A method and system for generating balance and coordination training strategies based on machine learning

Through a machine learning-based method, using deep convolutional neural networks and long and short-term memory networks, a balanced coordination training strategy generation model is built, which solves the problem of lack of personalization and intelligence in the existing technology, and improves the elderly's Balance and coordination ability and training efficiency.

CN118588238BActive Publication Date: 2024-12-06BEIJING VOCATIONAL COLLEGE OF SOCIAL MANAGEMENT
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202410772607.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-17
Publication Date
2024-12-06
Estimated Expiration
2044-06-17

AI Technical Summary

Technical Problem

The existing balance and coordination training methods lack personalization and intelligence, and are difficult to meet the specific needs of different elderly people. They cannot effectively improve the balance and coordination ability of the elderly, which increases the risk of falls and injuries.

Method used

Using a machine learning-based method, by collecting user's motion data and medical health data, preprocessing and building a balanced coordination training strategy generation model, using deep convolutional neural networks and long and short-term memory networks to extract motion characteristics and facial emotions characteristics, and calculate balance Coordinate contribution rate and preferences, optimize training strategies, and generate personalized balanced and coordinated training solutions.

Benefits of technology

The accuracy and accuracy of the balanced coordination training strategy are improved, and personalized customization for different elderly people is achieved, adapting to different needs, saving resources, improving work efficiency, and real-time action correction and customization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118588238B_ABST
    Figure CN118588238B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for generating a balance coordination training strategy based on machine learning, including collecting the user's motion data and medical health data, preprocessing the motion data, formulating a scientific balance coordination plan through the medical health data, the motion data including daily motion data and balance coordination motion data, obtaining a fixed balance coordination value according to the daily motion data, screening the balance coordination motion data based on balance coordination preference to obtain a balance coordination action, using the balance coordination action to adjust the scientific balance coordination plan to obtain an optimized balance coordination plan, constructing a balance coordination training strategy generation model according to the fixed balance coordination value and the optimized balance coordination plan, and optimizing the balance coordination training strategy generation model. This method can not only improve the accuracy of generating a balance coordination training strategy based on machine learning, but also has good interpretability, and can be directly applied to a balance coordination training strategy generation system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of strategies, and in particular to a method and system for generating a balance and coordination training strategy based on machine learning. Background Art

[0002] As they age, the physical functions of the elderly gradually decline, and their balance and coordination abilities also weaken, increasing their risk of falling and injury. Existing balance and coordination training methods are mostly based on traditional physical training. Although they can improve the balance ability of the elderly to a certain extent, they lack personalization and intelligence and are difficult to meet the specific needs of different elderly people.

[0003] In recent years, machine learning technology has achieved remarkable results in many fields, especially in the field of medical health, where it can provide personalized solutions based on the specific circumstances of individuals. Therefore, applying machine learning technology to balance and coordination training for the elderly has broad application prospects and important practical significance. Then, how to optimize the current balance and coordination training strategy based on personalized balance and coordination characteristics becomes a problem to be solved. Summary of the invention

[0004] The purpose of the present invention is to provide a method for generating a balance and coordination training strategy based on machine learning.

[0005] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0006] The present invention comprises the following steps:

[0007] Collect the user's exercise data and medical health data, pre-process the exercise data, and formulate a scientific balance and coordination plan based on the medical health data; the exercise data includes daily exercise data and balance and coordination exercise data;

[0008] Obtaining a fixed balance coordination value according to the daily exercise data, and filtering the balance coordination exercise data based on the balance coordination preference to obtain a balance coordination action;

[0009] The scientific balance coordination scheme is adjusted by the balance coordination action to obtain an optimized balance coordination scheme, and a balance coordination training strategy generation model is constructed according to the fixed balance coordination value and the optimized balance coordination scheme;

[0010] The balance and coordination training strategy generation model is optimized, the data to be generated is input into the optimized balance and coordination training strategy generation model, and the balance and coordination training strategy is output.

[0011] Furthermore, the method for obtaining a fixed balance value according to the daily exercise data includes:

[0012] Static balance test and muscle activity analysis were used to obtain the balance and coordination effects of daily exercise data. The balance and coordination effects were weighted and summed according to the importance of position balance to obtain a comprehensive fixed balance and coordination value.

