist of data where each training point is labeled with a category. The labeled training data is input into the neural network, and the error can be calculated by comparing the network's output (category) with the label of the training data. As another example, in the case of unsupervised learning for data classification, the error can be calculated by comparing the input training data with the neural network's output. The calculated error is backpropagated in the neural network (i.e., from the output layer to the input layer), and through backpropagation, the connection weights of each node in each layer of the neural network can be updated. The amount of change in the connection weights of each node being updated can be determined by the learning rate. The neural network's calculation of the input data and the backpropagation of the error can constitute a learning cycle (epoch). The learning rate can be applied differently depending on the number of iterations of the neural network's learning cycle. For example, a high learning rate can be used in the early stages of training to quickly achieve a certain level of performance and increase efficiency, while a low learning rate can be used in the later stages to improve accuracy.

[0049] In the training of neural networks, training data is generally a subset of real-world data (i.e., the data intended to be processed using the trained neural network); therefore, a training cycle may exist where errors on the training data decrease but errors on real-world data increase. Overfitting is a phenomenon in which errors on real-world data increase due to excessive training on the training data. For example, a neural network trained on cats by showing it yellow cats may fail to recognize cats other than yellow ones as cats, which can be a type of overfitting.

[0050] Overfitting can cause an increase in errors in machine learning algorithms. Various optimization methods can be used to prevent such overfitting. To prevent overfitting, methods such as increasing the training data, regularization, dropout (which disables some nodes in the network during training), and the use of batch normalization layers can be applied.

[0051] The survey unit (160) can extract questions regarding the association factor with the highest weight from the survey DB (170) and provide them to the user terminal (200). The survey unit (160) can extract a user's lifestyle pattern schedule through general questions including the user's age and the problem to be solved, and can provide a customized lifestyle pattern schedule to the user without data input through questions regarding the association factor with the highest weight derived from an artificial intelligence learning model. The survey unit (160) can extract questions regarding the association factor with the highest weight from among multiple questions stored in the survey DB (170) and provide them to the user.

[0052] The recommendation unit (180) can recommend non-prescription commercial drugs through symptom analysis based on a large language model (LLM). Based on the survey results input from the user terminal (200), it can generate customized lifestyle patterns and non-prescription commercial drug ingredients and provide them to the user terminal. When a question is input, the LLM performs a document search based on similarity and keywords, evaluates the documents, and if it is determined that the question is not useful, it regenerates the question and performs a document search again. If the document evaluation passes, it generates an answer based on the extracted document, evaluates the answer, and if it is determined that the answer is not useful, it regenerates the answer; if the answer is useful, it passes.

[0053] The control unit (190) can control each configuration of the operating server. The control unit is a processor composed of one or more cores and may include a processor for data analysis and deep learning, such as a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), or a tensor processing unit (TPU) of a computing device. The processor can read a computer program stored in a storage space and perform data processing for learning according to one embodiment of the present disclosure. According to one embodiment of the present disclosure, the processor can perform calculations for learning a neural network. The processor can perform calculations for learning a neural network, such as processing input data for learning in deep learning (DL), extracting features from input data, calculating errors, and updating the weights of the neural network using backpropagation. At least one of the CPU, GPGPU, and TPU of the processor can process the learning of a network function. For example, the CPU and GPGPU can together process the learning of a network function and data classification using the network function. In addition, in one embodiment of the present invention, processors of a plurality of computing devices can be used together to process the learning of a network function and data classification using the network function. Furthermore, a computer program executed on a computing device according to one embodiment of the present invention may be a CPU, GPGPU, or TPU executable program.

[0054] According to the embodiment, the control unit (190) may include a feeding / sleep information receiving module, a feeding / sleep information storage module, an infant information storage module, a feeding / sleep recommended value calculation module, a feeding / sleep recommended value transmission module, an automatic alarm control module, an AI question receiving module, an AI question database 120i, an AI question interpretation module 120j, and an AI answer transmission module.

