Exercise safety training scheme recommendation method and system

Abstract

According to the exercise safety training scheme recommendation method and system provided by the invention, the target user training state information of the to-be-analyzed target and the important knowledge vector information of the to-be-analyzed target are fused, so that the secondary knowledge vector information of the to-be-analyzed target and the important knowledge vector information of the to-be-analyzed target are integrated; more knowledge vector information of the to-be-analyzed target is extracted, and the stability of the rock can be accurately determined, so that the accuracy and reliability of rock training effect recommendation can be improved; besides, through the target user training state information of the to-be-analyzed target, the to-be-analyzed target is recommended to obtain an exercise safety training scheme recommendation result, so that the recommendation accuracy is ensured, the user can be ensured to obtain personalized training, and the optimal training effect is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of training effect recommendation, and in particular to a method and system for recommending sports safety training programs. Background Art

[0002] With the continuous improvement of the quality of life, users are more and more concerned about their physical condition. Users go to gyms, yoga studios, etc. for sports training to ensure that their physical condition is at its best. However, each user's physical condition is different. Targeted training plans are needed for different physical conditions. However, how to recommend personalized training plans for different users is a technical problem that urgently needs to be solved. Summary of the Invention

[0003] In order to improve the technical problems existing in the related technologies, the present application provides a method and system for recommending sports safety training programs.

[0004] In a first aspect, a method for recommending a sports safety training program is provided, comprising: extracting knowledge vectors of the training injury risk level of the personalized user sports data to be performed through a hazard level prediction network for the training injury risk level, and obtaining knowledge vector information of the training injury risk level corresponding to the target to be analyzed, as secondary knowledge vector information of the target to be analyzed; performing prediction processing in combination with the secondary knowledge vector information of the target to be analyzed, and obtaining target user training status information of the target to be analyzed; extracting important knowledge vectors of the personalized user sports data to be performed, and obtaining important knowledge vector information of the target to be analyzed; fusing the target user training status information of the target to be analyzed in the personalized user sports data to be performed with the important knowledge vector information of the target to be analyzed, and obtaining a fusion result of the target to be analyzed; and performing recommendation processing in combination with the fusion result of the target to be analyzed, and obtaining a sports safety training program recommendation result.

[0005] In the present application, the prediction network for predicting personalized user motion data includes multiple danger level prediction networks, and the multiple danger level prediction networks correspond to different training injury risk levels respectively; the prediction processing is performed in combination with the secondary knowledge vector information of the target to be analyzed to obtain the target user training status information of the target to be analyzed, including: through the danger level prediction network of the training injury risk level, and in combination with the knowledge vector information of the training injury risk level corresponding to the training injury risk level of the target to be analyzed, a secondary prediction processing is performed to obtain the target user training status information of the target to be analyzed corresponding to the training injury risk level; the target user training status information of the target to be analyzed corresponding to multiple training injury risk levels is fused to obtain the target user training status information of the target to be analyzed.

[0006] In the present application, the secondary prediction processing is performed on the knowledge vector information of the training injury risk level corresponding to the target to be analyzed to obtain the target user training status information of the target to be analyzed corresponding to the training injury risk level, including: performing knowledge vector elimination processing on the knowledge vector information of the training injury risk level corresponding to the target to be analyzed to obtain the eliminated knowledge vector information; performing weighted processing on the eliminated knowledge vector information to obtain the target user training status information of the training injury risk level corresponding to the target to be analyzed.

[0007] In the present application, the target user training status information corresponding to the multiple training injury risk levels of the target to be analyzed is fused to obtain the target user training status information of the target to be analyzed, including: when there is target user training status information corresponding to two different training injury risk levels, the distinction between the target user training status information of the two different training injury risk levels is used as the target user training status information of the target to be analyzed; when there is target user training status information corresponding to at least three different training injury risk levels, the target user training status information of the at least three different training injury risk levels is fused to obtain the target user training status information of the target to be analyzed.

[0008] In the present application, each of the danger level prediction networks includes a plurality of staged prediction networks, and the plurality of staged prediction networks correspond to different attributes of the target to be analyzed; the danger level prediction network of the training injury risk level performs a knowledge vector extraction process on the personalized user motion data to be performed, and obtains the knowledge vector information of the training injury risk level corresponding to the target to be analyzed, including: for each attribute of the target to be analyzed, performing the following processing: performing a knowledge vector extraction process on the personalized user motion data to be performed through the staged prediction network corresponding to the attribute, and obtaining the knowledge vector information of the target to be analyzed corresponding to the attribute; the training injury risk level prediction network performs a knowledge vector extraction process on the personalized user motion data to be performed, and obtains the knowledge vector information of the target to be analyzed corresponding to the attribute; the training injury risk level prediction network performs a knowledge vector extraction process on the personalized user motion data to be performed, and obtains the knowledge vector information of the target to be analyzed corresponding to the attribute. A hazard level prediction network for training injury risk level is used, and secondary prediction processing is performed in combination with the knowledge vector information of the target to be analyzed corresponding to the training injury risk level to obtain the target user training status information of the target to be analyzed at the training injury risk level, including: for each attribute of the target to be analyzed, the following processing is performed: through a staged prediction network corresponding to the attribute, and secondary prediction processing is performed in combination with the knowledge vector information of the target to be analyzed corresponding to the attribute to obtain the target user training status information of the target to be analyzed corresponding to the attribute; the target user training status information of multiple attributes corresponding to the target to be analyzed is fused to obtain the target user training status information of the target to be analyzed corresponding to the training injury risk level.

