Growth stage-customized growth prediction method using hybrid growth prediction model

Abstract

A growth stage-customized growth prediction method using a hybrid growth prediction model, according to embodiments of the present invention, may comprise the steps of: collecting body composition data of a subject; classifying, on the basis of the current age of the subject in months, a growth stage of the subject into any one of first to third growth intervals; reviewing whether the body composition data of the subject has been sufficiently collected; and applying, in correspondence with the classified growth interval of the subject, a hybrid growth prediction model in which a plurality of growth prediction models are integrated, to derive a predicted growth graph of the subject. The first to third growth intervals may be classified according to quantitative values of previously collected body composition data of sample subjects, the plurality of growth prediction models may include first to third growth prediction models constructed on the basis of different predictor variables, and the hybrid growth prediction model may be controlled to derive the predicted growth graph of the subject by selectively applying, from among the first to third growth prediction models, a growth prediction model corresponding to the current growth interval of the subject.

Description

Growth stage-specific growth prediction method using a hybrid growth prediction model

[0001] The present invention relates to a growth stage-tailored growth prediction method using a hybrid growth prediction model, and more specifically, to a technology for effectively monitoring and predicting a subject's growth process by classifying the subject's growth stage and deriving a growth graph by applying an integrated hybrid growth prediction model corresponding to each growth stage.

[0002] Predicting a subject's growth is recognized as an essential element for the health management of children and adolescents. While basic physical measurements such as height and weight are generally used as primary indicators for growth assessment, these single indicators alone are insufficient to adequately reflect an individual's specific growth characteristics. In particular, since the growth process is complexly influenced by an individual's genetic and environmental factors, as well as changes in body composition, more precise and personalized prediction methods are required.

[0003] Traditional growth prediction methods have primarily relied on a single growth model, applying the same criteria to all subjects or estimating growth changes using only limited variables. However, these approaches have failed to adequately reflect subjects' current growth stages or body composition data, demonstrating limitations in the accuracy and reliability of their predictions.

[0004] Recently, more sophisticated growth predictions have become possible by combining machine learning algorithms with body composition data. Body composition data provides crucial information for analyzing a subject's growth status from multiple perspectives, as it includes not only height and weight but also various factors such as muscle mass, body fat mass, protein mass, mineral mass, and basal metabolic rate. Growth prediction utilizing this data can be used to detect growth abnormalities early or as a decision-support tool for healthy growth management.

[0005] Meanwhile, currently used growth prediction technologies suffer from the problem of failing to adequately reflect the individual growth characteristics and stage-specific differences of subjects. Furthermore, applying the same growth model uniformly to all growth stages ignores the unique characteristics of each stage, resulting in reduced reliability of the prediction results. Accordingly, there is a need for a methodology that quantitatively classifies growth stages based on subjects' body composition data and applies a prediction model appropriate for each stage to derive more accurate and customized prediction results.

[0006] One of the various objectives of the present invention is to provide a method for deriving a predicted growth graph and a predicted final height optimized for a subject by classifying the subject's growth stage into first to third growth segments based on body composition data and applying different growth prediction models corresponding to each growth segment.

[0007] One of the various objectives of the present invention is to provide a technology that quantitatively classifies the growth stages of a subject and effectively derives a predicted growth graph of the subject by applying a growth prediction model suitable for each growth stage.

[0008] One of the various challenges of the present invention is to detect abnormal signs in a subject's growth process at an early stage and to provide customized information for healthy growth management.

[0009] A growth stage customized growth prediction method using a hybrid growth prediction model according to exemplary embodiments of the present invention may include the steps of: collecting body composition data of a subject; classifying the subject's growth stage into one of a first to third growth interval based on the subject's current age in months; reviewing whether sufficient body composition data of the subject has been collected; and deriving a predicted growth graph of the subject by applying a hybrid growth prediction model in which a plurality of growth prediction models are integrated, corresponding to the classified growth interval of the subject. The first to third growth intervals may be distinguished according to the quantitative values ​​of the body composition data of previously collected sample subjects, and the plurality of growth prediction models may each consist of a first to third growth prediction model constructed based on different prediction variables. The hybrid growth prediction model may be controlled to derive a predicted growth graph of the subject by selectively applying a growth prediction model among the first to third growth prediction models that corresponds to the subject's current growth interval.

[0010] The above first to third growth intervals may be distinguished using the first month age and the second month age as boundary values, wherein the first month age is the month age at which the quantitative value of the body composition data of the sample subjects is less than or equal to a pre-set first reference value, and the second month age is the month age at which the quantitative value of the body composition data of the sample subjects is greater than or equal to a pre-set second reference value.

[0011] The first growth period may be a period corresponding to the first month age or lower, the third growth period may be a period corresponding to the second month age or higher, and the second growth period may be a period corresponding to the first growth period and the second growth period.

[0012] The first growth prediction model may be a growth prediction model constructed by utilizing height, weight, BMI, and each of these quantiles as a first prediction variable dataset, the second growth prediction model may be a growth prediction model constructed by utilizing the first prediction variable dataset together with a second prediction variable dataset including protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, basal metabolic rate, and each of these quantiles, and the third growth prediction model may be a model constructed by linearly combining the first growth prediction model and the second growth prediction model.

[0013] If the current growth stage of the subject is classified as the first growth interval, and body composition data for each of the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are all missing, the step of deriving the predicted growth graph can be controlled to derive a predicted growth graph corresponding to the total months of the first to third growth intervals by applying the first growth prediction model.

[0014] When the current growth stage of the subject is classified as the first growth interval, and body composition data for each of the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are all collected, the step of deriving the predicted growth graph can be controlled to derive a predicted growth graph corresponding to the entire age in months of the first to third growth intervals by applying the third growth prediction model.

[0015] If the current growth stage of the subject is classified as the first growth interval and some body composition data among the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are missing, the step of deriving the predicted growth graph may be controlled to derive a predicted growth graph corresponding to the total age in months of the first to third growth intervals by applying the second growth prediction model.