[0013] Furthermore, the method of obtaining the balancing and coordination action by screening the balancing and coordination movement data based on the balancing and coordination preference includes:

[0014] Deep convolutional neural network and long short-term memory network are used to extract motion features and facial emotion features of balance and coordination motion data;

[0015] Obtain facial emotion scores based on muscle activity analysis and calculate the eigenvalues ​​of motion features:

[0016]

[0017] The number of features is n, and the characteristic value of the w-th motion feature is The wth motion feature of the cth attribute is The standard value of the wth motion feature of the cth attribute is The average value of the motion feature is The number of attributes is p, the distance of attribute data is d, and the control factor of the wth motion feature of the cth attribute is α w,c ;

[0018] Calculate the contribution rate of Hengxie:

[0019]

[0020] The balance contribution rate of the wth motion feature is The balancing effect of the wth motion feature is h w , the coordination error value of the w-th motion feature is σ w ;

[0021] Calculate the balance preference of motion features:

[0022]

[0023] The trade-off preference of the w-th motion feature is The initial facial emotion score of the w-th motion feature and the ending facial emotion score of the w-th motion feature are C w,o , the natural constant is e, and the emotion score error of the w-th motion feature is τ w , the membership degree of the rth attribute of the wth motion feature is z w (r), the number of attributes is N, and the fuzzy element of the k-th membership degree is s κ , the fuzzy element of the ιth membership degree is s ι ;

[0024] The motion feature coordination preferences of the actions are summed to obtain the action coordination preference, and the actions with action coordination preference greater than 0.347 are output as coordinated actions.

[0025] Furthermore, the method of using the balancing action to adjust the scientific balancing scheme to obtain an optimized balancing scheme includes:

[0026] Calculate the composite score for the alternative moves:

[0027]

[0028] The comprehensive score of the i-th alternative action is The contribution rate weight is ρ 1 , the sentiment weight is ρ 2 , the balance contribution rate of the ith alternative action is The initial facial emotion score C of the i-th alternative action i,o , the ending facial emotion score of the ith alternative action is C i,en ;

[0029] Sort the alternative actions in descending order according to the comprehensive scores, and use the first alternative action to replace the non-balancing action in the scientific balancing solution to obtain the optimized action;

[0030] According to the goal of the scientific balance plan, the training time of the optimized action is adjusted to obtain the optimized balance plan.

[0031] Furthermore, a method for constructing a balance coordination training strategy generation model according to the fixed balance value and the optimized balance scheme includes:

[0032] The target coordination amount is obtained by subtracting the fixed coordination value from the target of the scientific coordination plan, and the training time of the optimized coordination plan is adjusted according to the target coordination amount to obtain the training plan;

[0033] Use scoring to obtain user satisfaction and calculate the training strategy score:

[0034]

[0035] The scientific coordination solution goal of the jth user is K j , the fixed trade-off value for the jth user is The first weight is The second weight is The error value of the jth user is The training strategy score of the jth user is

[0036] The objective function of the balanced coordination training strategy generation model is constructed according to the training strategy function, and the expression is:

[0037]

[0038] The objective function of the jth user is Training strategy score and predict the training strategy score The loss function is

[0039] The balance coordination training strategy generation model includes random forest algorithm, convolutional neural network algorithm, long short-term memory network algorithm, spectral clustering algorithm, and Q learning;

[0040] The random forest algorithm divides the input data into training and test sets in a ratio of 4:1;

[0041] The convolutional neural network algorithm automatically learns the motion feature representation of the training set to obtain the balance motion feature, and then fuses the balance motion features according to the action to obtain the balance action.

[0042] The long short-term memory network algorithm obtains the balance coordination effect value according to the time change of the balance coordination action on the user;

[0043] Spectral clustering algorithm, using the balance effect value to cluster the balance actions and obtain the action classification;

[0044] Q-learning generates a balanced coordination training policy by learning the best policy for the agent to take actions in a given state.