[0009] In the present application, the secondary prediction processing is performed on the knowledge vector information of the target to be analyzed corresponding to the attribute to obtain the target user training status information of the target to be analyzed corresponding to the attribute, including: performing knowledge vector elimination processing on the knowledge vector information of the target to be analyzed corresponding to the attribute to obtain the eliminated knowledge vector information; performing weighted processing on the eliminated knowledge vector information to obtain the target user training status information of the target to be analyzed corresponding to the attribute.

[0010] In the present application, the weighted processing of the knowledge vector information after elimination is performed to obtain the target user training status information corresponding to the attribute of the target to be analyzed, including: mapping the knowledge vector information after elimination to the target user training status information to obtain a training effect coefficient arrangement; and using the target user training status information corresponding to the maximum coefficient in the training effect coefficient arrangement as the target user training status information corresponding to the attribute of the target to be analyzed.

[0011] In the present application, before the secondary knowledge vector extraction processing is performed on the personalized user motion data to be performed, it also includes: performing differentiation processing on the personalized user motion data to be performed to obtain multiple training effect factor range data of the personalized user motion data to be performed; performing secondary knowledge vector extraction processing on the personalized user motion data to be performed to obtain secondary knowledge vector information of the target to be analyzed in the personalized user motion data to be performed, including: performing secondary knowledge vector extraction processing on each of the training effect factor range data to obtain the secondary knowledge vector information of the target to be analyzed in the training effect factor range data; performing prediction processing in combination with the secondary knowledge vector information of the target to be analyzed to obtain the target user training status information of the target to be analyzed, including: performing prediction processing through a prediction network and combining the secondary knowledge vector information of the target to be analyzed in each of the training effect factor range data to obtain the target user training status information of the target to be analyzed corresponding to each of the training effect factor range data; combining the target user training status information of the target to be analyzed corresponding to multiple ranges to obtain the target user training status information of the target to be analyzed.

[0012] In the present application, the target user training status information of the target to be analyzed corresponding to multiple ranges is combined and processed to obtain the target user training status information of the target to be analyzed, including: fusing the target user training status information of the target to be analyzed corresponding to multiple ranges to obtain fused target user training status information; and performing COUNTIF function processing on the fused target user training status information to obtain the target user training status information of the target to be analyzed.

[0013] In the present application, the target user training status information corresponding to the target to be analyzed in multiple ranges is combined and processed to obtain the target user training status information of the target to be analyzed, including: based on the user physical condition of the target to be analyzed represented by the target user training status information, determining the target user training status information with the largest user physical condition from the target user training status information corresponding to multiple training effect factor range data of the target to be analyzed, and using the target user training status information representing the largest user physical condition as the target user training status information of the target to be analyzed.

[0014] In the present application, the recommendation processing is performed in combination with the fusion result of the target to be analyzed to obtain the sports safety training program recommendation result, including: mapping the fusion result of the target to be analyzed to the training effect coefficient arrangement of the sports safety training program recommendation result; and taking the training injury risk level corresponding to the maximum coefficient in the training effect coefficient arrangement as the sports safety training program recommendation result.

[0015] In the present application, before the secondary knowledge vector extraction processing is performed on the personalized user motion data that needs to be performed, it also includes: obtaining original personalized user motion data in response to a recommendation instruction for the personalized user motion data that needs to be performed; performing at least one of the following processing on the original personalized user motion data: simplifying each data cluster in the original personalized user motion data, and using the simplified original personalized user motion data as the personalized user motion data that needs to be performed; eliminating the influencing information in the original personalized user motion data, and using the eliminated original personalized user motion data as the personalized user motion data that needs to be performed; analyzing the target to be analyzed in the original personalized user motion data, and using the analyzed original personalized user motion data as the personalized user motion data that needs to be performed.

[0016] In a second aspect, a sports safety training program recommendation system is provided, comprising a processor and a memory communicating with each other, wherein the processor is configured to read a computer program from the memory and execute the program to implement the above method.

[0017] The embodiment of the present application provides a method and system for recommending a sports safety training program. By fusing the target user training status information of the target to be analyzed with the important knowledge vector information of the target to be analyzed, the secondary knowledge vector information of the target to be analyzed and the important knowledge vector information of the target to be analyzed are integrated to extract more knowledge vector information of the target to be analyzed, and the stability of the rock can be accurately determined, thereby improving the accuracy and reliability of the rock training effect recommendation. In addition, the target to be analyzed is recommended based on the target user training status information, and the sports safety training program recommendation result is obtained to ensure the accuracy of the recommendation, thereby ensuring that the user receives personalized training and achieves the best training effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 This is a flowchart of a method for recommending a sports safety training program provided in an embodiment of the present application. DETAILED DESCRIPTION

[0020] In order to better understand the above technical solution, the technical solution of this application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of this application and the specific knowledge vectors in the embodiments are detailed descriptions of the technical solution of this application, rather than limitations on the technical solution of this application. In the absence of conflict, the embodiments of this application and the technical knowledge vectors in the embodiments can be combined with each other.