[0016] The growth stage customized growth prediction method using the above-described hybrid growth prediction model may further include the steps of: re-collecting body composition data of the subject after a predetermined period has elapsed; reclassifying the growth stage of the subject into either the second growth segment or the third growth segment; reviewing whether sufficient body composition data for predicting the subject's growth has been collected; and re-deriving the predicted growth graph of the subject by reapplying the above-described hybrid growth prediction model.

[0017] The step of redrawing the predicted growth graph above can be controlled to redraw a predicted growth graph corresponding to some months after the second growth interval or after the third growth interval.

[0018] In the step of reviewing the collection of the body composition data, if all body composition data for each of the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are collected, the step of re-deriving the predicted growth graph may be controlled to apply the second growth prediction model, and if at least one of the body composition data for the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate is missing, the step of re-deriving the predicted growth graph may be controlled to apply the third growth prediction model.

[0019] In the case where body composition data for each of the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are all missing, the growth stage-customized growth prediction method using the hybrid growth prediction model may further include a step of comparing and reviewing the percentiles for each of the subject's height, weight, and BMI, and the percentile comparison and review step may be performed by examining whether the re-collected percentiles of the subject's height, weight, and BMI fall within the error range with the percentiles of the predicted height, predicted weight, and predicted BMI of the previously derived predicted growth graph.

[0020] In the above percentile comparison review step, if the respective percentiles of the subject's re-collected height, weight, and BMI fall within the error range of the respective percentiles of the predicted height, predicted weight, and predicted BMI of the previously derived predicted growth graph, the step of re-deriving the predicted growth graph may be controlled to re-derive the predicted growth graph by applying the same growth prediction model applied to the derivation of the previously derived predicted growth graph.

[0021] In the above percentile comparison review step, if the respective percentiles of the subject's re-collected height, weight, and BMI do not fall within the error range with the respective percentiles of the predicted height, predicted weight, and predicted BMI of the previously derived predicted growth graph, the step of re-deriving the predicted growth graph may be controlled to re-derive the predicted growth graph by applying the third growth prediction model.

[0022] The growth stage-tailored growth prediction method using the above-described hybrid growth prediction model may further include the step of comparing a predicted growth graph derived at a point in time when the subject's growth stage corresponds to the first growth interval with a predicted growth graph re-derived at a point in time when the subject's growth stage corresponds to the second growth interval or the third growth interval, and the step of analyzing the subject's growth trend based on the comparison result and modifying or additionally proposing the subject's growth management plan according to the analysis result.

[0023] The step of modifying or additionally proposing the above growth management plan may be performed by proposing at least one of a diet, exercise plan, nutritional supplementation recommendations, sleep management, mental health management, posture correction guide, activity promotion plan, participation in social activities, management of environmental factors, and periodic growth data monitoring based on the results of the analysis of the subject's growth trend.

[0024] According to the growth stage customized growth prediction method using a hybrid growth prediction model according to exemplary embodiments of the present invention, growth intervals are classified into first to third growth intervals based on quantitative values ​​derived from the subject's body composition data, and then first to third growth prediction models are applied to each of the first to third growth intervals to derive a predicted growth graph or a predicted final height.

[0025] At this time, the first growth prediction model targets subjects of relatively low age in months and is constructed using height, weight, BMI, and their respective quantiles as the first predictor variable dataset. In the case of subjects of low age in months, body composition data may not be sufficiently accumulated or growth changes may not yet be distinct, so it may not be effective for predicting final height. Therefore, the technical feature of the first growth prediction model is that it focuses on analyzing the growth trends of subjects of low age in months and predicting early growth patterns based on relatively simple variables.

[0026] On the other hand, the second growth prediction model targets subjects of high age and enables more sophisticated predictions by utilizing a second prediction variable dataset that includes protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, basal metabolic rate, and each of these quantiles in addition to the first prediction variable dataset.

[0027] Meanwhile, the third growth prediction model is constructed by linearly combining the first and second growth prediction models and is applied to subjects in the period between the first and second months. The third growth prediction model is a hybrid model designed for a natural connection between the first and second growth periods, and it can ensure the continuity and stability of the predicted growth graph by harmoniously reflecting the prediction characteristics of the two growth periods.

[0028] FIG. 1 is a configuration diagram illustrating the configuration of a growth prediction system according to exemplary embodiments of the present invention.

[0029] FIG. 2 is a flowchart illustrating a growth prediction method according to exemplary embodiments of the present invention.

[0030] FIG. 3 is an illustrative diagram for explaining the process of constructing a predictor variable data set according to exemplary embodiments of the present invention.

[0031] FIG. 4 is an illustrative diagram for explaining the process of deriving predicted body composition data according to exemplary embodiments of the present invention.

[0032] FIGS. 5 and FIGS. 6 are drawings showing predicted growth graphs according to the prior art and an embodiment of the present invention, respectively.

[0033] FIG. 7 is a flowchart illustrating a growth prediction method according to other embodiments of the present invention.

[0034] FIGS. 8 to 11 are drawings showing predicted growth graphs according to other embodiments of the present invention.

[0035] Specific embodiments of the present invention will be described below. The following detailed description is provided to facilitate a comprehensive understanding of the methods, devices, and / or systems described herein. However, this is merely illustrative and the present invention is not limited thereto.

[0036] In describing the embodiments of the present invention, detailed descriptions of known technologies related to the present invention are omitted if it is determined that such detailed descriptions may unnecessarily obscure the essence of the invention. Furthermore, the terms described below are defined in consideration of their functions within the present invention, and these may vary depending on the intentions or practices of the user or operator. Therefore, such definitions should be based on the content throughout this specification. Terms used in the detailed description are intended merely to describe the embodiments of the present invention and should not be limiting in any way. Unless explicitly stated otherwise, expressions in the singular form include the meaning of the plural form. In this description, expressions such as "include" or "comprise" are intended to refer to certain characteristics, numbers, steps, actions, elements, parts thereof, or combinations thereof, and should not be interpreted to exclude the existence or possibility of one or more other characteristics, numbers, steps, actions, elements, parts thereof, or combinations thereof other than those described.

[0037] In addition, terms such as first, second, A, B, (a), (b), etc. may be used when describing the components of the embodiments of the present invention. These terms are used merely to distinguish the components from other components, and the essence, order, or sequence of the components is not limited by the terms.

