A method, system and device for early warning of obesity risk in adolescents

Abstract

The present invention discloses a method, system and device for early warning of adolescent obesity risk, and the method comprises the following steps: S1, establishing an adolescent obesity monitoring early warning database R; S2, performing data collection, including basic data entry and physical fitness test; S3, performing data processing and marking on the database R; S4, establishing a training data set and a test data set for an adolescent obesity monitoring early warning model; S5, establishing an obesity risk classification prediction model, and using an improved sparrow search algorithm to optimize the submodel; S6, performing classification prediction of adolescent obesity risk. The data collection of the present invention is more convenient and efficient, and the original data set is divided into subdata sets according to age and gender and an adolescent obesity prediction submodel is established, which can improve the overall prediction accuracy.

Description

Technical Field

[0001] The present invention relates to the field of adolescent health weight risk prediction and assessment, and in particular to a method, system and device for adolescent obesity risk warning. Background Art

[0002] Currently, there is a lack of corresponding indicators and methods for weight and obesity risk assessment. The most widely used indicator, BMI, is a common standard for measuring the degree of fatness and health of the human body. However, BMI is an outcome evaluation indicator that can only reflect the current state of the test subject and cannot predict whether the currently healthy test subject will have obesity risks in the future. In the field of adolescent weight and obesity assessment, there are large differences in the physical health of adolescents of different genders and at different growth stages, making it difficult to uniformly and effectively screen potential obese people and thus provide early warning and early intervention for early obesity weight risks.

[0003] A Chinese patent application numbered CN112086190A discloses a method for monitoring and predicting adolescent obesity based on big data fitting. The method uses daily data that causes obesity and metabolic data of adolescents as input and establishes an adolescent obesity prediction model, and predicts the possibility of obesity among adolescents through the adolescent obesity prediction model.

[0004] However, the data collection cost required by this application plan is relatively high, and it is not suitable for obesity risk screening and early warning of large-scale adolescent populations. Moreover, this application plan combines the data of adolescents of all ages and genders into one model, ignoring the large differences in physical health performance in growth and development among adolescents of different genders and age groups, which will significantly reduce the overall accuracy of the model. Summary of the invention

[0005] In order to solve the problems of high data collection cost and low model accuracy in the existing adolescent obesity risk prediction, the present invention proposes a method, system and device for adolescent obesity risk warning to solve the above problems.

[0006] The present application discloses a method for early warning of adolescent obesity risk, comprising the following steps:

[0007] S1. Establish a database R for monitoring and early warning of obesity in adolescents to store basic data and physical fitness test data of adolescents;

[0008] S2, data collection, including basic data entry and physical fitness testing;

[0009] S3, processing and marking the database R;

[0010] S4. Establish training data set and test data set for adolescent obesity monitoring and early warning model;

[0011] S5. Establish an obesity risk classification prediction model and optimize the sub-model;

[0012] S6. Classify and predict adolescent obesity risk based on the model obtained in S5.

[0013] Preferably, S2 comprises the following steps:

[0014] The basic data include basic personal information: name, ID number or student ID number as a unique identifier, gender, age, grade, family weight information: parents' height and weight, height and weight of all family members living together, personal diet and exercise information: dietary characteristics, calorie intake, protein intake, number of meals per day, water intake, exercise type, exercise duration, exercise frequency, etc.;

[0015] Each physical fitness test requires the entry of basic data and the completion of all test items. The data of each physical fitness test constitutes a physical fitness test record and is saved in database R; physical fitness test items include height, weight, 1-minute sit-ups, pull-ups, vital capacity, 800 meters, 1000 meters, 1-minute rope skipping, standing long jump, sit-ups, 50-meter run, and 50*8 round-trip run.

[0016] Preferably, S3 comprises the following steps:

[0017] S31. Score the physical fitness test values ​​of all items to obtain item scores, and calculate the BMI value of the adolescent, the BMI value of the adolescent's parents, and the BMI values ​​of all family members living together;

[0018] S32. According to the BMI value, weight is divided into normal, underweight, overweight and obese;

[0019] S33, adding a field for marking the weight classification attribute of the adolescent, the adolescent's parents and all family members living together to the data, and the field is one of normal, underweight, overweight and obese;

[0020] S34. Add a field to the data to record the number of obese people among the teenagers’ parents and all family members living with them.