[0045] Furthermore, the method for optimizing the balance and coordination training strategy generation model includes:

[0046] Randomly divide the particle population into two equal sub-populations and calculate the fitness of the particles:

[0047]

[0048] The fitness of the ath particle is The objective function value of the ath particle is The number of constraints is m, and the penalty factor of the u-th constraint is β u , the u-th constraint function of the a-th particle is k u (a), the predicted value of the objective function is The variable control constant is b;

[0049] Sort the particles of the sub-populations in descending order according to their fitness, take the particle with the best fitness in the first sub-population as the first individual, and take the particle with the best fitness in the second sub-population as the second individual;

[0050] Calculate the first and second parameters:

[0051]

[0052]

[0053] The maximum number of iterations is t max , the number of iterations is t, and the first parameter is ψ 1 , the second parameter is ψ 2 , the interference factor is μ, the number of particles in the subpopulation is n, and the influence constant is c;

[0054] When the first parameter is greater than or equal to 0.25 and the second parameter is greater than 0.6, update the particle position:

[0055]

[0056]

[0057] The position of the ath main particle in the t+1th iteration is S a,y (t+1), the position of the ath main particle in the tth iteration is S a,y (t), the first random number is v 1 , the second random number is v 2 , the optimal position of the main particle in the tth iteration is S y,bs (t), the worst position of the main particle at the tth iteration is S y,wt (t), the position of the ath particle in the t+1th iteration is S a,f (t+1), the position of the ath particle in the tth iteration is S a,f (t), the optimal position of the secondary particle is S f,bs , the worst position of the secondary particle is S f,wt ;

[0058] If the random number is greater than 0.6, update the particle position:

[0059]

[0060]

[0061] The position of the ath main particle in the t+1th iteration is The position of the ath particle in the t+1th iteration is

[0062] On the contrary, iterate continuously until the update conditions are met and output the optimal position.

[0063] In a second aspect, a system for generating a balance and coordination training strategy based on machine learning includes:

[0064] Data collection module: used to collect the user's sports data and medical health data, pre-process the sports data, and formulate a scientific coordination plan based on the medical health data;

[0065] Data processing module: used for obtaining fixed balance coordination values ​​according to the daily exercise data, and filtering the balance coordination exercise data based on balance coordination preferences to obtain balance coordination actions;

[0066] Model building module: used to use the balance coordination action to adjust the scientific balance coordination scheme to obtain an optimized balance coordination scheme, and to build a balance coordination training strategy generation model according to the fixed balance coordination value and the optimized balance coordination scheme;

[0067] An optimization output module is used to optimize the balance coordination training strategy generation model, input the data to be generated into the optimized balance coordination training strategy generation model, and output the balance coordination training strategy. In a second aspect, the embodiment of the present application further provides an electronic device, including:

[0068] A processor; and a memory arranged to store computer executable instructions, which when executed cause the processor to perform the method steps described in the first aspect.

[0069] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple applications, the electronic device executes the method steps described in the first aspect.

[0070] The beneficial effects of the present invention are:

[0071] The present invention is a method and system for generating a balance and coordination training strategy based on machine learning. Compared with the prior art, the present invention has the following technical effects:

[0072] The present invention can improve the accuracy of balance and coordination training strategy generation through preprocessing, formulating scientific balance and coordination plans, obtaining fixed balance and coordination values, screening balance and coordination actions, optimizing balance and coordination plans, model construction and optimizing model steps, thereby improving the accuracy of balance and coordination training strategy generation, and optimizing the balance and coordination training strategy generation, which can greatly save resources and improve work efficiency, and can realize the intelligent generation of balance and coordination training strategies, and perform action correction and personalized customization of the balance and coordination training strategy generation in real time, which is of great significance to the balance and coordination training strategy generation, can adapt to the balance and coordination training strategy generation of different standards and different balance and coordination training strategy generation needs, and has a certain universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1A flow chart of the steps of a method for generating a balance and coordination training strategy based on machine learning according to the present invention;

[0074] Figure 2 It is a schematic diagram of the structure of an electronic device in an embodiment of this specification. DETAILED DESCRIPTION

[0075] The present invention is further described below by means of specific embodiments. The illustrative embodiments and descriptions of the present invention are used to explain the present invention but are not intended to limit the present invention.