[0038] FIG. 1 is a configuration diagram illustrating the configuration of a growth prediction system according to exemplary embodiments of the present invention.

[0039] Referring to FIG. 1, the growth prediction system (100) of the present invention is composed of various components to classify the growth stage of a subject and to derive a growth graph and final height by applying a suitable prediction model according to the classified stage.

[0040] Specifically, the growth prediction system may be composed of a data processing unit (100) and a prediction model unit (200).

[0041] The data processing unit (100) may include a data collection unit (110), a growth stage classification unit (120), a predictor variable data set construction unit (130), a predicted growth graph derivation unit (140), a predicted final height derivation unit (150), a growth age estimation unit (160), and a control unit (170).

[0042] The data collection unit (110) performs the role of collecting body composition data necessary to analyze the growth stage of a subject. This data includes height, weight, BMI, muscle mass, body fat mass, protein mass, mineral mass, basal metabolic rate, etc. These data are used to evaluate the growth status of the subject from various angles and serve as basic data for growth prediction in subsequent stages. The collected data is converted into quantitative values ​​and used as input variables for a prediction model through subsequent processing stages.

[0043] The growth stage classification unit (120) classifies the growth stage of a subject into first to third growth segments based on collected body composition data. These segments are determined based on the subject's age in months and quantitative values ​​of body composition data, and each segment has different characteristics depending on the growth pattern. For example, the first growth segment is an early growth stage and includes subjects of low age in months, and the second growth segment is a rapid growth segment and includes a period when growth changes appear rapidly. The third growth segment is a stable growth segment and includes a period when growth changes gradually decrease.

[0044] The predictor variable data set construction unit (130) performs the role of generating a dataset for growth prediction. The collected body composition data consists of basic variables such as height, weight, and BMI, as well as extended variables such as protein mass, muscle mass, and body fat mass. Each variable is normalized by calculating a quantile, and this data is provided as an input variable for a prediction model used in a subsequent step.

[0045] The predicted growth graph derivation unit (140) generates a growth graph that visually represents the growth trend of the subject. This graph includes the subject's growth rate change, growth limit, and expected growth time point based on the time axis. This graph is used to intuitively understand the subject's current growth status and to predict future growth trends.

[0046] The predicted final height derivation unit (150) derives the subject's final height based on the classified growth segments and the prediction model applied to the segments. In this step, the subject's body composition data and the model's training data are combined to predict the height at the end of growth.

[0047] The growth age estimation unit (160) estimates the growth age of the subject by analyzing the relationship between the subject's actual age and the growth stage. This stage provides additional information to more precisely evaluate the subject's growth status and increase the reliability of the prediction result.

[0048] The control unit (170) integrally controls all components within the system and manages the flow and processing of data. The control unit coordinates the interactions between each component and manages the overall process so that prediction results can be generated in a timely manner.

[0049] The prediction model unit (200) may include a prediction model unit (200) comprising a first growth prediction model (210), a second growth prediction model (220), and a third growth prediction model (230) to provide appropriate prediction results according to each growth interval.

[0050] The first growth prediction model (210) uses basic variables such as height, weight, and BMI to provide prediction results suitable for the early growth period.

[0051] The second growth prediction model (220) performs a prediction suitable for the rapid growth period by including additional variables such as muscle mass, body fat mass, and protein mass in addition to the variables of the first growth prediction model (210).

[0052] The third growth prediction model (230) linearly combines the first and second growth prediction models (210, 220) to derive a prediction result in a stable growth range.

[0053] FIG. 2 is a flowchart illustrating a growth prediction method according to exemplary embodiments of the present invention.

[0054] Referring to FIG. 2, the growth prediction method of the present invention includes a series of steps for analyzing the growth stage of a subject and applying a prediction model suitable for the corresponding growth stage to derive a growth graph and a final height. Specifically, the growth prediction method of the present invention consists of a data collection step (S1), a growth stage classification step (S2), and a predicted growth graph and / or predicted final height derivation step (S3).

[0055] The data collection step (S1) is a process of collecting body composition data of a subject, and this data provides key information for growth prediction. Body composition data includes not only basic variables such as height, weight, and BMI (Body Mass Index), but also extended variables such as protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate. This data is used as basic data to evaluate the subject's growth status from various angles, and is converted into a normalized data set by calculating the quantiles of each variable.

[0056] In the growth stage classification step (S2), the growth stage of the subject can be classified into three segments based on the collected body composition data.

[0057] The first growth interval is an early growth stage of less than 1 month (e.g., 75 months), which includes subjects of lower age. Since body composition data may not be sufficiently effective for predicting final height in this interval, a first growth prediction model (210) is applied using a first prediction variable dataset consisting of height, weight, BMI, and their quantiles as input data. Through this, a predicted growth graph and / or predicted final height is derived.

[0058] The third growth stage is a late growth stage exceeding the second month of age (e.g., 90 months), and includes subjects of higher age. Since body composition data is very effective for predicting final height in this stage, in addition to the first predictor variable data set consisting of height, weight, BMI, and their quantiles, a second predictor variable data set consisting of protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, basal metabolic rate, and their quantiles is used as input. A second growth prediction model (220) using the first predictor variable data set and the second predictor variable data set as input data is used to derive a predicted growth graph and / or a predicted final height.

[0059] The second growth period is an intermediate growth stage from the first month (75 months) or older to the second month (90 months) or younger, and includes subjects of the middle age. In this period, body composition data is valid for final height prediction, but cases where both the first prediction variable data set (Height, Weight, BMI and their quantiles) and the second prediction variable data set (Protein mass, Muscle mass, Mineral mass, Body fat mass, Skeletal muscle mass, Basal metabolic rate and their quantiles) are collected are limited. Accordingly, a third growth prediction model (230) constructed by linearly combining the first growth prediction model (210) and the second growth prediction model (220) is used to derive a predicted growth graph and / or a predicted final height. The third growth prediction model (230) harmoniously combines the advantages of the two models to ensure the continuity and accuracy of the prediction.