[0021] Preferably, S4 comprises the following steps:

[0022] S41, let the number of teenagers participating in the test in database R be a, let the qth teenager among a teenagers have p in database R. (q) records, for this p (q) The records are sorted according to the test time;

[0023] S42, record the test time t1 as row vector at1 , the record of the next test time t2 is row vector a t2 , the data changes of all numeric fields except the unchanged fields during the test time t1 to test time t2 are row vectors

[0024] S43. Merge Δa and a t2 And remove the unchanged fields to get a row vector Δa′ of length n. If the weight of the adolescent changes from non-obese to obese, from normal to overweight, or from underweight to overweight during the period from t1 to t2, then the supervised learning label corresponding to the model training of Δa′ is marked as 1, and the rest are marked as 0.

[0025] S44, for all p (q) records to perform the calculation Δa′, composed of (p (q) -1)*n-dimensional matrix A q , merge the matrices A corresponding to all teenagers q , we get a matrix consisting of m rows and n columns of data and the corresponding label vector, where

[0026] S45. Divide the matrix and label vector obtained in S44 into d different data sets according to age and gender, and perform normalization processing. Then randomly divide the normalized data to obtain d groups of training data sets and test data sets, where d = b*c, b is the number of age value types, and c = 2 is the number of gender value types.

[0027] Preferably, S5 comprises the following steps:

[0028] According to the d sets of training data sets and test data sets obtained in S45, an obesity risk classification prediction model {M1, M2, …, M i , …, M d}, and use the improved sparrow search algorithm to optimize the sub-model M through the following steps i :

[0029] S51, initialize the parameters of the sparrow search algorithm, including the number of sparrows N, the maximum number of iterations T, the model penalty regularization term parameter λ, and the nonlinear mapping function f θ The parameter vector θ of (x) is used to randomly generate a sparrow population based on the following formula:

[0030] X (N,D) =rand(N,D)×(u c,d -l c,d )+l c,d ;

[0031] Among them, (N, D) represents the search space of the sparrow population, D represents the dimension of the optimization problem, X represents the position of the sparrow, and u c,d and l c,d Respectively represent the upper and lower limits of the search space;

[0032] S52. Calculate the initial fitness of N sparrow individuals using the following formula:

[0033]

[0034]

[0035] Among them, m represents the number of samples in the training data set, θ j Represents the nonlinear mapping function f θ The jth coefficient of (x), represents the probability that the kth sample is a positive example, and n is the nonlinear mapping function f θ The number of coefficients of (x), y (k) is the true label of the kth sample data;

[0036] S53, dividing the population into discoverers and followers, calculating the initial fitness of all sparrow individuals according to S52, and sorting them from small to large in terms of fitness, taking the individuals in the first r1% as discoverers, and the rest as followers;

[0037] S54, update the discoverer position, the update formula is:

[0038]

[0039] in, represents the d-dimensional position of the i-th individual in the t-th generation in the population, Q is a random number from a standard normal distribution, R2 is a uniform random number in [0,1], ST is the warning threshold, and its value range is [0.5,1.0];

[0040] S55, update the follower position, the update formula is:

[0041]

[0042] in, is the position of the sparrow with the best position in the current iteration;

[0043] S56. Randomly select r2% of individuals from the population as sentinels and update their positions using the following formula:

[0044]

[0045] Among them, β is a random number that conforms to the standard normal distribution, f iis the fitness value of the current sparrow individual, f g is the fitness value of the individual sparrow with the best position in the current iteration, is the position of the sparrow with the worst position in the current iteration;

[0046] S57, repeat S53-S56 until the iteration ends, output the optimized parameters of the model, input the obtained parameters into the obesity risk classification prediction model, and use the test data set to verify the model, and calculate the actual prediction accuracy Verify the classification accuracy of the model and compare it with the preset accuracy threshold r. The model training is completed and can be used for adolescent obesity risk classification prediction. Otherwise, it is necessary to recheck the sample data or modify the parameters λ and θ in S51 to retrain the model.

[0047] Preferably, S6 comprises the following steps:

[0048] If the field used to mark the adolescent weight classification attribute obtained in S3 is obesity, then "obesity" is output; otherwise, the model input data corresponding to the adolescent is obtained according to S4, and the data is input into the corresponding adolescent obesity risk prediction sub-model M according to the age and gender in the input data. i In S52, the probability p corresponding to the data is calculated using the formula in S52. If p>0.5, the prediction result is marked as 1, and the output prompts that the adolescent "is at risk of obesity."