[0076] The present invention provides a method and system for generating a balance and coordination training strategy based on machine learning, comprising the following steps:

[0077] like Figure 1 As shown, in this embodiment, the following steps are included:

[0078] Collect the user's sports data and medical health data, pre-process the sports data, and formulate a scientific balance plan based on the medical health data; the sports data includes daily sports data and balance and coordination sports data; daily sports data includes exercise steps, walking distance, exercise time, exercise type, duration, and preferred exercise; balance and coordination sports data includes balance training, coordination training, balance effect, coordination effect, heart rate, and strength; medical health data includes gender, age, weight, height, BIM, chronic diseases, acute diseases, past medical history, orthopedic disease history, family medical history, blood pressure, blood sugar, blood lipids, liver function, living habits, and mental health;

[0079] In the actual evaluation, the exercise data are 8532 steps, 5.2 kilometers of walking distance, 45 minutes of exercise time, fast walking Tai Chi, balance training 10 minutes per day, coordination training three times a week 15 minutes each time, balance effect improved by 3%, coordination effect improved by 2.1%, heart rate 120, strength 1.3 kg, duration 20 minutes, preference for Tai Chi; medical health data are female, age 72 years old, weight 62 kg, height 159 cm, BIM24, chronic disease hypertension, acute disease no, past medical history no, orthopedic disease history osteoporosis knee rheumatoid, family history of heart disease, blood pressure 130 / 85 mmHg, fasting blood sugar 4.8 mmol / L, blood lipid total cholesterol 4.5 mmol / L, normal liver function, living habits no smoking and occasional drinking, good mental health;

[0080] The scientific balance program includes 5 minutes of warm-up, 20 seconds of Tai Chi standing and single-leg standing, 20 minutes of Sanda, 10 minutes of synchronous arm and leg exercise, 5 minutes of leg weight training with 1.5 kg, and 5 minutes of relaxation exercise.

[0081] Obtaining a fixed balance coordination value according to the daily exercise data, and filtering the balance coordination exercise data based on the balance coordination preference to obtain a balance coordination action;

[0082] In the actual evaluation, the fixed Hengxie value is 5.1%, and the Hengxie action is Tai Chi;

[0083] The scientific balance coordination scheme is adjusted by the balance coordination action to obtain an optimized balance coordination scheme, and a balance coordination training strategy generation model is constructed according to the fixed balance coordination value and the optimized balance coordination scheme;

[0084] In the actual evaluation, the optimized Hengxie program is 6 minutes of warm-up, 20 seconds of alternating Tai Chi standing with single-leg standing, 10 minutes of Tai Chi, 10 minutes of synchronous arm and leg exercise, 5 minutes of double-leg weight training with 1 kg, and 8 minutes of relaxation exercise;

[0085] The balance and coordination training strategy generation model is optimized, the data to be generated is input into the optimized balance and coordination training strategy generation model, and the balance and coordination training strategy is output.

[0086] In this embodiment, the method for obtaining a fixed balance value according to the daily exercise data includes:

[0087] Static balance test and muscle activity analysis were used to obtain the balance and coordination effects of daily exercise data. The balance and coordination effects were weighted and summed according to the importance of position balance to obtain a comprehensive fixed balance and coordination value.

[0088] In this embodiment, the method of obtaining the balancing and coordination action by screening the balancing and coordination motion data based on the balancing and coordination preference includes:

[0089] Deep convolutional neural network and long short-term memory network are used to extract motion features and facial emotion features of balance and coordination motion data;

[0090] Obtain facial emotion scores based on muscle activity analysis and calculate the eigenvalues ​​of motion features:

[0091]

[0092] The number of features is n, and the characteristic value of the w-th motion feature is The wth motion feature of the cth attribute is The standard value of the wth motion feature of the cth attribute is The average value of the motion feature is The number of attributes is p, the distance of attribute data is d, and the control factor of the wth motion feature of the cth attribute is α w,c ;

[0093] Calculate the contribution rate of Hengxie:

[0094]

[0095] The balance contribution rate of the wth motion feature is The balancing effect of the wth motion feature is h w , the coordination error value of the w-th motion feature is σ w ;

[0096] Calculate the balance preference of motion features:

[0097]

[0098] The trade-off preference of the w-th motion feature is The initial facial emotion score of the w-th motion feature and the ending facial emotion score of the w-th motion feature are C w,o , the natural constant is e, and the emotion score error of the w-th motion feature is τ w , the membership degree of the rth attribute of the wth motion feature is z w (r), the number of attributes is N, and the fuzzy element of the k-th membership degree is s κ , the fuzzy element of the ιth membership degree is s ι ;

[0099] The motion feature coordination preferences of the actions are summed to obtain the action coordination preference, and the actions with action coordination preference greater than 0.347 are output as coordinated actions.