[0060] In the step of deriving the predicted growth graph and / or final height (S3), after a suitable prediction model is applied according to the growth stage classification, a predicted growth graph that visually represents the subject's growth trend and a predicted final height at the time of growth termination are derived. The predicted growth graph includes changes in growth rate, growth limits, and the time of expected growth attainment on the time axis, and helps to intuitively understand and monitor the subject's growth status. In addition, the predicted final height quantitatively provides the expected height at the time of growth termination, serving as a decision support tool related to healthy growth management.

[0061] FIG. 3 is an illustrative diagram for explaining the process of constructing a predictor variable data set according to exemplary embodiments of the present invention.

[0062] Referring to FIG. 3, the process of constructing a predictor variable dataset according to the present invention includes steps for systematically organizing key data for predicting a subject's growth and utilizing it as input data for a prediction model. This process focuses on constructing a dataset that considers data normalization and correlations between variables in order to more precisely analyze and predict the subject's growth status.

[0063] As shown in FIG. 3, in the data collection step (S1), body composition data of the subject is collected. The body composition data consists of various variables such as gender, age in months, height, weight, BMI, protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate. The body composition data is essential for evaluating the subject's growth status from various angles and provides basic information useful for analyzing growth patterns.

[0064] For example, as illustrated in FIG. 3, when the subject is 83 months old, the height can be measured as 115.3 cm, the weight as 24.7 kg, and the BMI as 18.6, and additionally, the protein mass (3.3 kg), muscle mass (15.7 kg), mineral mass (1.1 kg), body fat mass (8.1 kg), skeletal muscle mass (7.9 kg), and basal metabolic rate (729 kcal) can be measured.

[0065] Quantiles are calculated for each variable in order to transform the collected body composition data into normalized variables. Quantiles are values ​​representing the relative position of specific data within the overall data distribution, and they contribute to improving the performance of predictive models through data normalization.

[0066] For example, in Fig. 3, the following quantile values ​​can be calculated.

[0067] Key Quantum: 0.139

[0068] Weight percentile: 0.666

[0069] BMI Quartile: 0.903

[0070] Protein mass quantile: 0.188

[0071] Muscle mass percentile: 0.166

[0072] Mineral mass quantile: 0.123

[0073] Body fat mass percentile: 0.912

[0074] Skeletal muscle mass quantile: 0.164

[0075] Basal metabolic rate quantile: 0.154

[0076] In exemplary embodiments, the respective quantiles of height, weight, and BMI can be calculated using the Korean Child and Adolescent Growth Chart (1st Growth Chart) provided by the Korea Disease Control and Prevention Agency. In one embodiment, when the subject's age in months is 83 months, height is 115.3 cm, weight is 24.7 kg, and BMI is measured as 18.6, the quantile values ​​included in the 1st predictor variable data set can be derived as the quantiles corresponding to each value (height: 0.139, weight: 0.666, BMI: 0.903) by referring to the 1st Growth Chart.

[0077] In exemplary embodiments, the quantiles of protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate can be calculated through a body composition cohort growth chart (second growth chart) constructed using previously collected growth cohort data. In one embodiment, when the subject's body composition data is measured as protein mass (3.3 kg), muscle mass (15.7 kg), mineral mass (1.1 kg), body fat mass (8.1 kg), skeletal muscle mass (7.9 kg), and basal metabolic rate (729 kcal), the quantile values ​​included in the second predictor variable data set can be derived by referring to the second growth chart as the quantiles corresponding to each value (protein mass quantile: 0.188, muscle mass quantile: 0.166, mineral mass quantile: 0.123, body fat mass quantile: 0.912, skeletal muscle mass quantile: 0.164, basal metabolic rate quantile: 0.154).

[0078] The above quantile data reflects the relative importance of each variable rather than its absolute value, and is used as a predictor variable data set that is input data for the first to third growth prediction models (210, 220, 230).

[0079] FIG. 4 is an illustrative diagram for explaining the process of deriving predicted body composition data according to exemplary embodiments of the present invention.

[0080] Referring to FIG. 4, the process of deriving predicted body composition data according to an exemplary embodiment of the present invention specifically describes the process of constructing a predictor variable data set based on the subject's body composition data and calculating a predicted growth graph and / or a predicted final height through this.

[0081] In Figure 3, for example, when the subject is 83 months old, the height is measured as 115.3 cm, the weight as 24.7 kg, and the BMI as 18.6, and additional body composition data are measured, including protein mass (3.3 kg), muscle mass (15.7 kg), mineral mass (1.1 kg), body fat mass (8.1 kg), skeletal muscle mass (7.9 kg), and basal metabolic rate (729 kcal). These data provide basic information for evaluating growth status from various angles and are used as input variables for a prediction model in subsequent steps.

[0082] The first predictor variable data set and the second predictor variable data set calculated above are used to predict body composition values ​​at the time of growth completion by assuming that each body composition data maintains the same quantile until the time of growth completion of the subject.

[0083] For example, in Figure 4, for convenience of explanation, the point of growth completion is assumed to be 90 months of age. Assuming that the current measured quantiles of each body component remain unchanged until the subject reaches the point of growth completion, the body component values ​​corresponding to those quantiles can be calculated to predict the body component data at the point of growth completion.

[0084] Specifically, based on the body composition data of the aforementioned subject, the body composition values ​​at the time of growth completion were predicted, and the data at the time of growth completion was calculated under the assumption that the quantiles of each body component remain the same. If the height quantile remains at 0.139, the height at the time of growth completion is predicted to be approximately 125 cm, and if the weight quantile is 0.666, the weight is calculated to be approximately 25.7 kg. If the BMI quantile is 0.903, the BMI at the time of growth completion is expected to be approximately 16.4.

[0085] In addition, if the protein mass quantile is maintained at 0.188, the protein mass is predicted to be approximately 4.29 kg, and if the muscle mass quantile is 0.166, the muscle mass is predicted to be approximately 20.17 kg. If the mineral mass quantile is maintained at 0.123, the mineral mass is calculated to be approximately 1.32 kg, and if the body fat mass quantile is 0.912, the body fat mass is calculated to be approximately 4.87 kg. If the skeletal muscle mass quantile is 0.164, the skeletal muscle mass is predicted to be approximately 10.54 kg, and if the basal metabolic rate quantile is maintained at 0.154, the basal metabolic rate is expected to be approximately 828.2 kcal.