[0055] The feeding / sleep information receiving module can be configured to receive the infant's feeding / sleep information in real time from a user terminal. It can also be configured to store this information in a feeding / sleep information storage module. The feeding / sleep information includes details regarding the actual amount fed, the actual feeding time, and the actual sleep time for the day. It is information that allows for accurate determination of when and how much was fed, and when and for how long the infant slept. This feeding / sleep information can be collected by the user.

[0056] The infant information storage module is a component that stores information regarding the infant's date of birth, gender, age in life, daily weight, and daily height. Information such as the infant's date of birth and gender can be pre-stored by the user terminal, and the age in life can be automatically calculated and stored daily. Additionally, daily weight and daily height are collected and received daily through the user terminal and can be stored in the infant information storage module.

[0057] The feeding / sleep recommendation calculation module can be configured to automatically calculate the currently required feeding and sleep recommendations based on feeding / sleep information and infant data. It can calculate recommendations for feeding times based on the actual amount and time of feeding on the day. Additionally, these feeding amounts or times can be calculated by considering whether the infant is currently sleeping. Furthermore, regarding sleep duration, recommendations can be calculated by comprehensively considering factors such as how much sleep was done today and whether it is advisable to wake the infant to feed if they are currently sleeping. In particular, for infants whose day-night sleep patterns fluctuate, the system can be configured to provide sleep recommendations by considering whether to induce sleep or keep the infant awake to restore their sleep pattern to normal. These feeding and sleep recommendations may vary depending on the infant's age in weeks, current weight and height, and whether they are appropriate for the duration of life.

[0058] The feeding / sleep recommendation calculation module can be configured to automatically adjust the feeding / sleep recommendations based on whether they are within the normal range, depending on the infant's developmental status.

[0059] The feeding / sleep recommendation transmission module can be configured to transmit feeding / sleep recommendations according to the above algorithm to a user terminal in real time.

[0060] The automatic alarm control module can be configured to set achievement conditions for each step above, automatically generate and output a report on whether the achievement has been achieved, and provide it to a user terminal.

[0061] The AI ​​question receiving module can be configured to receive a user's question from a user terminal. The user terminal can be configured to receive the user's question via a microphone or touch input, etc. The AI ​​question database can be configured to store the AI ​​question received from the AI ​​question receiving module.

[0062] The AI ​​question database can be configured to link all feeding / sleep information regarding the infant's current condition, the status of compliance with recommendations, and all information regarding the infant's age, gender, weight, height, etc., in response to the AI ​​question.

[0063] The AI ​​question interpretation module can be configured to generate an AI response by synthesizing AI questions stored in the AI ​​question database, all associated feeding / sleep information, compliance status of recommendations, infant age, gender, weight, height, etc. The AI ​​response transmission module can be configured to transmit the AI ​​response to a user terminal in real time.

[0064]

[0065] FIGS. 3 and FIGS. 4 are drawings illustrating a judgment method using cosine similarity according to an embodiment of the present invention.

[0066] Referring to Figures 3 and 4, similarity can be calculated according to the developmental stage by month using the cosine similarity between the actual input data for infants and the lifestyle patterns by month.

[0067] FIGS. 5 to 9 are drawings illustrating application service operation screens according to an embodiment of the present invention.

[0068] Referring to FIGS. 5 to 9, FIG. 5 is an application service screen operating on a user terminal, which is a screen (Fig. 5(a)) that displays a recommended timetable by age according to the characteristics of infants and toddlers, and can provide a screen (Fig. 5(b)) that finds and suggests an optimal timetable when an icon is selected at the top right of the screen.

[0069] FIGS. 6 to 8 show a screen in which a survey unit extracts a question regarding the association factor with the highest weight from among multiple questions stored in a survey DB and provides it to the user in order to propose an optimal lifestyle pattern schedule. For example, questions such as "whether weaning has started," "how many times nighttime feedings have been done," and "how many times daytime feedings have been done," which are determined to be related to the association factor with the highest weight, can be asked and answers can be received from the user terminal. FIGS. 9(a) to 9(b) show a screen displaying a customized lifestyle pattern schedule generated by a recommendation unit based on the answers received from the user terminal.