[0100] In this embodiment, the method of using the balancing action to adjust the scientific balancing scheme to obtain an optimized balancing scheme includes:

[0101] Calculate the composite score for the alternative moves:

[0102]

[0103] The comprehensive score of the i-th alternative action is The contribution rate weight is ρ 1 , the sentiment weight is ρ 2 , the balance contribution rate of the ith alternative action is The initial facial emotion score C of the i-th alternative action i,o , the ending facial emotion score of the ith alternative action is C i,en ;

[0104] Sort the alternative actions in descending order according to the comprehensive scores, and use the first alternative action to replace the non-balancing action in the scientific balancing solution to obtain the optimized action;

[0105] According to the goal of the scientific balance plan, the training time of the optimized action is adjusted to obtain the optimized balance plan.

[0106] In this embodiment, the method for constructing a balance coordination training strategy generation model according to the fixed coordination value and the optimized coordination scheme includes:

[0107] The target coordination amount is obtained by subtracting the fixed coordination value from the target of the scientific coordination plan, and the training time of the optimized coordination plan is adjusted according to the target coordination amount to obtain the training plan;

[0108] Use scoring to obtain user satisfaction and calculate the training strategy score:

[0109]

[0110] The scientific coordination solution goal of the jth user is K j , the fixed trade-off value for the jth user is The first weight is The second weight is The error value of the jth user is The training strategy score of the jth user is

[0111] The objective function of the balanced coordination training strategy generation model is constructed according to the training strategy function, and the expression is:

[0112]

[0113] The objective function of the jth user is Training strategy score and predict the training strategy score The loss function is

[0114] The balance coordination training strategy generation model includes random forest algorithm, convolutional neural network algorithm, long short-term memory network algorithm, spectral clustering algorithm, and Q learning;

[0115] The random forest algorithm divides the input data into training and test sets in a ratio of 4:1;

[0116] The convolutional neural network algorithm automatically learns the motion feature representation of the training set to obtain the balance motion feature, and then fuses the balance motion features according to the action to obtain the balance action.

[0117] The long short-term memory network algorithm obtains the balance coordination effect value according to the time change of the balance coordination action on the user;

[0118] Spectral clustering algorithm, using the balance effect value to cluster the balance actions and obtain the action classification;

[0119] Q-learning generates a balanced coordination training policy by learning the best policy for the agent to take actions in a given state.

[0120] In this embodiment, the method for optimizing the balance and coordination training strategy generation model includes:

[0121] Randomly divide the particle population into two equal sub-populations and calculate the fitness of the particles:

[0122]

[0123] The fitness of the ath particle is The objective function value of the ath particle is The number of constraints is m, and the penalty factor of the u-th constraint is β u , the u-th constraint function of the a-th particle is k u (a), the predicted value of the objective function is The variable control constant is b;

[0124] Sort the particles of the sub-populations in descending order according to their fitness, take the particle with the best fitness in the first sub-population as the first individual, and take the particle with the best fitness in the second sub-population as the second individual;

[0125] Calculate the first and second parameters:

[0126]

[0127]

[0128] The maximum number of iterations is t max , the number of iterations is t, and the first parameter is ψ 1 , the second parameter is ψ 2 , the interference factor is μ, the number of particles in the subpopulation is n, and the influence constant is c;

[0129] When the first parameter is greater than or equal to 0.25 and the second parameter is greater than 0.6, update the particle position:

[0130]

[0131]

[0132] The position of the ath main particle in the t+1th iteration is S a,y (t+1), the position of the ath main particle in the tth iteration is S a,y (t), the first random number is v 1 , the second random number is v 2 , the optimal position of the main particle in the tth iteration is S y,bs (t), the worst position of the main particle at the tth iteration is S y,wt (t), the position of the ath particle in the t+1th iteration is S a,f(t+1), the position of the ath particle in the tth iteration is S a,f (t), the optimal position of the secondary particle is S f,bs , the worst position of the secondary particle is S f,wt ;

[0133] If the random number is greater than 0.6, update the particle position:

[0134]

[0135]

[0136] The position of the ath main particle in the t+1th iteration is The position of the ath particle in the t+1th iteration is

[0137] On the contrary, iterate continuously until the update conditions are met and output the optimal position.