[0021] See also Figure 1 , shows a method for recommending a sports safety training program, which may include the technical solutions described in the following steps 101-105.

[0022] In step 101, a risk level prediction network for training injury risk levels is used to extract the knowledge vectors of the training injury risk levels of the personalized user motion data to be analyzed, and the knowledge vector information of the training injury risk levels corresponding to the target to be analyzed is obtained as the secondary knowledge vector information of the target to be analyzed.

[0023] The training injury risk level can be divided into levels 1 to 10, with the larger the value, the higher the risk level. The knowledge vector can be understood as a feature.

[0024] For example, the hazard level prediction network is a type of artificial intelligence network that can analyze the coefficient of rock training effect; among them, artificial intelligence (AI) is an interdisciplinary and emerging discipline based on computer science, which is an interdisciplinary integration of multiple disciplines such as computer science, psychology, and philosophy. It studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. It attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a way similar to human intelligence. Research in this field includes robots, speech recognition, image recognition, natural language processing and expert systems.

[0025] As an example of obtaining the required personalized user motion data, the user's body state information can be obtained by analyzing the user's body through a user state detection device, and the collected personalized user motion data that needs to be performed is sent to the terminal. The terminal forwards the required personalized user motion data to the server, so that the server extracts the secondary knowledge vector of the required personalized user motion data, and obtains the secondary knowledge vector information of the target to be analyzed in the required personalized user motion data. Subsequent analysis and processing can be performed based on the secondary knowledge vector information of the target to be analyzed.

[0026] In some possible implementations, before performing secondary knowledge vector extraction processing on the personalized user motion data to be performed, the method further includes: obtaining original personalized user motion data in response to a recommendation instruction for the personalized user motion data to be performed; performing at least one of the following processing on the original personalized user motion data: simplifying each data cluster in the original personalized user motion data and using the simplified original personalized user motion data as the personalized user motion data to be performed; removing influencing information from the original personalized user motion data and using the removed original personalized user motion data as the personalized user motion data to be performed; analyzing the target to be analyzed in the original personalized user motion data and using the analyzed original personalized user motion data as the personalized user motion data to be performed. The simplification processing can be understood as nondimensionalization (or dimensionless), which refers to removing some or all units of an equation involving physical quantities by replacing them with an appropriate variable in order to simplify experiments or calculations.

[0027] For example, raw personalized user motion data is collected, and the terminal forwards the raw personalized user motion data to the server. After receiving the raw personalized user motion data, the server preprocesses the raw personalized user motion data. The preprocessed personalized user motion data is more suitable for subsequent personalized user motion data analysis. Preprocessing can be understood as debugging the raw personalized user motion data through secondary knowledge vectors. For example, each data cluster in the raw personalized user motion data is simplified, and the simplified raw personalized user motion data is used as the personalized user motion data to be analyzed. The simplification ensures that certain knowledge vectors of the personalized user motion data have invariant properties under a given transformation; the influencing information in the raw personalized user motion data is eliminated, and the eliminated raw personalized user motion data is used as the personalized user motion data to be analyzed. The elimination can remove random influencing information in the personalized user motion data; the target to be analyzed in the raw personalized user motion data is analyzed, and the analyzed raw personalized user motion data is used as the personalized user motion data to be analyzed. The analysis can conduct targeted analysis of the information in the personalized user motion data to improve the inaccuracy of the personalized user motion data.

[0028] In step 102, prediction processing is performed in combination with the secondary knowledge vector information of the target to be analyzed to obtain the target user training state information of the target to be analyzed.

[0029] For example, after the server obtains the secondary knowledge vector information of the target to be analyzed, it can perform prediction processing based on the secondary knowledge vector information of the target to be analyzed to obtain the target user training status information of the target to be analyzed, so as to subsequently perform subsequent personalized user motion data recommendation operations based on the target user training status information of the target to be analyzed.

[0030] Among them, the secondary knowledge vector is understood as feature information that has no direct impact on the user training plan.

[0031] An embodiment of the present invention provides a method for recommending a sports safety training program, wherein step 101 includes step 1011A: In step 1011A, the following processing is performed for any one of a plurality of training injury risk levels: a knowledge vector extraction process of the training injury risk level is performed on the personalized user sports data to be analyzed using a risk level prediction network for the training injury risk level, and knowledge vector information of the training injury risk level corresponding to the target to be analyzed is obtained as secondary knowledge vector information of the target to be analyzed;

[0032] Furthermore, step 102 includes steps 1021A to 1022A: in step 1021A, a hazard level prediction network of the trained injury risk level is used, and a secondary prediction process is performed in combination with the knowledge vector information of the trained injury risk level of the target to be analyzed to obtain the target user training status information of the trained injury risk level of the target to be analyzed; in step 1022A, the target user training status information of multiple training injury risk levels of the target to be analyzed is fused to obtain the target user training status information of the target to be analyzed.

[0033] It should be understood that the prediction network used for personalized user motion data prediction includes multiple danger level prediction networks, and the multiple danger level prediction networks correspond to different training injury risk levels; when the prediction processing is performed in combination with the secondary knowledge vector information of the target to be analyzed, the problem of inaccurate prediction is improved, thereby ensuring the accuracy of the target user training status information of the target to be analyzed.