[0086] In other words, the present invention relates to a customized growth prediction method that classifies a subject's growth stage into first to third growth segments based on quantitative body composition data, and derives a predicted growth graph and final height by applying a growth prediction model suitable for each growth segment. The present invention enables accurate and highly reliable prediction by considering the characteristics of the subject's growth stage and body composition data, and provides customized information necessary for healthy growth management.

[0087] The core of the present invention is to divide the growth stage of a subject into first, second, and third growth intervals using the first and second months as boundary values, and to apply a growth prediction model suitable for each interval.

[0088] Specifically, the first growth stage is an early growth stage of one month or less, and a first growth prediction model (210) is applied using a first prediction variable data set consisting of height, weight, BMI and quantiles thereof as input; the third growth stage is a late growth stage of over two months, and a second growth prediction model (220) is applied using a first prediction variable data set and a second prediction variable data set consisting of protein mass, muscle mass, body fat mass, mineral mass, skeletal muscle mass, basal metabolic rate and quantiles thereof as input; and the second growth stage is a mid-term growth stage of one month or more and two months or less, and a third growth prediction model (230) constructed by linearly combining the first and second growth prediction models is applied.

[0089] At this time, the first to third growth prediction models (210, 220, 230) may utilize machine learning algorithms (random forest, support vector machine, neural network, etc.) to use collected body composition data and growth trend data as training data. In this process, normalization of the data and correlation between variables are considered, and the model may provide prediction results optimized for each segment by reflecting the characteristics of each growth stage.

[0090] The predicted growth graph provided as the final output visualizes the subject's growth changes along a time axis, including changes in growth rate, growth limits, and the time of expected growth attainment, and can be provided to allow for an intuitive understanding of the subject's growth status. Additionally, the predicted final height provided as the final output quantitatively provides the expected height at the time of growth completion, which can be provided to offer standard information necessary for health management.

[0091] Meanwhile, the above-mentioned first to third growth prediction models (210, 220, 230) may additionally calculate a growth abnormality warning value capable of detecting the possibility of a subject's growth abnormality, and may automatically generate a warning message recommending additional examination or expert consultation if the predicted growth value exceeds a threshold. Thus, it is possible to detect a subject's growth abnormality early and support decision-making for healthy growth management.

[0092] Furthermore, during the process of collecting and analyzing body composition data, quantiles for basic variables such as height, weight, and BMI can be derived using the Korean Growth Chart for Children and Adolescents (Growth Chart 1), while quantiles for extended variables such as protein mass, muscle mass, body fat mass, mineral mass, skeletal muscle mass, and basal metabolic rate can be derived using the Body Composition Cohort Growth Chart (Growth Chart 2). The predicted body composition data calculates final body composition values ​​by assuming that the same quantiles are maintained at the time of growth completion, and based on this data, customized growth prediction results for the subject can be derived.

[0093] FIGS. 5 and FIGS. 6 are drawings showing predicted growth graphs according to a conventional method and an embodiment of the present invention, respectively. More specifically, FIG. 5 is a drawing showing a predicted growth graph according to a conventional method, and FIG. 6 is a drawing showing a predicted growth graph according to an embodiment of the present invention.

[0094] Referring to Fig. 5, the conventional predicted growth graph shows unnatural results because, even if the subject's growth stage is divided by age in months, a suitable growth prediction model is not applied to each growth segment.

[0095] Specifically, as shown in FIG. 5, the connection between the first growth section and the second growth section, and the connection between the second growth section and the third growth section, are expressed discontinuously due to a sudden change in slope. This discontinuous expression hinders the natural flow of the growth curve and can be seen as a result of not sufficiently reflecting the difference in growth rates between each section.

[0096] In particular, the discontinuity in these connections is attributed to insufficient sample size of collected data or the failure to properly apply customized growth prediction models for each growth segment classification. In other words, a smooth transition in the growth pattern could not be achieved because there was insufficient data to reflect the characteristics and speed differences of each growth segment, or because the models for each growth stage were not optimized for each segment.

[0097] That is, the conventional predicted growth graph shown in Fig. 5 exhibits a problem in that the connection between the first growth segment and the second growth segment, and the connection between the second growth segment and the third growth segment, are both expressed discontinuously due to abrupt changes in slope. This discontinuous expression may hinder the natural flow of growth changes and potentially lower the reliability of the subject's prediction results.

[0098] To address these issues, the present invention applies a technique to continuously and naturally connect growth curves. By compensating for data gaps between growth segments and reflecting differences in growth rates by segment in the modeling, the design ensures that the flow of growth changes continues smoothly.

[0099] Specifically, referring to FIG. 6, the connection between the first growth section and the second growth section, and the connection between the second growth section and the third growth section, are all expressed very naturally. Specifically, the predicted growth graph for the first growth section is derived by applying the first growth prediction model (210), and the predicted growth graph for the third growth section is derived by applying the second growth prediction model (220). Additionally, the predicted growth graph for the second growth section is derived by applying the third growth prediction model (230), which is a linear combination of the first growth prediction model (210) and the second growth prediction model (220).

[0100] Figure 6 shows the result of successfully applying the growth prediction model for each growth segment of the present invention. The change in growth rate between each segment is appropriately reflected, and the data is expressed in the form of a single continuous, smooth curve, which more accurately and reliably represents the growth trend of the subject. This natural connection of the graph was made possible through the harmony between the quantitative characteristics of the growth data and the segment-specific prediction model, and can be utilized as an effective decision support tool for managing the healthy growth of the subject.

[0101] As described above, according to the growth stage customized growth prediction method using a hybrid growth prediction model according to exemplary embodiments of the present invention, growth intervals are classified into first to third growth intervals based on quantitative values ​​derived from the subject's body composition data, and then first to third growth prediction models (210, 220, 230) are applied to each of the first to third growth intervals to derive a predicted growth graph and / or a predicted final height.