[0049] The present application discloses a teenager obesity risk warning device, which executes the teenager obesity risk warning method.

[0050] Preferably, the adolescent obesity risk warning device includes a data acquisition module, a data processing module and a risk prediction module.

[0051] The present application discloses a teenager obesity risk warning system, which includes the teenager obesity risk warning device.

[0052] Preferably, the adolescent obesity risk warning system also includes an intelligent terminal and multiple physical fitness testing devices.

[0053] Beneficial effects of the present invention:

[0054] (1) The present invention uses adolescent physical health test data and change data per unit time as feature data of the obesity prediction model training data set. The data required by the model can be automatically collected and uploaded by various test devices without the need for a large amount of manpower support. Data collection is more convenient and efficient, and can be applied to obesity risk screening scenarios for large-scale adolescent populations.

[0055] (2) The present invention takes into account the differences in physical health performance of adolescents of different ages and genders in their growth and development, divides the original data set into sub-datasets according to age and gender, and establishes an adolescent obesity prediction sub-model, which can improve the overall prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a schematic diagram of the process of predicting obesity risk in adolescents according to an embodiment of the present invention;

[0057] Figure 2 Schematic diagram of the structure of the adolescent obesity risk prediction system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0058] In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and examples.

[0059] The present application embodiment discloses a method for early warning of obesity risk in adolescents, the method flow is as follows Figure 1 As shown, the following steps are included:

[0060] S1. Establish a database R for monitoring and early warning of obesity in adolescents to store basic data and physical fitness test data of adolescents;

[0061] S2. Data collection, including basic data entry and physical fitness test; basic data is collected through intelligent terminal entry (basic data must be filled in during the first physical fitness test, and can be skipped if there is no change during the second physical fitness test). Basic data includes personal basic information: name, ID number or student ID number as a unique identifier, gender, age, grade, family weight status: parents' height and weight, height and weight of all family members living together, personal daily diet and exercise status: dietary characteristics, calorie intake, protein intake, number of meals per day, water intake, exercise type, exercise duration, exercise frequency, etc.;

[0062] The same adolescent can take multiple tests at any different time. Each physical test requires the entry of basic data and the completion of all test items. The data of each physical test constitutes a physical test record and is saved in database R. The physical test is collected through automated physical test equipment. The test items include height, weight, 1-minute sit-ups, pull-ups, vital capacity, 800 meters, 1000 meters, 1-minute rope skipping, standing long jump, sit-ups, 50-meter run, and 50*8 round-trip run. During the test, multiple corresponding items are tested according to the grade of the adolescent being tested.

[0063] Students of different ages and genders were tested as shown in Table 1. "Male, female" in the table means that both male and female students of the same age will be tested on the item, "male" means that male students of the same age will be tested on the item while female students will not participate, "female" means that female students of the same age will be tested on the item while male students will not participate, and blank means that neither male nor female students of the same age will participate in the item. Multiple corresponding items were tested.

[0064] Table 1 Test items for each age and gender

[0065]

[0066]

[0067] S3. After collecting the above basic data and physical fitness test data from a large range of teenagers, the database R is processed and marked, including the following steps:

[0068] S31. Score the physical fitness test values ​​of all items to obtain item scores, calculate the BMI value of the adolescent, the BMI value of the adolescent's parents, and the BMI values ​​of all family members living together, and the calculation formula of the BMI value is: BMI = weight / height squared (weight unit: kilograms, height unit: meters);

[0069] S32. According to the BMI value, the body weight is divided into normal, underweight, overweight and obese. The classification standards for different genders and ages are shown in Table 2 and Table 3:

[0070] Table 2 Classification standards of body mass index (BMI) for boys (unit: kg / m 2 )

[0071]

[0072]

[0073] Table 3 Classification standards of body mass index (BMI) for girls (unit: kg / m 2 )