[0034] Among them, the prediction network for personalized user motion data prediction includes multiple danger level prediction networks, and the multiple danger level prediction networks correspond to different training injury risk levels. For any training injury risk level among the multiple training injury risk levels, the knowledge vector information for the training injury risk level in the personalized user motion data to be analyzed is extracted through the decision layer in the danger level prediction network corresponding to the training injury risk level, and combined with the knowledge vector information of the training injury risk level corresponding to the target to be analyzed, the target to be analyzed is subjected to a secondary prediction through the danger level prediction network corresponding to the training injury risk level to obtain the target user training status information corresponding to the training injury risk level of the target to be analyzed, that is, to obtain targeted target user training status information, and the targeted target user training status information is fused to obtain target user training status information for all training injury risk levels, so that accurate target recommendations can be made based on the target user training status information for all training injury risk levels.

[0035] In some possible implementations, secondary prediction processing is performed on the knowledge vector information of the training injury risk level of the target to be analyzed to obtain target user training status information corresponding to the training injury risk level of the target to be analyzed, including: performing knowledge vector elimination processing on the knowledge vector information of the training injury risk level of the target to be analyzed to obtain eliminated knowledge vector information; and performing weighted processing on the eliminated knowledge vector information to obtain target user training status information corresponding to the training injury risk level of the target to be analyzed. The weighted processing may include a level calculation method.

[0036] It should be understood that when secondary prediction processing is performed in combination with the knowledge vector information of the training injury risk level corresponding to the target to be analyzed, the problem of influence is improved, so that the target user training status information of the training injury risk level corresponding to the target to be analyzed can be accurately obtained.

[0037] Furthermore, through the pooling layer in the danger level prediction network corresponding to the training injury risk level, the knowledge vector information of the training injury risk level corresponding to the target to be analyzed is subjected to knowledge vector elimination to obtain the eliminated knowledge vector information so as to delete unimportant knowledge vector information, and through the prediction layer in the danger level prediction network corresponding to the training injury risk level, the eliminated knowledge vector information is weightedly processed to obtain the target user training status information of the training injury risk level corresponding to the target to be analyzed.

[0038] In terms of some possible implementation embodiments, the target user training status information corresponding to multiple training injury risk levels of the target to be analyzed is fused to obtain the target user training status information of the target to be analyzed, including: when there is target user training status information corresponding to two different training injury risk levels, the distinction between the target user training status information of the two different training injury risk levels is used as the target user training status information of the target to be analyzed; when there is target user training status information corresponding to at least three different training injury risk levels, the target user training status information of at least three different training injury risk levels is fused to obtain the target user training status information of the target to be analyzed.

[0039] It should be understood that when the target user training status information corresponding to multiple training injury risk levels of the target to be analyzed is fused, the problem of fusion error is improved, so that the target user training status information of the target to be analyzed can be accurately obtained.

[0040] Furthermore, when there are two different training injury risk levels, target user training status information corresponding to the two different training injury risk levels will be generated, and the distinction between the target user training status information of the two different training injury risk levels will be used as the target user training status information of the target to be analyzed, so as to perform knowledge vector extraction and learning through comparative learning, and better extract the commonalities within the domain and the distinctions between domains; when there are at least three different training injury risk levels, target user training status information corresponding to at least three different training injury risk levels will be generated, and the target user training status information of the at least three different training injury risk levels will be fused, and the fusion result will be used as the target user training status information of the target to be analyzed, so that the target user training status information of the target to be analyzed includes the target user training status information for all training injury risk levels, so that accurate personalized user motion data recommendations can be made based on the target user training status information of the target to be analyzed.

[0041] In some possible implementations, the training injury risk level is further subdivided, and a knowledge vector of the training injury risk level is extracted from the personalized user motion data to be analyzed using a danger level prediction network for the training injury risk level to obtain knowledge vector information of the training injury risk level corresponding to the target to be analyzed. The method includes: performing the following processing on each attribute of the target to be analyzed: extracting the knowledge vector of the personalized user motion data to be analyzed using a staged prediction network for the corresponding attribute to obtain knowledge vector information of the attribute corresponding to the target to be analyzed;

[0042] Furthermore, by training a hazard level prediction network for injury risk levels and performing secondary prediction processing in combination with the knowledge vector information of the training injury risk level corresponding to the target to be analyzed, the target user training status information of the training injury risk level of the target to be analyzed is obtained, including: for each attribute of the target to be analyzed, performing the following processing: through a staged prediction network for the corresponding attribute and performing secondary prediction processing in combination with the knowledge vector information of the corresponding attribute of the target to be analyzed, the target user training status information of the corresponding attribute of the target to be analyzed is obtained; the target user training status information of multiple attributes corresponding to the target to be analyzed is fused to obtain the target user training status information of the training injury risk level of the target to be analyzed.

[0043] For example, the attributes can be understood as improving lung function, increasing muscle strength and endurance, improving cardiopulmonary health, promoting metabolism, strengthening physical fitness, improving sleep, improving mental health, etc.

[0044] It should be understood that each danger level prediction network includes multiple staged prediction networks, and the multiple staged prediction networks correspond to different attributes of the target to be analyzed; when the knowledge vector of the training injury risk level is extracted and processed for the personalized user motion data required, the problem of inaccurate extraction is improved, so that the knowledge vector information of the training injury risk level corresponding to the target to be analyzed can be accurately obtained.