[0102] At this time, the first growth prediction model (210) targets subjects of relatively low age and is constructed using height, weight, BMI, and each of these quantiles as the first prediction variable data set. In the case of low age, body composition data is not sufficiently accumulated or growth changes are not yet clearly evident, so it may not be effective for predicting final height. Therefore, the technical feature of the first growth prediction model (210) is that it focuses on analyzing the growth trend of subjects of low age and predicting the initial growth pattern based on relatively simple variables.

[0103] On the other hand, the second growth prediction model (220) targets subjects of high age and enables more sophisticated prediction by utilizing a second prediction variable data set that includes protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, basal metabolic rate and each of these quantiles in addition to the first prediction variable data set.

[0104] Meanwhile, the third growth prediction model (230) is a model constructed by linearly combining the first growth prediction model (210) and the second growth prediction model (220), and is applied to subjects of a month age corresponding to the first month age and the second month age. The third growth prediction model (230) is a hybrid model designed for a natural connection between the first growth interval and the second growth interval, and can ensure the continuity and stability of the predicted growth graph by harmoniously reflecting the prediction characteristics of the two growth intervals.

[0105] FIG. 7 is a flowchart illustrating a growth prediction method according to other embodiments of the present invention.

[0106] Referring to FIG. 7, a growth stage customized growth prediction method using a hybrid growth prediction model according to other embodiments of the present invention may include the steps of: collecting body composition data of a subject (S10); classifying the growth stage of the subject into one of a first to third growth interval based on the subject's current age (S20); reviewing whether sufficient body composition data of the subject has been collected (S30); and deriving a predicted growth graph of the subject by applying a hybrid growth prediction model in which a plurality of growth prediction models are integrated corresponding to the classified growth interval of the subject (S40).

[0107] The first to third growth intervals may be distinguished according to the quantitative values ​​of the body composition data of the sample subjects collected, and the plurality of growth prediction models may each be composed of first to third growth prediction models (210, 220, 230) built based on different prediction variables, and the hybrid growth prediction model may be controlled to derive a predicted growth graph of the subject by selectively applying a growth prediction model among the first to third growth prediction models (210, 220, 230) that corresponds to the subject's current growth interval.

[0108] The above first to third growth intervals may be distinguished using the first month age and the second month age as boundary values, wherein the first month age is the month age at which the quantitative value of the body composition data of the sample subjects is less than or equal to a pre-set first reference value, and the second month age is the month age at which the quantitative value of the body composition data of the sample subjects is greater than or equal to a pre-set second reference value.

[0109] The first growth period may be a period corresponding to the first month age or lower, the third growth period may be a period corresponding to the second month age or higher, and the second growth period may be a period corresponding to the first growth period and the second growth period.

[0110] The first growth prediction model (210) may be a growth prediction model constructed by utilizing height, weight, BMI and each of these quantiles as a first prediction variable data set, and the second growth prediction model (220) may be a growth prediction model constructed by utilizing the first prediction variable data set together with a second prediction variable data set including protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, basal metabolic rate and each of these quantiles, and the third growth prediction model (230) may be a model constructed by linearly combining the first growth prediction model and the second growth prediction model.

[0111] In one embodiment, when the current growth stage of the subject is classified as the first growth segment and all body composition data for the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are missing, the predicted growth graph derivation step (S40) may be controlled to derive a predicted growth graph corresponding to the total months of the first to third growth segments by applying the first growth prediction model (210).

[0112] In another embodiment, when the subject's current growth stage is classified into the first growth segment and all body composition data for the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are collected, the predicted growth graph derivation step (S40) may be controlled to derive a predicted growth graph corresponding to the total months of the first to third growth segments by applying the third growth prediction model (230).

[0113] In another embodiment, if the subject's current growth stage is classified as the first growth interval and some body composition data among the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are missing, the predicted growth graph derivation step (S40) may be controlled to derive a predicted growth graph corresponding to the total months of the first to third growth intervals by applying the second growth prediction model (220).

[0114] Meanwhile, the growth stage customized growth prediction method using the hybrid growth prediction model may further include the step of re-collecting the subject's body composition data after a predetermined period has elapsed (S50), the step of reclassifying the subject's growth stage into either the second growth segment or the third growth segment (S60), the step of reviewing whether sufficient body composition data for predicting the subject's growth has been collected (S70), and the step of re-deriving the subject's predicted growth graph by reapplying the hybrid growth prediction model (S80).

[0115] At this time, the predicted growth graph re-derivation step (S80) may be controlled to re-derivate a predicted growth graph corresponding to a portion of the months after the second growth section or the third growth section. That is, the predicted growth graph re-derivation step (S80) may be controlled so as not to re-derivate a predicted growth graph for the first growth section for which a predicted growth graph has already been derived, and may be controlled to re-derivate only a predicted growth graph corresponding to a portion of the months after the second growth section or the third growth section, which is a future point in time, based on newly collected body composition data.

[0116] In exemplary embodiments, in the body composition data collection review step (S70), if all body composition data for each of the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are collected, the predicted growth graph re-derivation step (S80) may be controlled to apply the second growth prediction model (220), and if at least one of the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate body composition data is missing, the predicted growth graph re-derivation step (S80) may be controlled to apply the third growth prediction model (230).

[0117] Meanwhile, in the step of reviewing the collection of body composition data (S70), if all body composition data for the subject’s protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are missing, the growth stage customized growth prediction method using the hybrid growth prediction model may further include a step (not shown) of comparing and reviewing the percentiles for the subject’s height, weight, and BMI, respectively. The percentile comparison and review step may be performed by reviewing whether the re-collected percentiles of the subject’s height, weight, and BMI fall within the error range with the percentiles of the predicted height, predicted weight, and predicted BMI of the previously derived predicted growth graph.

[0118] In one embodiment, in the percentile comparison review step, if the respective percentiles of the subject's re-collected height, weight, and BMI fall within the error range of the respective percentiles of the predicted height, predicted weight, and predicted BMI of the previously derived predicted growth graph, the step of re-deriving the predicted growth graph may be controlled to re-derive the predicted growth graph by applying the same growth prediction model applied to the derivation of the previously derived predicted growth graph. That is, in this case, the subject's re-collected body composition data can be utilized as a valid indicator to confirm that it maintains consistency with the existing predicted growth graph, and this can be credible for the following reasons.