[0138] In a second aspect, a system for generating a balance and coordination training strategy based on machine learning includes:

[0139] Data collection module: used to collect the user's sports data and medical health data, pre-process the sports data, and formulate a scientific coordination plan based on the medical health data;

[0140] Data processing module: used for obtaining fixed balance coordination values ​​according to the daily exercise data, and filtering the balance coordination exercise data based on balance coordination preferences to obtain balance coordination actions;

[0141] Model building module: used to use the balance coordination action to adjust the scientific balance coordination scheme to obtain an optimized balance coordination scheme, and to build a balance coordination training strategy generation model according to the fixed balance coordination value and the optimized balance coordination scheme;

[0142] Optimization output module: used to optimize the balance and coordination training strategy generation model, input the data to be generated into the optimized balance and coordination training strategy generation model, and output the balance and coordination training strategy.

[0143] Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 2 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0144] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0145] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0146] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a balance coordination training strategy generation device based on machine learning at the logical level. The processor executes the program stored in the memory and is specifically used to execute any of the above-mentioned balance coordination training strategy generation methods based on machine learning.

[0147] The above application Figure 1The embodiment shown discloses a method for generating a balance coordination training strategy based on machine learning, which can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0148] The electronic device may also perform Figure 1 A machine learning-based balance and coordination training strategy generation method is proposed and implemented Figure 1 The functions of the illustrated embodiment will not be described in detail in the embodiments of the present application.

[0149] An embodiment of the present application also proposes a computer-readable storage medium, which stores one or more programs, and the one or more programs include instructions. When the instructions are executed by an electronic device including multiple applications, any one of the aforementioned balance coordination training strategy generation methods based on machine learning is executed.