[0074] age normal low body weight overweight obesity Aged 6 and under 7 13.3～17.3 ≤13.2 17.4～19.2 ≥19.3 Aged 7 and under 8 13.5～17.8 ≤13.4 17.9～20.2 ≥20.3 Aged 8 or above but under 9 years old 13.6～18.6 ≤13.5 18.7～21.1 ≥21.2 Aged 9 and under 10 13.7～19.4 ≤13.6 19.5～22.0 ≥22.1 Aged 10 or above but under 11 13.8～20.5 ≤13.7 20.6～22.9 ≥23.0 Aged 11 or above but under 12 14.2～20.8 ≤14.1 20.9～23.6 ≥23.7 Aged 12 or above but under 13 14.8～21.7 ≤14.7 21.8～24.4 ≥24.5 Aged 13 or above but under 14 15.3～22.2 ≤15.2 22.3～24.8 ≥24.9 Aged 14 or above but under 15 16.0～22.6 ≤15.9 22.7～25.1 ≥25.2 Aged 15 or above and under 16 16.5～22.7 ≤16.4 22.8～25.2 ≥25.3 Aged 15 or above and under 16 16.9～23.2 ≤16.8 23.3～25.4 ≥25.5 Aged 15 or above and under 16 17.1～23.3 ≤17.0 23.4～25.7 ≥25.8 18 years and above 17.2～23.9 ≤17.1 24.0～27.9 ≥28.0

[0075] S33, adding a field for marking the weight classification attribute of the adolescent, the adolescent's parents and all family members living together to the data, and the field is one of normal, underweight, overweight and obese;

[0076] S34. Add a field to the data to record the number of obese people among the teenagers’ parents and all family members living with them.

[0077] S4. Establish training data set and test data set for adolescent obesity monitoring and early warning model;

[0078] S41, let the number of teenagers participating in the test in database R be a, let the qth teenager among a teenagers have p in database R. (q) Records (p (q) >1), for this p (q) The records are sorted according to the test time;

[0079] S42, record the record of an arbitrary test time t1 as row vector a t1 , the record of the next test time t2 is row vector a t2 , from test time t1 to test time t2, except for the unchanged fields such as name and gender, the data changes of other numeric fields are row vectors

[0080] S43. Merge Δa and a t2 And remove the unchanged fields such as name, student ID or ID number to get a row vector Δa′ of length n. If the weight of the adolescent changes from non-obese to obese, from normal to overweight, or from underweight to overweight during the period from t1 to t2, then the supervised learning label corresponding to the model training of Δa′ is marked as 1, and the rest are marked as 0;

[0081] S44, for all p (q) records to perform the calculation Δa′, composed of (p (q) -1)*n-dimensional matrix A q , merge the matrices A corresponding to all teenagers q , we get a matrix consisting of m rows and n columns of data and the corresponding label vector, where

[0082] S45. Divide the matrix and label vector obtained in S44 into d different data sets according to age and gender, and perform normalization processing. Then randomly divide the normalized data to obtain d groups of training data sets and test data sets, where d = b*c, b is the number of age value types, and c = 2 is the number of gender value types.

[0083] S5. According to the d groups of training data sets and test data sets obtained in S45, an obesity risk classification prediction model {M1, M2, ..., M i , …, M d}, and optimize the sub-model M by the following steps i :

[0084] S51, initialize the parameters of the sparrow search algorithm, including the number of sparrows N, the maximum number of iterations T, the model penalty regularization term parameter λ, and the nonlinear mapping function f θ The parameter vector θ of (x) is used to randomly generate a sparrow population based on the following formula:

[0085] X(N,D) =rand(N,D)×(u c,d -l c,d )+l c,d ;

[0086] Among them, (N, D) represents the search space of the sparrow population, D represents the dimension of the optimization problem, X represents the position of the sparrow, and u c,d and l c,d Respectively represent the upper and lower limits of the search space;

[0087] S52. Calculate the initial fitness of N sparrow individuals using the following formula:

[0088]

[0089] Among them, m represents the number of samples in the training data set, θ j Represents the nonlinear mapping function f θ The jth coefficient of (x), represents the probability that the kth sample is a positive example, and n is the nonlinear mapping function f θ The number of coefficients of (x), y (k) is the true label of the kth sample data;

[0090] S53, dividing the population into discoverers and followers, calculating the initial fitness of all sparrow individuals according to S52, and sorting them from small to large in terms of fitness, taking the individuals in the first r1% as discoverers, and the rest as followers;

[0091] S54, update the discoverer position, the update formula is:

[0092]