[0070]

[0071] The 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 invention should be determined by the technical spirit of the appended claims.

Claims

1. Regarding methods for recommending AI-based solutions for infant and toddler health management, A step in which a data collection unit collects lifestyle pattern data including feeding data or sleep data of an infant; A step in which a lifestyle pattern extraction unit refines developmental data and lifestyle pattern data from collected lifestyle pattern data, and matches the refined lifestyle pattern data and developmental data to extract multiple lifestyle patterns by age in months; A step in which the associated factor extraction unit analyzes multiple extracted lifestyle patterns by age to extract problems and lifestyle pattern associated factors by age and lifestyle pattern; A step in which a similarity measurement unit measures the similarity between actual input data and lifestyle patterns by age; A step in which the learning unit inputs actual input data and lifestyle patterns by age into an artificial intelligence neural network model to perform learning and derives weights for each associated factor; and An AI-based solution recommendation method for infant and toddler health management that includes a step in which a recommendation unit recommends over-the-counter medications through symptom analysis based on a large language model (LLM).

2. In Paragraph 1, A step of the survey unit extracting a question regarding the association factor with the highest weight from the database and providing it to the user terminal; and An AI-based solution recommendation method for infant health management comprising the step of a recommendation unit generating customized lifestyle patterns and non-prescription commercial drug ingredients based on survey results input from a user terminal and providing them to the user terminal.

3. In Paragraph 2, The above-mentioned similarity measurement unit sets variables of standard lifestyle patterns of infants and toddlers by age in months to A1 to A9, sets variables of actual input data of the user to B1 to B9, calculates cosine similarity by applying a separate scale control table, and the scale control table can be set as weights to C1 to C9, where variables C1 to C9 are weights based on expert knowledge and frequency of major problem occurrence by age in months, and measures similarity according to the following similarity calculation formula, characterized by an artificial intelligence-based solution recommendation method for infant and toddler health management. AA = SQRT[(A1*C1) 2 +(A2*C2) 2 +...+(A9*C9) 2 ] BB = SQRT[(B1*C1) 2 +*B2*C2) 2 +... +(b9*C9) 2 ] MM = [(A1*C1*B1*C1) + (A2*C2*B2*C2)+...+(A9*C9*B9*C9)] Similarity = AA * BB / MM 4. In Paragraph 3, The above survey unit extracts a user’s lifestyle pattern timetable through general questions including the user’s age and the problem to be solved, and extracts a question regarding the association factor with the highest weight among multiple questions stored in the survey DB (170) and provides it to the user. This is an artificial intelligence-based solution recommendation method for infant health management.

5. In an artificial intelligence-based solution recommendation system for infant and toddler health management, The above operating server is, A data collection unit that collects lifestyle pattern data including feeding data or sleep data of infants and toddlers; A lifestyle pattern extraction unit that refines developmental data and lifestyle pattern data from collected lifestyle pattern data, and extracts multiple lifestyle patterns by month by matching the refined lifestyle pattern data and developmental data; An association factor extraction unit that analyzes multiple extracted lifestyle patterns by age to extract problems and lifestyle pattern association factors by age and lifestyle pattern; A similarity measuring unit that measures the similarity between actual input data and lifestyle patterns by age; A learning unit that inputs actual input data and lifestyle patterns by age into an artificial intelligence neural network model to perform learning and derives weights for each associated factor; A survey unit that extracts questions regarding the association factor with the highest weight from a database and provides them to a user terminal; and An AI-based solution recommendation system for infant and toddler health management comprising a recommendation unit that generates customized lifestyle patterns and non-prescription drug ingredients based on survey results input from a user terminal and provides them to the user terminal.

Images