[0045] Furthermore, each danger level prediction network includes multiple staged prediction networks, and the multiple staged prediction networks correspond to different attributes of the target to be analyzed; for any attribute of the multiple attributes under a certain training injury risk level, the knowledge vector information for the attribute in the personalized user motion data to be performed is extracted through the decision layer in the staged prediction network of the corresponding attribute, and combined with the knowledge vector information of the attribute corresponding to the target to be analyzed, a secondary prediction is made on the target to be analyzed through the danger level prediction network corresponding to the attribute to obtain the target user training status information of the attribute corresponding to the target to be analyzed, that is, to obtain targeted target user training status information, and the targeted target user training status information is fused to obtain target user training status information for all attributes, and the target user training status information corresponding to multiple attributes of the target to be analyzed is fused to obtain the target user training status information corresponding to the training injury risk level of the target to be analyzed, so that accurate target recommendation can be made based on the target user training status information for all attributes in the future.

[0046] In terms of some possible implementation embodiments, secondary prediction processing is performed on the knowledge vector information of the corresponding attributes of the target to be analyzed to obtain the target user training status information of the corresponding attributes of the target to be analyzed, including: performing knowledge vector elimination processing on the knowledge vector information of the corresponding attributes of the target to be analyzed to obtain the eliminated knowledge vector information; performing weighted processing on the eliminated knowledge vector information to obtain the target user training status information of the corresponding attributes of the target to be analyzed.

[0047] It should be understood that when the secondary prediction processing is performed in combination with the knowledge vector information of the corresponding attribute of the target to be analyzed, the problem of the influence of the information is improved, so that the target user training status information of the corresponding attribute of the target to be analyzed can be accurately obtained.

[0048] Furthermore, the knowledge vector information of the attribute corresponding to the target to be analyzed is subjected to knowledge vector elimination through the pooling layer in the staged prediction network corresponding to the attribute, and the eliminated knowledge vector information is obtained to delete unimportant knowledge vector information, and the eliminated knowledge vector information is weighted through the prediction layer in the staged prediction network corresponding to the attribute, thereby obtaining the target user training status information of the attribute corresponding to the target to be analyzed.

[0049] In terms of some possible implementation embodiments, the knowledge vector information after elimination is weighted to obtain the target user training status information of the corresponding attribute of the target to be analyzed, including: mapping the knowledge vector information after elimination to the target user training status information to obtain the training effect coefficient arrangement; and using the target user training status information corresponding to the maximum coefficient in the training effect coefficient arrangement as the target user training status information of the corresponding attribute of the target to be analyzed.

[0050] It should be understood that when weighted processing is performed on the knowledge vector information after elimination, the problem of inaccurate arrangement of the training effect coefficients is improved, so that the target user training status information corresponding to the attribute of the target to be analyzed can be accurately obtained.

[0051] Furthermore, the eliminated knowledge vector information is mapped to the target user training state information to obtain a training effect coefficient arrangement, in which there are multiple different coefficients of the target user training state information corresponding to the attribute.

[0052] A method for recommending a sports safety training program provided by an embodiment of the present invention, in order to accurately obtain the target user training status information of the target to be analyzed, also includes step 106: in step 106, the personalized user sports data to be performed is differentiated and processed to obtain multiple training effect factor range data of the personalized user sports data to be performed; then step 101 includes step 1011B: in step 1011B, secondary knowledge vector extraction processing is performed on each training effect factor range data to obtain secondary knowledge vector information of the target to be analyzed in the training effect factor range data; wherein, the secondary knowledge vector extraction processing can be understood as key content extraction or identification.

[0053] It should be understood that when the target user training status information corresponding to the target to be analyzed in a plurality of said ranges is combined and processed, the duplication problem is improved, so that the target user training status information of the target to be analyzed can be accurately obtained.

[0054] Furthermore, step 102 includes steps 1021B to 1022B: In step 1021B, a prediction process is performed using the prediction network and combined with the secondary knowledge vector information of the target to be analyzed in each training effect factor range data to obtain target user training state information corresponding to each training effect factor range data of the target to be analyzed; in step 1022B, the target user training state information corresponding to multiple ranges of the target to be analyzed is combined to obtain target user training state information of the target to be analyzed. Here, prediction can be understood as prediction.

[0055] It should be understood that when the prediction process is performed in combination with the secondary knowledge vector information of the target to be analyzed, the problem of inaccurate prediction process is improved, so that the target user training state information of the target to be analyzed can be accurately obtained.

[0056] For example, the personalized user motion data to be analyzed is first differentiated to obtain multiple training effect factor range data for the personalized user motion data to be analyzed. Then, through the prediction network used for personalized user motion data prediction, the secondary knowledge vector information of the target to be analyzed in each training effect factor range data is combined to perform predictions to obtain the target user training state information corresponding to each training effect factor range data for the target to be analyzed. Finally, the target user training state information corresponding to multiple ranges of the target to be analyzed is integrated to obtain the target user training state information for the target to be analyzed. By obtaining the staged target user training state information corresponding to each training effect factor range data, the target user training state information of all the targets to be analyzed in the personalized user motion data to be analyzed can be accurately obtained, avoiding the omission of the staged target user training state information in the personalized user motion data to be analyzed.