[0093] in, represents the d-dimensional position of the i-th individual in the t-th generation in the population, Q is a random number from a standard normal distribution, R2 is a uniform random number in [0,1], ST is the warning threshold, and its value range is [0.5,1.0];

[0094] S55, update the follower position, the update formula is:

[0095]

[0096] in, is the position of the sparrow with the best position in the current iteration;

[0097] S56. Randomly select r2% of individuals from the population as sentinels and update their positions using the following formula:

[0098]

[0099] Among them, β is a random number that conforms to the standard normal distribution, f i is the fitness value of the current sparrow individual, f g is the fitness value of the individual sparrow with the best position in the current iteration, is the position of the sparrow with the worst position in the current iteration;

[0100] S57, repeat S53-S56 until the iteration ends, output the optimized parameters of the model, input the obtained parameters into the obesity risk classification prediction model, and use the test data set to verify the model, and calculate the actual prediction accuracy Verify the classification accuracy of the model and compare it with the preset accuracy threshold r. The model training is completed and can be used for adolescent obesity risk classification prediction. Otherwise, it is necessary to recheck the sample data or modify the parameters λ and θ in S51 to retrain the model.

[0101] S6. Classify and predict adolescent obesity risk based on the model obtained in S5. The prediction process is as follows: Figure 1 As shown, if the field used to mark the adolescent weight classification attribute obtained by S3 is obesity, then "obesity" is output; otherwise, the model input data corresponding to the adolescent is obtained according to S4, and the data is input into the corresponding adolescent obesity risk prediction sub-model M according to the age and gender in the input data. i In S52, the probability p corresponding to the data is calculated using the formula in S52. If p>0.5, the prediction result is marked as 1, and the output prompts that the adolescent "is at risk of obesity."

[0102] In one embodiment, a device for early warning of adolescent obesity risk is disclosed. The device executes the above-mentioned method for early warning of adolescent obesity risk, and includes a data collection module, a data processing module and a risk prediction module.

[0103] In one embodiment, a youth obesity risk early warning system is disclosed, such as Figure 2 As shown, the adolescent obesity risk warning system includes the adolescent obesity risk warning device mentioned above, and also includes an intelligent terminal and multiple automated physical fitness testing devices. The automated physical fitness testing devices include a height and weight tester, a 50-meter run tester, a 50*8 round-trip run tester, a 1-minute sit-up tester, a pull-up tester, a vital capacity tester, an 800 / 1000-meter tester, a 1-minute skipping tester, a standing long jump tester, and a sit-and-reach tester.