[0119] First, the fact that the re-collected height, weight, and BMI data of the subject fall within the quantiles and error range of the existing predicted growth graph indicates that the subject's growth trend is sufficiently accurately reflected by the existing prediction model, and this data verification procedure contributes to increasing the reliability of the growth prediction.

[0120] Furthermore, the reason for reapplying the same growth prediction model is based on the evidence that the existing model can reliably predict the subject's growth pattern, which guarantees that the existing growth prediction model can consistently reflect growth trends consistent with the collected data.

[0121] Furthermore, by comparing existing prediction models with new data, it is possible to prevent the unnecessary application of different models or complicating the prediction process. This is important for maintaining model efficiency and deriving accurate results.

[0122] Furthermore, the predicted growth graph is a useful tool for visually verifying a subject's growth trend; if the subject's re-collected data matches the existing graph, it demonstrates that the growth trend is continuous and that the existing prediction remains valid for future growth.

[0123] In addition, the review based on whether it falls within the margin of error is based on statistical methodology, and since this is carried out according to quantitative criteria rather than simple subjective judgment, it enables objective evaluation and can increase data reliability.

[0124] In addition, by reusing existing models, additional computational burden and resources associated with applying new models can be saved, thereby increasing the processing efficiency of the misunderstanding system and providing users with fast and accurate prediction results.

[0125] In another embodiment, in the percentile comparison review step, if the respective percentiles of the subject's re-collected height, weight, and BMI do not fall within the error range with the respective percentiles of the predicted height, predicted weight, and predicted BMI of the previously derived predicted growth graph, the step of re-deriving the predicted growth graph may be controlled to re-derive the predicted growth graph by applying the second growth prediction model (220). That is, in this case, to resolve the discrepancy between the subject's re-collected body composition data and the existing predicted growth graph and to provide more sophisticated and reliable prediction results, an appropriate methodology may be applied based on the following grounds.

[0126] First, the fact that the re-collected height, weight, and BMI data are inconsistent with the existing predicted growth graph by an error margin exceeding the margin of error may mean that the predictor variables or data weights used in the existing model (e.g., the first growth prediction model (210)) did not sufficiently reflect the subject's latest growth data. Therefore, applying a more advanced third growth prediction model (230) to compensate for the data inconsistency strengthens the reliability of the data processing.

[0127] At this time, the third growth prediction model (230) is a composite model constructed by linearly combining the first growth prediction model (210) and the second growth prediction model (220), and integrates the strengths of each model.

[0128] In addition, compared to the first growth prediction model (210) which uses only height, weight, BMI, and their quantiles as input variables, the third growth prediction model (230) additionally utilizes the subject's body composition data, thereby more comprehensively reflecting the major variables affecting growth. In particular, the third growth prediction model (230) can provide an optimal solution to resolve data discrepancies by linearly combining the prediction results of the first and second growth prediction models (210, 220).

[0129] In addition, the fact that the re-collected data of the subject is inconsistent with the existing prediction results may suggest that a change has occurred in the subject's growth pattern, and the third growth prediction model (230) can more accurately reflect this change by including at least partially the additional body composition data used in the second growth prediction model (220).

[0130] In addition, the linear combination method of the third growth prediction model (230) can provide a flexible structure that can utilize the mutually complementary characteristics of the first growth prediction model (210) and the second growth prediction model (220), thereby allowing the growth data of the subject to be analyzed from various angles and the correlations between various variables to be reflected in an optimized manner.

[0131] In addition, the results derived through the third growth prediction model (230) do not rely solely on height, weight, and BMI, but strengthen the connection with additional body composition data, thereby ensuring both the accuracy and reliability of the growth prediction.

[0132] Meanwhile, although not explicitly stated, the growth stage-tailored growth prediction method using the above-mentioned hybrid growth prediction model may further include the step of comparing a predicted growth graph derived at a point in time when the subject's growth stage corresponds to the first growth interval with a predicted growth graph re-derived at a point in time when the subject's growth stage corresponds to the second growth interval or the third growth interval, and the step of analyzing the subject's growth trend based on the comparison result and modifying or additionally proposing the subject's growth management plan according to the analysis result.

[0133] The step of modifying or additionally proposing the above growth management plan may be performed by proposing at least one of a diet, exercise plan, nutritional supplementation recommendations, sleep management, mental health management, posture correction guide, activity promotion plan, participation in social activities, management of environmental factors, and periodic growth data monitoring based on the results of the analysis of the subject's growth trend.

[0134] FIGS. 8 to 11 are drawings showing predicted growth graphs according to other embodiments of the present invention.

[0135] Referring to FIGS. 8 through 11, the predicted growth graph according to other embodiments of the present invention shows the connection between the first growth section and the second growth section, and the connection between the second growth section and the third growth section, all very naturally. Specifically, the predicted growth graph for the first growth section is derived by applying the first growth prediction model (210), and the predicted growth graph for the third growth section is derived by applying the second growth prediction model (220). Additionally, the predicted growth graph for the second growth section is derived by applying the third growth prediction model (230), which is a linear combination of the first growth prediction model (210) and the second growth prediction model (220).

[0136] Figure 11 shows the result of successfully applying the growth prediction model for each growth segment of the present invention. The change in growth rate between each segment is appropriately reflected, and the data is expressed in the form of a single continuous, smooth curve, which more accurately and reliably represents the growth trend of the subject. This natural connection of the graph was made possible through the harmony between the quantitative characteristics of the growth data and the segment-specific prediction model, and can be utilized as an effective decision support tool for managing the healthy growth of the subject.

[0137] As described above, according to the growth stage customized growth prediction method using a hybrid growth prediction model according to exemplary embodiments of the present invention, growth intervals are classified into first to third growth intervals based on quantitative values ​​derived from the subject's body composition data, and then first to third growth prediction models (210, 220, 230) are applied to each of the first to third growth intervals to derive a predicted growth graph and / or a predicted final height.

[0138] However, the concept of the present invention is not necessarily limited thereto, and the apparatus / method / system according to the exemplary embodiments of the present invention may be applied to various product / technology fields in addition to the aforementioned product / technology fields.