[0150] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0151] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0152] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0153] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0154] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0155] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0156] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0157] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0158] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0159] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for generating a balance and coordination training strategy based on machine learning, characterized in that: The following steps are involved: Collect the user's exercise data and medical health data, pre-process the exercise data, and formulate a scientific balance and coordination plan based on the medical health data; the exercise data includes daily exercise data and balance and coordination exercise data; Obtaining a fixed balance coordination value according to the daily exercise data, and filtering the balance coordination exercise data based on the balance coordination preference to obtain a balance coordination action; The scientific balance coordination scheme is adjusted by the balance coordination action to obtain an optimized balance coordination scheme, and a balance coordination training strategy generation model is constructed according to the fixed balance coordination value and the optimized balance coordination scheme; Optimizing the balance and coordination training strategy generation model, inputting the data to be generated into the optimized balance and coordination training strategy generation model, and outputting the balance and coordination training strategy; The method for obtaining a fixed balance value according to the daily exercise data includes: The balance and coordination effects of daily exercise data were obtained by using static balance test and muscle activity analysis. The balance and coordination effects were weighted and summed according to the importance of positional balance to obtain a comprehensive fixed balance and coordination value. The method of screening the balance coordination motion data based on the balance coordination preference to obtain the balance coordination action includes: Deep convolutional neural network and long short-term memory network are used to extract motion features and facial emotion features of balanced coordination motion data; Obtain facial emotion scores based on muscle activity analysis and calculate the eigenvalues ​​of motion features: The number of features is The characteristic value of the w-th motion feature is The wth motion feature of the cth attribute is The standard value of the wth motion feature of the cth attribute is The average value of the motion feature is The number of attributes is p, and the distance between attribute data is The control factor of the cth attribute wth motion feature is α w,c ; Calculate the contribution rate of Hengxie: The balance contribution rate of the wth motion feature is The balancing effect of the wth motion feature is h w , the coordination error value of the w-th motion feature is σ w ; Calculate the balance preference of motion features: The trade-off preference of the w-th motion feature is θ w , the initial facial emotion score of the w-th motion feature, and the ending facial emotion score of the w-th motion feature is C w,o , the natural constant is e, and the emotion score error of the w-th motion feature is τ w , the membership degree of the rth attribute of the wth motion feature is The number of attributes is N, and the fuzzy element of the k-th membership degree is s κ , the fuzzy element of the ιth membership degree is s ι ; The motion feature coordination preference of the action is summed up to obtain the action coordination preference, and the action with the action coordination preference greater than 0.347 is output as the coordinated action; The method of using the balancing action to adjust the scientific balancing scheme to obtain an optimized balancing scheme includes: Calculate the composite score for the alternative moves: The comprehensive score of the i-th alternative action is The contribution rate weight is ρ1, the emotion weight is ρ2, and the balance contribution rate of the ith alternative action is The initial facial emotion score C of the i-th alternative action i,o , the ending facial emotion score of the ith alternative action is C i,en ; Sort the alternative actions in descending order according to the comprehensive scores, and use the first alternative action to replace the non-balancing action in the scientific balancing solution to obtain the optimized action; According to the goal of the scientific balance program, the training time of the optimized action is adjusted to obtain the optimized balance program; The method for constructing a balance coordination training strategy generation model according to the fixed balance coordination value and the optimized balance coordination scheme includes: The target coordination amount is obtained by subtracting the fixed coordination value from the target of the scientific coordination plan, and the training time of the optimized coordination plan is adjusted according to the target coordination amount to obtain the training plan; Use scoring to obtain user satisfaction and calculate the training strategy score: The scientific coordination solution goal of the jth user is K j , the fixed trade-off value for the jth user is The first weight is The second weight is The error value of the jth user is The training strategy score of the jth user is The objective function of the balanced coordination training strategy generation model is constructed according to the training strategy function, and the expression is: The objective function of the jth user is Training strategy score and predict the training strategy score The loss function is The balance coordination training strategy generation model includes random forest algorithm, convolutional neural network algorithm, long short-term memory network algorithm, spectral clustering algorithm, and Q learning; The random forest algorithm divides the input data into training and test sets in a ratio of 4:1; The convolutional neural network algorithm automatically learns the motion feature representation of the training set to obtain the balance motion feature, and then fuses the balance motion features according to the action to obtain the balance action. The long short-term memory network algorithm obtains the balance coordination effect value according to the time change of the balance coordination action on the user; Spectral clustering algorithm, using the balance effect value to cluster the balance actions and obtain the action classification; Q-learning generates a balanced coordination training policy by learning the best strategy for the agent to take actions in a given state; The method for optimizing the balance coordination training strategy generation model comprises: Randomly divide the particle population into two equal sub-populations and calculate the fitness of the particles: The fitness of the ath particle is The objective function value of the ath particle is The number of constraints is The penalty factor of the u-th constraint is β u , the u-th constraint function of the a-th particle is k u (a), the predicted value of the objective function is The variable control constant is b; Sort the particles of the sub-populations in descending order according to their fitness, take the particle with the best fitness in the first sub-population as the first individual, and take the particle with the best fitness in the second sub-population as the second individual; Calculate the first and second parameters: The maximum number of iterations is t max , the number of iterations is t, the first parameter is ψ1, the second parameter is ψ2, the interference factor is μ, the number of particles in the sub-population is n, and the influence constant is When the first parameter is greater than or equal to 0.25 and the second parameter is greater than 0.6, update the particle position: The position of the ath main particle in the t+1th iteration is S a,y (t+1), the position of the ath main particle in the tth iteration is S a,y (t), the first random number is v1, the second random number is v2, and the optimal position of the main particle in the tth iteration is S y,bs (t), the worst position of the main particle at the tth iteration is S y,wt (t), the position of the ath particle in the t+1th iteration is S a,f (t+1), the position of the ath particle in the tth iteration is S a,f (t), the optimal position of the secondary particle is S f,bs , the worst position of the secondary particle is S f,wt ; If the random number is greater than 0.6, update the particle position: The position of the ath main particle in the t+1th iteration is The position of the ath particle in the t+1th iteration is On the contrary, iterate continuously until the update conditions are met and output the optimal position.

2. A balance and coordination training strategy generation system based on machine learning, used to execute the balance and coordination training strategy generation method based on machine learning according to claim 1, characterized in that: include: Data collection module: used to collect the user's sports data and medical health data, pre-process the sports data, and formulate a scientific coordination plan based on the medical health data; Data processing module: used for obtaining fixed balance coordination values ​​according to the daily exercise data, and filtering the balance coordination exercise data based on balance coordination preferences to obtain balance coordination actions; Model building module: used to use the balance coordination action to adjust the scientific balance coordination scheme to obtain an optimized balance coordination scheme, and to build a balance coordination training strategy generation model according to the fixed balance coordination value and the optimized balance coordination scheme; Optimization output module: used to optimize the balance and coordination training strategy generation model, input the data to be generated into the optimized balance and coordination training strategy generation model, and output the balance and coordination training strategy.

3. An electronic device, comprising: processor; as well as A memory arranged to store computer executable instructions, which when executed cause the processor to perform the method of claim 1.

4. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, causes the electronic device to execute the method of claim 1.

Citation Information

Cited By

  • Deep learning-based balance coordination training strategy generation method and system

    CN121503733A