[0057] In terms of some possible implementation embodiments, target user training status information corresponding to multiple ranges of the target to be analyzed is combined and processed to obtain target user training status information of the target to be analyzed, including: fusing target user training status information corresponding to multiple ranges of the target to be analyzed to obtain fused target user training status information; and performing COUNTIF function processing on the fused target user training status information to obtain target user training status information of the target to be analyzed.

[0058] Furthermore, after the server obtains the target user training status information corresponding to each range of the target to be analyzed, it can first fuse the target user training status information corresponding to multiple ranges of the target to be analyzed (the multiple ranges can be part of all ranges or all ranges). Since the target user training status information of multiple ranges may be repeated, the fused target user training status information is deduplicated to remove the repeated target user training status information and obtain the target user training status information of the target to be analyzed without duplication.

[0059] In terms of some possible implementation embodiments, target user training status information corresponding to multiple ranges of the target to be analyzed is combined and processed to obtain the target user training status information of the target to be analyzed, including: according to the user physical condition of the item to be analyzed represented by the target user training status information, determining the target user training status information with the largest user physical condition from the target user training status information corresponding to multiple training effect factor range data of the target to be analyzed, and using the target user training status information representing the largest user physical condition as the target user training status information of the target to be analyzed.

[0060] It should be understood that when the target user training status information corresponding to multiple ranges of the target to be analyzed is combined and processed, the problem of unpredictable user physical condition in the user training status information is improved, so that the target user training status information of the target to be analyzed can be accurately obtained.

[0061] In step 103, important knowledge vector extraction processing is performed on the personalized user movement data to be analyzed to obtain important knowledge vector information of the target to be analyzed.

[0062] For example, important knowledge vectors are understood as feature information that directly affects the user's training plan.

[0063] Among them, important knowledge vectors are extracted from the personalized user movement data to obtain comprehensive knowledge vector information of the target to be analyzed.

[0064] In step 104, the target user training state information of the target to be analyzed in the personalized user motion data to be performed is fused with the important knowledge vector information of the target to be analyzed to obtain a fusion result of the target to be analyzed.

[0065] For example, after the server obtains the target user training status information of the target to be analyzed and the important knowledge vector information of the target to be analyzed, it fuses the target user training status information of the target to be analyzed and the important knowledge vector information of the target to be analyzed to obtain a fusion result of the target to be analyzed, so as to subsequently perform personalized user motion data recommendation processing based on the fusion result of the target to be analyzed.

[0066] In step 105, recommendation processing is performed in combination with the fusion result of the target to be analyzed to obtain a sports safety training program recommendation result.

[0067] Examples of sports safety training programs include:

[0068] 1: Primary exercise plan; 2: Intermediate exercise plan; 3: Advanced exercise plan; 4: Interval training and other plans.

[0069] For example, after the server obtains the fusion result of the target to be analyzed, it recommends the target to be analyzed based on the fusion result of the target to be analyzed to obtain a sports safety training program recommendation result.

[0070] In terms of some possible implementation embodiments, recommendation processing is performed based on the fusion results of the target to be analyzed to obtain a sports safety training program recommendation result, including: mapping the fusion results of the target to be analyzed to the training effect coefficient arrangement of the sports safety training program recommendation result; and taking the training injury risk level corresponding to the maximum coefficient in the training effect coefficient arrangement as the sports safety training program recommendation result.

[0071] It should be understood that when the recommendation process is performed in combination with the fusion results of the target to be analyzed, the problem of unreliable training effect coefficient is improved, so that the recommendation results of the sports safety training program can be accurately obtained.

[0072] In terms of some possible implementation embodiments, in order to analyze the training injury risk level of a target through a personalized user motion data processing network, the personalized user motion data processing network needs to be configured, and the configuration process includes: predicting and processing the personalized user motion data examples through the personalized user motion data processing network to obtain the target user training status information of the target to be analyzed; fusing the target user training status information of the target to be analyzed in the personalized user motion data examples with the important knowledge vector information of the target to be analyzed to obtain the fusion result of the target to be analyzed; performing recommendation processing based on the fusion result of the target to be analyzed to obtain a sports safety training plan recommendation result; constructing an evaluation index algorithm for the personalized user motion data processing network based on the sports safety training plan recommendation result and the training injury risk level identifier; updating the coefficients of the personalized user motion data processing network until the evaluation index algorithm converges, and using the updated coefficients of the personalized user motion data processing network when the evaluation index algorithm converges as the coefficients of the configured personalized user motion data processing network.

[0073] For example, a personalized user motion data processing network is used to extract secondary knowledge vectors from personalized user motion data examples to obtain secondary knowledge vector information of the target to be analyzed in the personalized user motion data examples. Prediction processing is performed based on the secondary knowledge vector information of the target to be analyzed to obtain target user training status information of the target to be analyzed. Important knowledge vector extraction processing is performed on personalized user motion data examples to obtain important knowledge vector information of the target to be analyzed. The target user training status information of the target to be analyzed in the personalized user motion data examples is fused with the important knowledge vector information of the target to be analyzed to obtain a fusion result of the target to be analyzed. Based on the fusion of the target to be analyzed The results are recommended and processed to obtain a recommended result of a sports safety training program. After determining the value of the evaluation index algorithm of the personalized user motion data processing network based on the recommended result of the sports safety training program and the training injury risk level identifier, it can be determined whether the value of the evaluation index algorithm of the personalized user motion data processing network exceeds the specified target value. When the value of the evaluation index algorithm of the personalized user motion data processing network exceeds the specified target value, the error information of the personalized user motion data processing network is determined based on the evaluation index algorithm of the personalized user motion data processing network, the error information is input into the personalized user motion data processing network for debugging, and each network coefficient is optimized during the debugging process.