[0139] Although various embodiments of the present invention have been described in detail above, those skilled in the art will understand that various modifications can be made to the above-described embodiments without departing from the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined by the claims set forth below as well as equivalents thereof.

Claims

1. Step of collecting the subject's body composition data; A step of classifying the growth stage of the subject into one of the first to third growth intervals based on the subject's current age in months; A step of reviewing whether sufficient body composition data of the above subject has been collected; and The method includes the step of deriving a predicted growth graph of the subject by applying a hybrid growth prediction model that integrates multiple growth prediction models corresponding to the classified growth intervals of the subject. The above first to third growth stages are distinguished according to the quantitative values ​​of the body composition data of the previously collected sample subjects, and The above plurality of growth prediction models each consist of first to third growth prediction models constructed based on different prediction variables, and A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that the hybrid growth prediction model is controlled to derive a predicted growth graph of the subject by selectively applying a growth prediction model among the first to third growth prediction models that corresponds to the subject's current growth interval.

2. In Paragraph 1, The above first to third growth intervals are distinguished using the first and second months as boundary values, and A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that the first month is an age in which the quantitative value of the body composition data of the sample subjects is less than or equal to a pre-set first reference value, and the second month is an age in which the quantitative value of the body composition data of the sample subjects is greater than or equal to a pre-set second reference value.

3. In Paragraph 2, A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that the first growth interval is a interval corresponding to the first month age or lower, the third growth interval is a interval corresponding to the second month age or higher, and the second growth interval is a interval corresponding to the first growth interval and the second growth interval.

4. In Paragraph 3, The above-mentioned first growth prediction model is a growth prediction model constructed by utilizing height, weight, BMI, and their respective quantiles as the first predictor variable dataset, and The above second growth prediction model is a growth prediction model constructed by utilizing the above first prediction variable dataset together with a second prediction variable dataset including protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, basal metabolic rate, and each of the respective quantiles. A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that the third growth prediction model is a model constructed by linearly combining the first growth prediction model and the second growth prediction model.

5. In Paragraph 4, If the current growth stage of the subject is classified as the first growth interval, and all body composition data for the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are missing, A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that the step of deriving the predicted growth graph is controlled to derive a predicted growth graph corresponding to the entire monthly age of the first to third growth intervals by applying the first growth prediction model.

6. In Paragraph 4, When the current growth stage of the subject is classified as the first growth interval, and body composition data for each of the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are all collected, A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that the step of deriving the predicted growth graph is controlled to derive a predicted growth graph corresponding to the entire month age of the first to third growth intervals by applying the third growth prediction model.

7. In Paragraph 4, If the current growth stage of the subject is classified as the first growth interval, and some body composition data among the subject's protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate are missing, A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that the step of deriving the predicted growth graph is controlled to derive a predicted growth graph corresponding to the entire month age of the first to third growth intervals by applying the second growth prediction model.

8. In Paragraph 1, A step of re-collecting body composition data of the subject after a predetermined period has elapsed; A step of reclassifying the growth stage of the subject into either the second growth segment or the third growth segment; A step of reviewing whether sufficient body composition data has been collected for the growth prediction of the above-mentioned subject; and A growth stage customized growth prediction method using a hybrid growth prediction model, characterized by further including the step of re-deriving the predicted growth graph of the subject by reapplying the hybrid growth prediction model.

9. In Paragraph 8, A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that the step of re-deriving the predicted growth graph is controlled to re-derivate a predicted growth graph corresponding to a portion of the months after the second growth interval or after the third growth interval.

10. In Paragraph 9, In the review stage of the collection of the above body composition data, When all body composition data for the protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate of the subject are collected, the step of re-deriving the predicted growth graph is controlled to apply the second growth prediction model, and A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that if at least one of the body composition data of the subject’s protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate is missing, the step of re-deriving the predicted growth graph is controlled to apply the third growth prediction model.

11. In Paragraph 9, If all body composition data for protein mass, muscle mass, mineral mass, body fat mass, skeletal muscle mass, and basal metabolic rate of the above subject are missing, The method further includes the step of comparing and reviewing the percentiles for the height, weight, and BMI of the subject, respectively; A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that the above-mentioned percentile comparison review step is performed by reviewing whether the respective percentiles of the subject's re-collected height, weight, and BMI fall within the error range with the respective percentiles of the predicted height, predicted weight, and predicted BMI of the previously derived predicted growth graph.

12. In Paragraph 11, A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that, in the above-mentioned percentile comparison review step, if the respective percentiles of the subject's re-collected height, weight, and BMI fall within the error range of the respective percentiles of the previously derived predicted growth graph's predicted height, predicted weight, and predicted BMI, the above-mentioned predicted growth graph re-derivation step is controlled to re-derivate the predicted growth graph by applying the same growth prediction model applied to the derivation of the previously derived predicted growth graph.

13. In Paragraph 11, A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that, in the above-mentioned percentile comparison review step, if the respective percentiles of the subject's re-collected height, weight, and BMI are not included within the error range with the respective percentiles of the previously derived predicted height, weight, and BMI, the above-mentioned predicted growth graph re-derivation step is controlled to re-derivate the predicted growth graph by applying the above-mentioned third growth prediction model.

14. In Paragraph 8, A step of comparing a predicted growth graph derived at a point in time when the growth stage of the subject corresponds to the first growth interval with a predicted growth graph re-derived at a point in time when the growth stage of the subject corresponds to the second growth interval or the third growth interval; and A growth stage customized growth prediction method using a hybrid growth prediction model, characterized by further including the step of analyzing the growth trend of the subject based on the above comparison results, and modifying or additionally proposing a growth management plan for the subject according to the above analysis results.

15. In Paragraph 14, A growth stage customized growth prediction method using a hybrid growth prediction model, characterized in that the step of modifying or additionally proposing the growth management plan is performed by proposing at least one of a diet, exercise plan, nutritional supplementation recommendations, sleep management, mental health management, posture correction guide, activity promotion plan, participation in social activities, management of environmental factors, and periodic growth data monitoring according to the results of the growth trend analysis of the subject.

Images