[0104] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for early warning of obesity risk in adolescents, characterized in that: The following steps are involved: S1. Establish a database R for monitoring and early warning of obesity in adolescents to store basic data and physical fitness test data of adolescents; S2, data collection, including basic data entry and physical fitness testing; S3, processing and marking the database R; S4. Establish training data set and test data set for adolescent obesity monitoring and early warning model; S41, let the number of teenagers participating in the test in database R be a, let the qth teenager among a teenagers have p in database R. (q) records, for this p (q) The records are sorted according to the test time; S42, record the test time t1 as row vector a t1 , the record of the next test time t2 is row vector a t2 , the data changes of all numeric fields except the unchanged fields during the test time t1 to test time t2 are row vectors S43. Merge Δa and a t2 And remove the unchanged fields to get a row vector Δa′ of length n. If the weight of the adolescent changes from non-obese to obese, from normal to overweight, or from underweight to overweight during the period from t1 to t2, then the supervised learning label corresponding to the model training of Δa′ is marked as 1, and the rest are marked as 0. S44, for all p (q) records to perform the calculation Δa′, composed of (p (q) -1)*n-dimensional matrix A q , merge the matrices A corresponding to all teenagers q , we get a matrix consisting of m rows and n columns of data and the corresponding label vector, where S45, dividing the matrix and label vector obtained in S44 into d different data sets according to age and gender for normalization, and then randomly dividing the normalized data to obtain d sets of training data sets and test data sets, where d = b*c, b is the number of age value types, and c = 2 is the number of gender value types; S5. Establish an obesity risk classification prediction model and optimize the sub-model; According to the d sets of training data sets and test data sets obtained in S45, an obesity risk classification prediction model {M1, M2, …, M i , …, M d }, and optimize the sub-model M by the following steps i : S51, initialize the parameters of the sparrow search algorithm, including the number of sparrows N, the maximum number of iterations T, the model penalty regularization term parameter λ, and the nonlinear mapping function f θ The parameter vector θ of (x) is used to randomly generate a sparrow population based on the following formula: X (N,D) =rand(N,D)×(u c,d -l c,d )+l c,d ; Among them, (N, D) represents the search space of the sparrow population, D represents the dimension of the optimization problem, X represents the position of the sparrow, and u c,d and l c,d Respectively represent the upper and lower limits of the search space; S52. Calculate the initial fitness of N sparrow individuals using the following formula: Among them, m represents the number of samples in the training data set, θ j Represents the nonlinear mapping function f θ The jth coefficient of (x), represents the probability that the kth sample is a positive example, and n is the nonlinear mapping function f θ The number of coefficients of (x), y (k) is the true label of the kth sample data; S53, dividing the population into discoverers and followers, calculating the initial fitness of all sparrow individuals according to S52, and sorting them from small to large in terms of fitness, taking the individuals in the first r1% as discoverers, and the rest as followers; S54, update the discoverer position, the update formula is: in, represents the d-dimensional position of the i-th individual in the t-th generation in the population, Q is a random number from a standard normal distribution, R2 is a uniform random number in [0,1], ST is the warning threshold, and its value range is [0.5,1.0]; S55. Update the follower position. The update formula is: in, is the position of the sparrow with the best position in the current iteration; S56. Randomly select r2% of individuals from the population as sentinels and update their positions using the following formula: Among them, β is a random number that conforms to the standard normal distribution, f i is the fitness value of the current sparrow individual, f g is the fitness value of the individual sparrow with the best position in the current iteration, is the position of the sparrow with the worst position in the current iteration; S57, repeat S53-S56 until the iteration ends, output the optimized parameters of the model, input the obtained parameters into the obesity risk classification prediction model, and use the test data set to verify the model, and calculate the actual prediction accuracy Verify the classification accuracy of the model and compare it with the preset accuracy threshold r. If the model training is completed, it can be used for adolescent obesity risk classification prediction. Otherwise, it is necessary to recheck the sample data or modify the parameters λ and θ in S51 to retrain the model. S6. Classify and predict adolescent obesity risk based on the model obtained in S5.

2. The method for early warning of adolescent obesity risk according to claim 1, characterized in that: The S2 comprises the following steps: The basic data include personal basic information, family weight information, and personal diet and exercise information; Each physical fitness test requires the entry of basic data and the completion of all test items. The data of each physical fitness test constitutes a physical fitness test record and is saved in database R.

3. A method for early warning of adolescent obesity risk according to claim 2, characterized in that: The S3 comprises the following steps: S31. Score the physical fitness test values ​​of all items to obtain item scores, and calculate the BMI value of the adolescent, the BMI value of the adolescent's parents, and the BMI values ​​of all family members living together; S32. According to the BMI value, weight is divided into normal, underweight, overweight and obese; S33, adding a field for marking the weight classification attribute of the adolescent, the adolescent's parents and all family members living together to the data, and the field is one of normal, underweight, overweight and obese; S34. Add a field to the data to record the number of obese people among the teenagers’ parents and all family members living with them.

4. The method for early warning of adolescent obesity risk according to claim 3, characterized in that: The S6 comprises the following steps: If the field used to mark the adolescent weight classification attribute obtained by S3 is obesity, then "obesity" is output; otherwise, the model input data corresponding to the adolescent is obtained according to S4, and the data is input into the corresponding adolescent obesity risk prediction sub-model M according to the age and gender in the input data. i In S52, the probability p corresponding to the data is calculated using the formula in S52. If p>0.5, the prediction result is marked as 1, and the output prompts that the adolescent "is at risk of obesity." 5. A device for early warning of adolescent obesity risk, characterized in that: The adolescent obesity risk warning device executes the adolescent obesity risk warning method as described in any one of claims 1-4.

6. The adolescent obesity risk warning device according to claim 5, characterized in that: It includes data acquisition module, data processing module and risk prediction module.

7. A teenager obesity risk early warning system, characterized in that: The adolescent obesity risk warning system includes the adolescent obesity risk warning device described in claim 6.

8. The adolescent obesity risk warning system according to claim 7, characterized in that: It also includes smart terminals and multiple physical fitness testing devices.

Images