[0074] Based on the above, a sports safety training program recommendation device is provided, which includes:

[0075] an information determination module, configured to extract knowledge vectors of the training injury risk levels from the personalized user motion data to be analyzed using a risk level prediction network for the training injury risk levels, and obtain knowledge vector information corresponding to the training injury risk levels of the target to be analyzed as secondary knowledge vector information of the target to be analyzed;

[0076] An information prediction module, configured to perform prediction processing based on the secondary knowledge vector information of the target to be analyzed to obtain target user training status information of the target to be analyzed;

[0077] A knowledge vector extraction module is used to extract important knowledge vectors from the personalized user motion data to obtain important knowledge vector information of the target to be analyzed;

[0078] A result fusion module is used to fuse the target user training state information of the target to be analyzed in the personalized user sports data to be performed with the important knowledge vector information of the target to be analyzed to obtain a fusion result of the target to be analyzed;

[0079] The result recommendation module is used to perform recommendation processing based on the fusion result of the target to be analyzed to obtain a sports safety training program recommendation result.

[0080] Based on the above, a sports safety training program recommendation system is shown, which includes a processor and a memory that communicate with each other. The processor is used to read a computer program from the memory and execute it to implement the above method.

[0081] Based on the above, a computer-readable storage medium is also provided, on which a computer program stored implements the above method when running.

[0082] In summary, based on the above scheme, by fusing the target user training status information of the target to be analyzed with the important knowledge vector information of the target to be analyzed, the secondary knowledge vector information of the target to be analyzed and the important knowledge vector information of the target to be analyzed are combined to extract more knowledge vector information of the target to be analyzed, which can accurately determine the stability of the rock, thereby improving the accuracy and reliability of the rock training effect recommendation; in addition, through the target user training status information of the target to be analyzed, the target to be analyzed is recommended, and the sports safety training plan recommendation result is obtained to ensure the accuracy of the recommendation, so as to ensure that the user receives personalized training and achieves the best training effect.

[0083] It should be understood that the system and its modules shown above can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated hardware. Those skilled in the art will understand that the above-mentioned methods and systems can be implemented using computer-executable instructions and / or contained in processor control code, for example, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the system and its modules of the present application. Not only can hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc. be implemented, they can also be implemented using software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software (for example, firmware).

[0084] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or any other possible beneficial effects.

Claims

1. A method for recommending a sports safety training program, characterized in that: The method comprises: By training a risk level prediction network for injury risk levels, extracting knowledge vectors of the training injury risk levels for the personalized user motion data to be performed, and obtaining knowledge vector information of the training injury risk levels corresponding to the target to be analyzed as secondary knowledge vector information of the target to be analyzed; Performing prediction processing in combination with the secondary knowledge vector information of the target to be analyzed to obtain target user training state information of the target to be analyzed; Performing important knowledge vector extraction processing on the personalized user motion data to be performed to obtain important knowledge vector information of the target to be analyzed; Fusing the target user training state information of the target to be analyzed in the personalized user sports data to be performed with the important knowledge vector information of the target to be analyzed to obtain a fusion result of the target to be analyzed; The recommendation process is performed in combination with the fusion result of the target to be analyzed to obtain a sports safety training program recommendation result.

2. The method according to claim 1, wherein The prediction network for personalized user motion data prediction includes multiple risk level prediction networks, and the multiple risk level prediction networks correspond to different training injury risk levels respectively; The step of performing prediction processing in combination with the secondary knowledge vector information of the target to be analyzed to obtain the target user training state information of the target to be analyzed includes: A secondary prediction process is performed using the risk level prediction network for the training injury risk level and in combination with the knowledge vector information of the training injury risk level corresponding to the target to be analyzed, to obtain the target user training status information corresponding to the training injury risk level of the target to be analyzed; The target user training status information of the target to be analyzed corresponding to the multiple training injury risk levels is fused to obtain the target user training status information of the target to be analyzed.

3. The method according to claim 2, wherein The performing secondary prediction processing on the knowledge vector information of the target to be analyzed corresponding to the training injury risk level to obtain the target user training status information of the target to be analyzed corresponding to the training injury risk level includes: performing knowledge vector elimination processing on the knowledge vector information of the target to be analyzed corresponding to the training injury risk level to obtain eliminated knowledge vector information; The eliminated knowledge vector information is weighted to obtain the target user training status information corresponding to the training injury risk level of the target to be analyzed.

4. The method according to claim 2, wherein The step of fusing the target user training status information corresponding to the plurality of training injury risk levels of the target to be analyzed to obtain the target user training status information of the target to be analyzed includes: When there are target user training status information corresponding to two different training injury risk levels, distinguishing between the two target user training status information corresponding to the two different training injury risk levels as the target user training status information of the target to be analyzed; When there is target user training status information corresponding to at least three different training injury risk levels, the target user training status information of the at least three different training injury risk levels is fused to obtain the target user training status information of the target to be analyzed.

5. The method according to claim 2, wherein Each of the risk level prediction networks includes multiple staged prediction networks, and the multiple staged prediction networks correspond to different attributes of the target to be analyzed; the risk level prediction network based on the training injury risk level performs knowledge vector extraction processing on the personalized user motion data to be performed, and obtains knowledge vector information corresponding to the training injury risk level of the target to be analyzed, including: For each attribute of the target to be analyzed, the following processing is performed: using a phased prediction network corresponding to the attribute, a knowledge vector extraction process is performed on the personalized user movement data to be analyzed to obtain knowledge vector information corresponding to the attribute of the target to be analyzed; The secondary prediction processing is performed by using the danger level prediction network of the training injury risk level and combining the knowledge vector information of the target to be analyzed corresponding to the training injury risk level to obtain the target user training status information of the target to be analyzed at the training injury risk level, including: For each attribute of the target to be analyzed, the following processing is performed: a secondary prediction process is performed using a phased prediction network corresponding to the attribute and in combination with the knowledge vector information of the target to be analyzed corresponding to the attribute, thereby obtaining target user training state information of the target to be analyzed corresponding to the attribute; The target user's training status information corresponding to multiple attributes of the target to be analyzed is fused to obtain the target user's training status information corresponding to the training injury risk level of the target to be analyzed.

6. The method according to claim 5, wherein The performing secondary prediction processing on the knowledge vector information of the attribute corresponding to the target to be analyzed to obtain the target user training state information corresponding to the attribute of the target to be analyzed includes: Performing knowledge vector elimination processing on the knowledge vector information of the target to be analyzed corresponding to the attribute to obtain eliminated knowledge vector information; Performing weighted processing on the eliminated knowledge vector information to obtain target user training status information corresponding to the attribute of the target to be analyzed; The step of performing weighted processing on the eliminated knowledge vector information to obtain the target user training status information corresponding to the attribute of the target to be analyzed includes: Mapping the eliminated knowledge vector information to the target user training status information to obtain a training effect coefficient arrangement; The target user training state information corresponding to the maximum coefficient in the arrangement of the training effect coefficients is used as the target user training state information corresponding to the attribute of the target to be analyzed.

7. The method according to claim 1, wherein Before performing the secondary knowledge vector extraction process on the personalized user motion data to be performed, the method further includes: performing differentiation processing on the personalized user motion data to be performed to obtain a plurality of training effect factor range data of the personalized user motion data to be performed; The performing of secondary knowledge vector extraction processing on the personalized user motion data to be performed to obtain secondary knowledge vector information of the target to be analyzed in the personalized user motion data to be performed includes: performing secondary knowledge vector extraction processing on each of the training effect factor range data to obtain secondary knowledge vector information of the target to be analyzed in the training effect factor range data; The performing prediction processing in combination with the secondary knowledge vector information of the target to be analyzed to obtain the target user training state information of the target to be analyzed includes: performing prediction processing through a prediction network and combining the secondary knowledge vector information of the target to be analyzed in each of the training effect factor range data to obtain the target user training state information of the target to be analyzed corresponding to each of the training effect factor range data; combining the target user training state information of the target to be analyzed corresponding to multiple ranges to obtain the target user training state information of the target to be analyzed; The step of combining the target user training status information corresponding to the target to be analyzed in the plurality of ranges to obtain the target user training status information of the target to be analyzed includes: Performing fusion processing on the target user training state information corresponding to the target to be analyzed in the plurality of said ranges to obtain fused target user training state information; Performing COUNTIF function processing on the fused target user training status information to obtain the target user training status information of the target to be analyzed; Among them, the target user training status information corresponding to the target to be analyzed in multiple ranges is combined and processed to obtain the target user training status information of the target to be analyzed, including: based on the user physical condition of the target to be analyzed represented by the target user training status information, determining the target user training status information with the largest user physical condition from the target user training status information corresponding to multiple training effect factor range data of the target to be analyzed, and using the target user training status information representing the largest user physical condition as the target user training status information of the target to be analyzed.

8. The method according to claim 1, wherein The recommendation processing is performed based on the fusion result of the target to be analyzed to obtain a sports safety training program recommendation result, including: Mapping the fusion results of the target to be analyzed to the training effect coefficient arrangement of the sports safety training program recommendation results; The training injury risk level corresponding to the maximum coefficient in the arrangement of the training effect coefficients is used as a recommendation result of a sports safety training program.

9. The method according to claim 1, wherein Before extracting the secondary knowledge vectors from the personalized user motion data, the method further includes: In response to a recommendation instruction for personalized user motion data to be performed, original personalized user motion data is obtained; and at least one of the following processes is performed on the original personalized user motion data: each data cluster in the original personalized user motion data is simplified, and the simplified original personalized user motion data is used as the personalized user motion data to be performed; Eliminating the influencing information in the original personalized user motion data, and using the original personalized user motion data after the elimination as the personalized user motion data to be processed; The target to be analyzed in the original personalized user motion data is analyzed and processed, and the analyzed original personalized user motion data is used as the personalized user motion data to be processed.

10. A sports safety training program recommendation system, characterized in that: The invention comprises a processor and a memory communicating with each other, wherein the processor is used to read a computer program from the memory and execute the computer program to implement the method according to any one of claims 1 to 9.

Images